# What else is changing in the oilfield? These are reviewed research notes, including updates to existing projects, adjacent operating technologies and historical context. They are not a count of unique deployments or additions to the AI register. Older operating periods and undated sources are labeled individually. The public record does not establish complete market coverage. Reviewed 7 October 2026. ## EQT — 2025 drilling/completion execution and per-foot cost performance R1-US-03 · Drilling & completions · FY2025 and Q4 2025 EQT reports fastest quarterly completions pace and most lateral footage drilled in 24 and 48 hours in Q4 2025; 2025 average well cost per foot was 13% lower YoY and 6% below internal expectations. Status: Observed FY2025 operating result; records repeated in Q4 2025. Economics: Average well-cost-per-foot KPI improved 13% YoY; not a dollar figure and not technology-attributed. Baseline: Internal expectation and year-over-year comparison are disclosed, but the release does not identify the technology or program that produced the difference, give absolute cost/foot, or adjust for well design, lateral length, basin/activity mix, or other operating changes. AI role: Unstated; do not attribute to AI or label as a specific non-AI technology. Limits: High-level management disclosure; no activity-normalized comparator, well design/geology mix, lateral length control, or attribution among crews, planning, compression, sand/water/power logistics and other factors. Next question: What is the activity- and design-normalized baseline behind the 13% YoY lower $/ft and 6% below plan, and which interventions caused it? - [EQT Reports Fourth Quarter and Full Year 2025 Results and Provides 2026 Guidance](https://ir.eqt.com/newsroom/news-releases/news-release-details/2026/EQT-Reports-Fourth-Quarter-and-Full-Year-2025-Results-and-Provides-2026-Guidance/default.aspx) — EQT; 2026-02-17; ‘Record Operational Efficiencies’ result near top of release. Access: Issuer results release opened; result section reviewed. - [Source reopening in Round 2](https://ir.eqt.com/newsroom/news-releases/news-release-details/2026/EQT-Reports-Fourth-Quarter-and-Full-Year-2025-Results-and-Provides-2026-Guidance/default.aspx) — See original source; 2026-02-17; 2026-02-17 release, 'Record Operational Efficiencies', lines 132-145. Access: Issuer results release opened; result section reviewed. ## EQT — Electric frac fleets using EQT-produced natural gas R1-US-04 · Drilling & completions · 2024 EQT reports 2.5 electric and 0.5 diesel frac fleets during 2024 and >24 million gallons of diesel eliminated; annual savings estimated at USD 61.15m. The same CDP table separately lists USD 7.262m investment required and <1 year payback. Keep these as issuer-reported fields; do not calculate ROI or assert that investment field is complete lifetime/fleet CAPEX. Status: Fleet electrification began in 2020; 2025 questionnaire reports on 2024 use/results. Treat as established 2024 operating evidence, not a newly introduced technology within 2025-2026. Economics: The $7.262m investment figure is the CDP response's 'investment required' field, not an independently reconciled full capital and lifecycle-cost bridge. EQT says implementation cost is expected to decrease as gas transport infrastructure develops. Baseline: Diesel fuel displacement is compared with diesel-powered frac operations. Emissions up to 70,000 tCO2e are expected/estimated versus historical diesel operation, not verified measured emissions savings. Vehicle delivery emissions were also not estimated. AI role: Non-AI physical power equipment as disclosed; no AI role stated. This classification describes the disclosed intervention, not every control system in the fleet. Limits: The savings/investment fields are issuer estimates, not independently controlled returns; the source does not reconcile the full cost boundary. The separate up-to-70,000-tonne CO2e avoidance is an energy-input comparison estimate, not a measured emissions outcome. Program started in 2020; source discloses 2024 results, not a 2025/2026 launch. Next question: Do the disclosed $61.15m annual savings and $7.262m “investment required” share the same fleet, time horizon, fuel-price assumptions and complete costs (gas, power, connection, rental, maintenance)? - [EQT Corporation 2025 CDP Questionnaire](https://esg.eqt.com/content/EQT%20Corporation%202025%20CDP%20Questionnaire.pdf) — EQT; 2025 (publication date to confirm from PDF metadata); Printed PDF p. 220, questionnaire section 7.55.2 (investment required, annual savings, payback); p. 47 energy input comparison (emissions estimate context). Access: Primary PDF text opened and printed page/section cross-checked in site/research/technology-scope/root-source-checks.json; screenshot attempt failed, so no visual PDF verification claimed.. - [Source reopening in Round 2](https://esg.eqt.com/content/EQT%20Corporation%202025%20CDP%20Questionnaire.pdf) — See original source; 2025 questionnaire, reporting year ended 2024-12-31; exact issue date not established; Printed p. 8 section 1.2 (USD); pp. 218-220 sections 7.55.1-7.55.2; row 2, lines 4873-4902. Access: Issuer-hosted PDF opened and text searched; currency and p. 220 fields verified. Existing local source check notes screenshots failed; no visual inspection claimed.. - [Additional root primary-source check](https://esg.eqt.com/content/EQT%20Corporation%202025%20CDP%20Questionnaire.pdf) — EQT; Date not established; Printed pages 47 and 220; sections 3.6.1.26 and 7.55.2, row 2. Access: Primary PDF text opened; screenshot attempt failed, so no visual PDF verification claimed.. ## EQT Corporation — Compression investments support 2026 production-outlook increase T10-R01-US-01 · Production & process operations · Q2 2026 results released 2026-07-21; full-year 2026 guidance updated at release. EQT reported 2026 production guidance raised by about 90 Bcfe, attributing it to better-than-expected benefits from compression investments that improved existing and new wells and shallowed decline rates. Full-year capital-spend guidance was reduced by $25 million, but the issuer did not provide compression-specific investment, cost savings, or a cost bridge. Status: Operating compression investments; Q2 2026 reported, while the increase is to full-year guidance and therefore forward-looking. Economics: Keep the guidance uplift, $25 million guidance reduction, and compression mechanism separate. Neither figure is an isolated compression ROI or a realized cash saving. Baseline: Production increase is versus prior 2026 guidance; Q2 sales volume was 634 Bcfe versus 568 Bcfe in Q2 2025, a companywide YoY comparison that includes multiple drivers. The $25m is guidance reduction, not an observed technology saving. AI role: No AI contribution disclosed. Physical compression investments are named; system pressure optimization is also cited in the operating result, without component-level detail. Limits: Issuer attribution and guidance. No compressor count, asset denominator, project capex, per-well uplift, uptime, matched baseline, or realized margin is disclosed. Do not count the $25m guidance reduction as compressor savings or add it to the full-year volume uplift. Next question: What compressor capex and operating cost, per-well production/decline response, and matched counterfactual support the ~90 Bcfe revised outlook, and how much of the $25m guidance reduction remains after separating unrelated efficiency/infrastructure effects? - [EQT Reports Second Quarter 2026 Results](https://ir.eqt.com/investor-relations/news/news-release-details/2026/EQT-Reports-Second-Quarter-2026-Results/default.aspx) — EQT Corporation; 2026-07-21; Highlights and CEO comments, lines 126–145; Q2 results show sales volumes lines 158–164.. Access: Issuer HTML opened. ## EQT Corporation — Olympus asset integration and deep Utica well-construction performance T20-R11-US-01 · Drilling & completions · Reported 21 October 2025 in EQT Q3 2025 results; specific well dates and the historic comparator period are not disclosed in the release. EQT said it achieved operational integration of all upstream and midstream assets acquired from Olympus Energy 34 days after closing, then drilled two deep Utica wells about 30% faster than Olympus' historic performance, saving $2 million per well. Status: Operating result reported after acquisition; all acquired assets reportedly operationally integrated 34 days after closing. Economics: Issuer-reported savings of $2 million per well for two wells. The release does not disclose absolute days, absolute well costs, scope of included cost, or whether savings are actual cash cost reduction versus comparison estimate. Baseline: Olympus historical performance; the number, lateral lengths, geology, service intensity, well designs, cost boundary and date of comparator wells are not specified. No randomized or matched-well analysis disclosed. AI role: No AI or named proprietary technology is identified as the cause. The claim concerns organizational/operational integration and well execution; absence of a disclosed AI role does not prove none was used. Limits: The result may reflect transferred operating practices, scale/integration, design, service market conditions or other factors. Do not attribute the improvement to a single technology or generalize beyond these two deep Utica wells. EQT's broader 2025 average well cost-per-foot disclosure is a separate, portfolio-level denominator. Next question: Can EQT disclose well-level actual days and cost scope, plus matched Olympus/EQT comparator wells, to identify which transferable practices drove the reported savings and how many eligible deep Utica wells remain? - [EQT Reports Third Quarter 2025 Results](https://ir.eqt.com/newsroom/news-releases/news-release-details/2025/EQT-Reports-Third-Quarter-2025-Results/default.aspx) — EQT Corporation; 2025-10-21; Release recent highlights, line 132; integration timing and two deep Utica wells. Per-unit disclosures and definitions at lines 264–270 are separate portfolio measures.. Access: Directly opened primary issuer HTML; relevant line and release context inspected.. ## EQT Corporation — Appalachian circular-water system reports >95% produced-water recycling in 2025 T20-R15-US-02 · Water, emissions & energy · Calendar 2025, disclosed July 20, 2026. EQT reports recycling more than 95% of produced water in 2025, attributing the ongoing result to its circular water-management strategy and approximately 245-mile integrated pipeline network. Status: Continuing operating program, with company-reported 2025 performance. Economics: No produced-water volume, pipeline capital/operating cost, treatment cost per barrel, freshwater purchases avoided, disposal cost avoided or measured completion cost impact disclosed in the issuer release. Baseline: No 2024 comparison, formula, water-quality specification, basin/pad boundary, recycled-water barrel denominator or third-party baseline is provided in the release; the statement should remain as issuer-reported percentage without an inferred ratio definition. AI role: No AI contribution disclosed. This is physical water handling/infrastructure and an operating practice; absence of AI mention is not evidence of no AI elsewhere in water operations. Limits: The source is a short issuer summary of its sustainability report; it does not show asset-level volumes or whether the percentage covers all produced water generated, water consumed in completions, or a specific operational subset. Do not claim the 245-mile network alone caused the rate or that the rate equals freshwater displacement. No process energy or disposal outcome is quantified. Next question: What 2025 produced-water volume and quality was recycled, at what treatment/pipeline cost per barrel and basin, and how much freshwater, third-party disposal and completion cost did that replace on eligible EQT pads? - [EQT Releases 2025 Sustainability Report, “Built for What’s Next”](https://ir.eqt.com/newsroom/news-releases/news-release-details/2026/-EQT-Releases-2025-Sustainability-Report-Built-for-Whats-Next-2026-8Z1uWdZ6uk/default.aspx) — EQT Corporation; 2026-07-20; Lines 127–143, especially 140–143 for produced-water recycling and network scope.. Access: Directly opened primary issuer release; relevant 2025 sustainability highlights inspected.. ## EOG Resources — iSense continuous leak-detection coverage in the Delaware Basin R1-US-06 · Water, emissions & energy · As of 2025-12-31 Coverage only. Source reports no alert precision/recall, leak repair rate, methane abatement, avoided loss, cost, or prior-year comparison. Status: Reported as achieved at year-end 2025. Economics: No cost or methane-reduction result disclosed alongside this coverage KPI. Baseline: Coverage metric, not an emissions reduction, leak-detection accuracy, response-time or avoided-loss result. No prior-year baseline included in headline. AI role: Not stated in this KPI disclosure; retain unclear and do not call it an AI result. Limits: Issuer-reported coverage only. Coverage does not establish detection performance, repair completion, emission abatement or economic payback; denominator is restricted to Delaware Basin gross oil handled at central tank batteries. Next question: What are alert precision/recall, repair closure, emissions avoided, loss prevention, and cost per covered gross-oil barrel? - [Sustainability – 2025 Highlights](https://www.eogresources.com/sustainability/) — EOG Resources; 2026 (2025 performance data); 2025 Highlights: iSense 99% coverage; footnote defines percentage of gross oil production at covered Delaware Basin central tank batteries as of year end. Access: Issuer sustainability page opened; KPI and footnote reviewed. - [Source reopening in Round 2](https://www.eogresources.com/sustainability/) — See original source; 2026 page with year-end 2025 performance; 2025 Highlights, iSense metric and footnote, lines 270-281. Access: Issuer sustainability page opened; highlight and footnote reviewed. ## EOG Resources — EOG keeps technology decisions close to its asset teams T20-R17-US-01 · Data, people & purchasing · Operating-model statements published March 2025; position posted 2025-10-29. EOG's March 2025 fact sheet describes in-house technical expertise, proprietary information technology and self-sourced materials as supporting well performance and cost control, and describes a decentralized, non-bureaucratic structure that enables asset-level value creation and multidisciplinary collaboration. A later Senior Drilling Engineer posting lists well planning/design, AFE preparation, drilling and cost analysis, vendor performance evaluation, new-technology application, cross-functional work with geology/reservoir/completions/land/construction/regulatory teams, and daily supervisory support for multiple rigs. Status: EOG fact sheet published March 2025; Senior Drilling Engineer position was posted 2025-10-29. The fact sheet describes company operating model; job posting describes the duties for a single open/posted role, not proof the role was filled or a measured organization-wide staffing structure. Economics: No cost, staffing headcount, software-development budget, decision-cycle time or productivity outcome is disclosed. This evidence helps identify workflow responsibility, not quantify the value of decentralization or proprietary IT. Baseline: No prior/alternative organizational structure or centralized-versus-decentralized performance comparator is supplied. The job posting is one role in one location, not a full org chart or staffing census. AI role: No AI role is stated in the sources. EOG describes proprietary information technology and data-driven decision-making at a high level; the job posting does not identify a named software or AI workflow. Limits: Corporate fact sheet language is high level. The job posting is a vacancy description and may not represent actual workload allocation, filled role, or all EOG basins. Proprietary IT is not identified by product, owner, maintenance model or AI content; do not infer AI leadership or superior readiness. Next question: Which well-planning, vendor, cost and daily rig decisions are made by asset-based engineers versus centralized technical/software teams, and what proprietary applications are owned and maintained in each layer? - [EOG Resources 2024 Operations & Financial Results Fact Sheet — Sustainable Value Creation](https://www.eogresources.com/documents/EOG_Fact_Sheet_2025.pdf) — EOG Resources; 2025-03; PDF p.1, lines 56–85: in-house expertise/proprietary IT, decentralized asset-level culture, multidisciplinary teams and data-driven decisions.. Access: Directly opened primary issuer PDF; relevant page inspected.. - [Senior Drilling Engineer — Midland, TX (Job ID 10871)](https://careers.eogresources.com/jobdetails.asp?apply=yes&jo_num=10871) — EOG Resources Careers; 2025-10-29; Lines 19–44, especially 36–37 for responsibilities and cross-functional interfaces.. Access: Directly opened primary EOG job listing; exact posted date and duties inspected.. ## EOG Resources — FY2025 average well-cost decline updates the existing lateral and drilling-efficiency program T10-R02-US-02 · Drilling & completions · FY2025 result published 2026-02-24; 2026 target stated in the same release. EOG's FY2025 average well costs decreased 7% across its multi-basin portfolio; 2026 low-single-digit reduction is guidance. The disclosure does not isolate savings from in-house motors or lateral length. Status: FY2025 reported outcome; 2026 cost-reduction language is a forward-looking target, not a realized result. Economics: A portfolio-average well-cost change is relevant to drilling capital productivity but is not a technology-specific cash saving, per-well dollar figure or return on the motor program. The 2026 target is a plan and should not be described as achieved. Baseline: FY2025 average well costs versus FY2024 average well costs, across a multi-basin portfolio; release does not provide an absolute $/well, basin/well denominator, lateral-length mix adjustment or comparable scope. The 2026 target is low-single-digit versus the FY2025 average-cost base as implied by the release. AI role: Issuer attributes the 2026 target to increased lateral lengths and other sustainable efficiencies, without identifying AI, a specific motor, or a single causal technology. Method mix is not disclosed. Limits: The reported change can reflect basin mix, lateral length, service pricing, well design and other efficiency actions. No attribution to the already-recorded in-house drilling-motor program or to AI is supported. Flat-oil production and capex are plan assumptions, not verified outcomes. Next question: Can EOG disclose basin-normalized $/well and $/lateral-foot changes, eligible well counts and the portion attributable to lateral length, motor performance, design, service pricing and other levers? - [EOG Resources Reports Fourth Quarter and Full Year 2025 Results](https://filecache.investorroom.com/mr5ir_eogresources2/444/4Q_2025_Earnings_Press_Release.pdf) — EOG Resources (issuer-hosted investor PDF); 2026-02-24; FY2025 results / 2026 capital plan paragraphs, PDF p. 1 (issuer-hosted release).. Access: Issuer-hosted PDF opened directly. ## EOG Resources — Extended laterals and in-house drilling motor program TECH-03 · Drilling & completions · 2025-02-27 CEO says better productivity plus sustainable efficiency improvements from extended laterals and EOG's in-house drilling motor program helped lower average well costs 6% in 2024. Status: Portfolio-wide, multi-basin average; number/share of wells using in-house motors, lateral-length distribution, and basin-level contributions are not stated. Economics: Not disclosed Baseline: Not disclosed AI role: non-AI drilling technology; no AI contribution disclosed Limits: The contribution of drilling motors cannot be isolated from extended laterals or other operating changes; no matched-well counterfactual disclosed. Next question: What well-cost reduction is attributable to motors vs longer laterals/productivity, by basin and design, with motor adoption share and incremental motor cost? - [EOG Resources Reports Fourth Quarter and Full-Year 2024 Results; Announces 2025 Capital Plan](https://investors.eogresources.com/2025-02-27-EOG-Resources-Reports-Fourth-Quarter-and-Full-Year-2024-Results-Announces-2025-Capital-Plan?asPDF=) — EOG Resources; 2025-02-27; Full-year highlights and CEO commentary. Access: See research audit. - [Source reopening in Round 2](https://investors.eogresources.com/2025-02-27-EOG-Resources-Reports-Fourth-Quarter-and-Full-Year-2024-Results-Announces-2025-Capital-Plan?asPDF=) — See original source; 2025-02-27; 2025-02-27 release, PDF p. 1 lines 49-54. Access: Issuer-hosted earnings PDF opened; CEO commentary reviewed. ## Devon Energy — D&C design, simulfrac and equipment efficiency bundle R1-US-02 · Drilling & completions · Q4/FY2025 metrics; earlier Q4 2024 call's savings are early wins at that time Q4 2025 slide 12 reports design improvements/new technology drove 9% faster drilling and 21% faster completions YoY; simulfrac application increased >30% YoY across portfolio; high-graded D&C equipment displaced ~40% of diesel and yielded >$25m savings; efficiency gains drove >$250m capital reduction for 2025, >5% improvement from original outlook. These are separate program metrics; >$250m is not an isolated technology result. The $600k/well early D&C savings stated in Feb 2025 were tied jointly to drilling/completion pace, supply-chain wins, self-sourced sand and simul-frac; they concern Williston progress and are not an AI-only result. Status: Operating program; Q4 2025 results report progress. Economics: >$250m reported 2025 capital reduction at initiative/program level; >$25m from high-graded equipment's diesel displacement; separate early 2024 call claim $600k/well bundled D&C savings. Baseline: 9%/21% are year-over-year rates without absolute duration, matched-well design, basin mix or counterfactual. >$250m is broad efficiency/program capital reduction; >$25m is equipment diesel displacement. Do not add these metrics or stack the earlier $600k/well figure into the 2025 total. AI role: AI models/agents are listed separately as flagging unproductive time and best practices. Source does not attribute D&C speed, equipment diesel displacement, simulfrac expansion, or aggregate capital reduction solely to AI. Limits: Issuer earnings deck and transcript. Several initiatives and commercial/supply-chain items are bundled; no counterfactual or independent cost audit. Separate the AI-monitoring line from physical D&C changes and do not double-count component metrics against the optimization headline. Next question: How much of >$250m capital reduction, 9% faster drilling, 21% faster completions and >$25m diesel savings is incremental to each named initiative, net of capex and not double-counted? - [Q4 2025 Earnings Presentation](https://investors.devonenergy.com/investors/events-presentations/event-details/2026/Q4-2025-Earnings-Presentation/default.aspx) — Devon Energy; 2026-02-17; Slide 12, Business Optimization Initiatives and Milestones. Access: Issuer IR presentation page verified; indexed slide text reviewed. - [Devon Q4 2024 Earnings Call Q&A Webcast Commentary Transcript](https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2024/q4/Q4-2024-DVN-Q-A-Webcast-Commentary-Transcript.pdf) — Devon Energy / Q4 CDN; 2025-02-19; PDF p. 4, lines 164-167. Access: Issuer-hosted transcript PDF opened. - [Source reopening in Round 2](https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2025/q4/Q4-2025-DVN-Earnings-Presentation.pdf) — See original source; 2026-02-17; Q4 2025 presentation dated 2026-02-17, printed slide 12 / PDF p. 11; initiatives and milestones. Access: Issuer-linked PDF opened; slide text reviewed. - [Source reopening in Round 2](https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2024/q4/Q4-2024-DVN-Q-A-Webcast-Commentary-Transcript.pdf) — See original source; 2025-02-19; Q4 2024 earnings call, 2025-02-19, PDF p. 4 lines 162-167; $600k/well early savings and named bundled causes. Access: Issuer-hosted PDF opened; relevant transcript lines reviewed. ## Devon Energy — Methane detection data integrated into field decision workflows, with automated repair work orders and SCADA-based AI alerts T10-R06-US-01 · Water, emissions & energy · 2024 deployment and coverage reported in Devon's 2025 Sustainability Report; issuer page adds 2025 new-facility monitoring practice but is undated/current. The 2025 report (stating default reporting period is year ended 2024) says Devon integrates methane detection data with operating systems/decision support centers, uses an AI program analyzing SCADA and decision-support-center data to alert action on operational issues, and automated leak-repair work orders. Continuous monitoring covered 46% at year-end 2024 and combined detection methods covered 100% of production; these are coverage, not emissions-reduction results. Status: Operating detection and dashboard/work-order workflow in the 2024 reporting year; AI program described as being deployed, not as a fully evaluated portfolio-wide system. Carbon accounting platform was in use for emissions inventory reporting. Economics: The software and workflow are enabling infrastructure for faster detection, mitigation and emissions reporting. No alert precision, time-to-work-order, time-to-repair, leak recurrence, production uptime, methane mass avoided, labor-hours saved or implementation cost is disclosed; no software ROI can be calculated. Baseline: Monitoring coverage and platform granularity are disclosed, not before/after performance. The 46% continuous-monitoring figure and 100% combined-tech coverage are distinct; neither is a measured emissions reduction. Equipment-level versus basin-level is a reporting-resolution comparison, not proof of more accurate direct emissions measurement or cash savings. AI role: Hybrid. Devon explicitly says it is deploying an AI program that analyzes SCADA and decision-support-center data and generates alerts for operational issues. The surrounding workflow includes continuous/emissions sensors, operator decision centers, dashboards, an automated work-order process and standard procedures. The source does not describe model type, production-control authority, or autonomous actuation; AI alerts are human action prompts as described. Limits: Do not merge emissions coverage or outcomes with EOG or Exxon. Report does not quantify repair time avoided, leaks prevented, methane volume, AI alert precision, or incremental benefit versus the non-AI workflow. Next question: What are the end-to-end metrics from sensor/model alert to verified repair—precision/recall, work-order creation and closeout time, leak recurrence and measured methane abatement—and what portion is attributable to software versus field staffing/process redesign? - [Devon Energy 2025 Sustainability Report](https://dvnweb-hfhahvbraba9feev.a01.azurefd.net/assets/documents/Sustainability/DVN_2025_SustainabilityReport.pdf) — Devon Energy; 2025 (exact issue date not confirmed in the PDF metadata checked); Printed pp. 9 and 15-16; PDF lines 678-700 (carbon accounting) and 1215-1258 (coverage, integration, AI alerts, dashboard, work orders).. Access: Issuer PDF opened directly. - [Reducing methane emissions](https://www.devonenergy.com/reducing-methane-emissions) — Devon Energy; Undated current issuer page (copyright 2026); Lines 107-121, including 2024 coverage, 2025 new-facility practice, AI program, dashboard and automated work order description.. Access: Issuer HTML opened directly. ## Devon Energy — 2025 optimization plan identifies vendor management and analytics as value levers, without naming technology suppliers or spend T10-R07-US-01 · Production & process operations · Plan targeting approximately 30% of estimated improvements by year-end 2025 and remaining value by year-end 2026; full annual improvement target stated at announcement. Devon announced a target of $1.0bn annual pre-tax free-cash-flow improvement, including $250m production-optimization target using advanced analytics and $300m capital-efficiency target including vendor management; these are plan estimates/targets with staged timing, not audited realized tech returns. Status: Plan announced 2025-04-22; by announcement the company said analytics/process automation advancements had been implemented. The $300m 2025 cash-flow uplift and $1bn annual pre-tax improvement goals were forecasts at publication, not verified outcomes in this source. Economics: The $300m and $250m figures are broad target buckets, not realized savings, tech spend, supplier revenue or ROI. Total $1bn target includes commercial and corporate actions. The separate approximately $300m cash-flow uplift anticipated by year-end 2025 combined marketing agreements and implemented technology, with no technology-only allocation in this announcement. Baseline: No audited baseline, asset/well population, supplier-level spending or savings mechanism is supplied in this plan release. Approximately 30%/remainder are expected milestone shares of forecast improvement, not measured technology deployment percentages. AI role: The production-optimization bucket cites advanced analytics; the release also says analytics and process automation had been implemented. It does not disclose AI use, a vendor/platform, or how any technology contributes to the value target. Limits: Issuer plan announcement and forward-looking targets. No named software/service vendor, procurement process, project capex, contract duration, deployment count, individual KPI, or confirmed realization is supplied. D&C efficiency material overlaps R1-US-02; analytics/work-order workflow overlaps T10-R06-US-01. Do not add the $300m/$250m buckets to observed D&C or AI results. Next question: Which vendor contracts, internal build costs and asset-level pilots sit behind the $300m capital-efficiency and $250m production-optimization buckets, and what portion of their target value was actually realized by the 2025 milestone? - [Devon Energy Unveils Value Enhancing Business Optimization Plan](https://investors.devonenergy.com/investors/press-releases/press-release-details/2025/Devon-Energy-Unveils-Value-Enhancing-Business-Optimization-Plan/default.aspx) — Devon Energy; 2025-04-22; Plan pathway and category buckets, lines 42-60.. Access: Issuer HTML opened directly. - [Devon Energy and Coterra Energy to Combine, Creating a Premier Shale Operator](https://investors.devonenergy.com/investors/press-releases/press-release-details/2026/Devon-Energy-and-Coterra-Energy-to-Combine-Creating-a-Premier-Shale-Operator/default.aspx) — Devon Energy; 2026-02-01; Transaction benefits, lines 50-66, especially AI platform strategy and $1bn synergy target.. Access: Issuer HTML opened directly. ## Devon Energy / Whitson — Devon calculates daily well pressure through Snowflake and Whitson T20-R16-US-01 · Data, people & purchasing · Undated vendor case; operating start and research-window eligibility are not established. Snowflake's case says Devon consolidates data from 30+ sources, gives employees access to 95% of application data, and supports more than 1,000 users and about 9 million queries per month. With Snowflake Secure Data Sharing, Devon sends time-series and wellbore-configuration data to Whitson; Whitson returns physics-based bottomhole-pressure calculations daily for 5,000+ wells. Calculations are used by production engineers to spot underperforming wells and by reservoir engineers to understand reservoir condition, site new wells and plan development. Status: Described as a live daily calculation/use case at the time of the undated customer story; deployment date and present status are not stated. Economics: The story states manual calculation across >5,000 wells did not scale and that installing a $50,000 physical pressure sensor in each existing well was uneconomical/unrealistic. It gives no actual sensor count avoided, software/data-platform cost, manual labor cost, calculation error rate, production gain, or realized savings. Do not multiply the quoted sensor price by the well count as realized savings. Baseline: Manual derivation was the prior workflow; the page provides no time-per-calculation, calculation accuracy, intervention, well-production or matched baseline before/after data. AI role: No AI is attributed to the bottomhole-pressure calculation. The customer story identifies Whitson's physics-based engine/digital-twin capabilities; Snowflake is the governed data-sharing layer. Devon says unified data is intended to support separate future AI/ML use cases, including geoscience work, which are not this workflow. Limits: Vendor-authored, undated customer story; current status/date of deployment is unknown. The 5,000+ figure is wells for which daily physics-based calculations are produced, not a documented production-uplift or model-accuracy denominator. No sensor count was avoided: $50,000 per sensor is an illustrative hardware alternative, not realized savings. Snowflake's site categorizes multiple products including AI, but this specific workflow is described as Whitson physics-based calculation via secure data sharing; no AI attribution is established. Next question: When was this workflow deployed and is it still active; what are the per-well calculation cost/accuracy, engineer hours, intervention changes and production impact relative to manual or sensor measurements, and what were Snowflake/Whitson integration and recurring costs? - [Devon Energy and Whitson Collaborate to Save Time and Increase Production by Automatically Calculating Bottomhole Pressure](https://www.snowflake.com/en/customers/all-customers/case-study/devon-energy/) — Snowflake customer story, with Devon Energy and Whitson quotations; Undated; accessed 2026-10-07; Lines 15–20 (headline metrics), 57–65 (platform reach), 66–80 (physics-based pressure calculation and workflow), 82–85 (future AI/ML data work).. Access: Directly opened full vendor-hosted case page; publication date not present in visible text.. ## Devon Energy — Devon puts digital and operational technology under its CTO T20-R17-US-02 · Data, people & purchasing · As announced 2025-01-13; no subsequent organization-chart update checked in this pass. Devon promoted Trey Lowe to SVP and Chief Technology Officer, while naming separate SVP roles for E&P asset management and E&P operations. The release says the CTO's team includes Devon Energy Ventures, operational technology, IT and other digital teams. Status: Leadership structure announced January 13, 2025; the announcement says Trey Lowe was promoted to SVP and Chief Technology Officer and describes his team scope. Economics: No technology organization budget, headcount, operating KPI or financial contribution disclosed. Baseline: No before/after organizational comparison or explicit decision-rights map disclosed. AI role: The source mentions no AI deployment or AI-specific decision rights. The CTO scope includes OT, IT and digital functions, without describing model development or production workflow controls. Limits: A leadership release documents executive scope, not day-to-day ownership of specific systems or asset workflows. Team makeup and reporting lines may have changed since Jan. 2025; this pass did not verify current org structure. Do not treat executive centralization as proof of operating autonomy, AI adoption or improved performance. Next question: Who currently owns production software, data infrastructure, model validation and on-asset workflow change, and which decisions remain with E&P asset/operations leaders? - [Devon Energy Announces Updates to Executive Leadership Team](https://investors.devonenergy.com/investors/press-releases/press-release-details/2025/Devon-Energy-Announces-Updates-to-Executive-Leadership-Team/) — Devon Energy; 2025-01-13; Lines 27–43, especially 35–43 for executive appointments and CTO team scope.. Access: Directly opened primary issuer release; leadership and team scope inspected.. ## Diamondback Energy — Continuous pumping reaches first full operating quarter T10-R01-US-02 · Drilling & completions · Q2 2026; release dated 2026-08-03. Diamondback reported average pumping time of 21.3 hours per day in its first full quarter of continuous pumping, corresponding to about 4,700 lateral feet completed per day. In the same operations paragraph, equipment cost per well fell about 14% quarter-over-quarter; the company does not expressly attribute the full cost decline to continuous pumping. Status: First full quarter of continuous pumping reported in Q2 2026; management says it is operating at scale, but fleet/well denominator is not supplied. Economics: Preserve 21.3 pumping hours/day and ~4,700 lateral feet/day as reported first-full-quarter metrics; ~14% QoQ decline is equipment cost per well, not attributed to continuous pumping alone. Flaring reduction and protected oil are attributed to gas-offload strategy, not pumping. Baseline: 21.3 pumping hours/day and ~4,700 lateral ft/day are Q2 2026 reported averages. Equipment cost comparison is Q2 vs Q1 2026. AI role: No AI role disclosed. Conventional completion process/equipment innovation; source provides no ML/AI attribution. Limits: Issuer-reported operating metrics; no crew/fleet count, eligible well denominator, prior pumping hours/day, stage count, well design adjustment, cash cost/ft, capital cost of custom fleets, or matched control. Commodity/pricing environment, activity and equipment mix can affect comparisons. Next question: Across how many fleets and wells was the 21.3 hours/day achieved, what was the prior/control pump-hours and $/lateral-foot baseline, and what custom equipment/integration costs were required? - [Diamondback Energy, Inc. Form 8-K, Exhibit 99.1 (Second Quarter 2026 Results)](https://ir.diamondbackenergy.com/static-files/3edc9cca-d7ee-45e5-b420-1b2b8fab639f) — Diamondback Energy; 2026-08-03; Q2 2026 Operational Performance, PDF p.22 / lines 719–725.. Access: Issuer-hosted PDF opened. ## Diamondback Energy — Diamondback is still testing chemical EOR; recovery gains remain undisclosed T20-R18-US-01 · Subsurface & recovery · Initial pilot/investment described as late 2025; first update published 2026-02-23, latest update published 2026-08-03. On 2026-02-23, Diamondback said it invested approximately $30 million in late 2025 on a surfactant pilot testing 60 wells, with positive results and additional tests planned in 2026. On 2026-08-03, it said its 50-well pilot program had been completed in H2 2025 and the second batch of well tests was still flowing back, with 'encouraging' results. The sources do not state whether 60 and 50 refer to the same cohort or different phases/populations. Status: February 23, 2026 letter: company reports approximately $30 million invested in late 2025 on a surfactant pilot testing 60 wells, with positive results and more testing planned. August 3, 2026 Q2 call: company says a 50-well pilot program was completed in H2 2025 and a second batch was still flowing back. The available sources do not reconcile the 60-well and 50-well counts or establish whether batches overlap; no commercial deployment established. Economics: Approximately $30 million is disclosed for the late-2025 pilot investment; the release does not clarify whether this is total pilot cost, product-only cost or capital plus operating cost. No incremental oil, recovery-factor uplift, treatment cost per well/barrel, decline response, payout, cash benefit or control-well comparison is disclosed. Baseline: No untreated control population, pre-treatment production baseline, production normalization, well count per second batch or measurement window is given. The 'positive'/'encouraging' language is qualitative issuer characterization. AI role: No AI role disclosed; the initiative is a physical surfactant-based recovery test and associated well evaluation. Limits: The 50-versus-60 well-count mismatch needs source clarification before using an eligible-well denominator. Flowback is an interim stage and not a recovery factor or lasting production outcome. The pilot investment cannot be translated to cost per treated well until scope is clear; no cost/benefit or scale decision can be derived. Primary operator disclosures, but no independent technical verification. Next question: Do the 50-well and 60-well counts describe one cohort or separate phases, what was included in the $30m, and what normalized oil response versus matched untreated wells persisted after flowback across each formation and treatment design? - [Letter to Stockholders Issued by Diamondback Energy, Inc. (FY2025 results)](https://ir.diamondbackenergy.com/news-releases/news-release-details/letter-stockholders-issued-diamondback-energy-inc-10) — Diamondback Energy; 2026-02-23; Lines 81–82; approximately $30m, late-2025 surfactant pilot testing 60 wells, positive results and planned additional tests.. Access: Directly opened primary issuer HTML; exact passage inspected.. - [Form 8-K — Second Quarter 2026 Shareholder Letter](https://ir.diamondbackenergy.com/static-files/3edc9cca-d7ee-45e5-b420-1b2b8fab639f) — Diamondback Energy; 2026-08-03; PDF pp. 22–23; lines 726–735; completed 50-well pilot program and second-batch flowback status.. Access: Directly opened primary issuer-hosted PDF; relevant operations passage inspected.. ## Diamondback Energy — Diamondback’s first chemical EOR pilot leads to follow-on testing T10-R08-US-01 · Subsurface & recovery · Pilot completed H2 2025; second-batch flowback underway in Q2 2026; issuer disclosure dated 2026-08-03. Diamondback says its 50-well chemical-EOR pilot was completed in H2 2025 and a second batch of well tests was flowing back in Q2 2026, with qualitative 'encouraging' language but no quantified recovery factor or production result in the cited passage. Status: Initial 50-well pilot completed in H2 2025; as of Q2 2026, second batch of well tests was flowing back. Commercial-scale deployment and the decision to invest behind the method were still prospective. Economics: No incremental oil recovery, cumulative response period, intervention cost, chemical use, production impact, decline-adjusted baseline or payout is disclosed. The source's characterization of recovery improvement as potentially high-return is forward-looking value framing, not observed project economics. Baseline: No control wells, untreated comparison group, pre-treatment production decline, treatment timing, eligible well inventory or sample count for the second batch disclosed. 'Encouraging' has no quantitative definition in the source. AI role: No AI role or test method disclosed in this operational status passage. Do not infer AI or assume the process is non-AI based on silence. Limits: Issuer status disclosure only. A 50-well count establishes pilot activity, not 50 successful responses or independent statistical support. Second-batch well count and result duration are absent. No chemicals, injection method, candidate selection, pressure/interference effects, water disposal consequences or failure cases disclosed. No evidence of technical failure or abandonment is provided; the commercial/scaling stage remains unresolved. Next question: What were the treatment design, treated and control well counts, oil-recovery uplift per well over a defined duration, chemical and intervention cost, adverse effects and gate criteria for a scaled capital commitment? - [Form 8-K for Diamondback Energy Inc. filed Aug. 3, 2026 (Q2 2026 results)](https://ir.diamondbackenergy.com/static-files/3edc9cca-d7ee-45e5-b420-1b2b8fab639f) — Diamondback Energy; 2026-08-03; Q2 2026 Operational Performance, PDF p. 22, paragraph on chemical EOR tests and second-batch flowback; lines 726-735.. Access: Issuer-hosted PDF opened directly. ## Diamondback Energy — Diamondback’s Barnett drilling record uses a narrower cost boundary T10-R09-US-01 · Drilling & completions · Q1 2026 operating result disclosed 2026-05-04. Company presentation's delineation/full-development comparison was available in 2026; source presents the ~$800/ft value as an expected outcome. Q1 2026: one fastest-in-company-history (~18-day) Barnett well at lowest Barnett cost to date (<$400/lateral foot). Separate 2026 investor material: current delineation wells at about $1,000/ft; full-field well costs expected to fall ~20% to ~$800/ft. These represent separate well populations and stages. Status: Observed single-well company record in Q1 2026; full-field cost reduction is a forward-looking estimate, not yet achieved at development scale. Economics: Keep <$400/ft single-well record out of the ~$1,000-to-$800 development estimate bridge. Neither reveals complete capex cost scope or eligible-population distribution; ~$800 is expected, not realized. Baseline: Record well compared to prior Diamondback Barnett wells; no absolute baseline or matched well details given. The full-field estimate compares ~$800/ft expected against ~$1,000/ft current delineation wells. Do not substitute the single-well <$400/ft outlier as the current cost baseline for the full-field target. AI role: No AI method attributed to the reported well cost; source emphasizes operational execution and speed. Do not infer AI role from parallel automation programs. Limits: Operator-reported; new-play delineation and full-field development have different geology, learning, pad and infrastructure scope. 'Lowest cost to date' is a single-well internal record. Q1 disclosure lacks lateral length, exact capex numerator and included cost components; full-field cost and return remain forecasts. Not proof of basin-wide economic parity without repeat wells and decline-adjusted production. Next question: What is the repeatable, fully loaded drilling-and-completion cost per completed lateral foot for the defined Barnett development inventory, with well count, lateral length, design, infrastructure, production type curve and facility cost included? - [Form 8-K for Diamondback Energy Inc. filed May 4, 2026 (Q1 2026 results)](https://ir.diamondbackenergy.com/static-files/a1520480-f925-42c2-8f51-4b4223086323) — Diamondback Energy; 2026-05-04; Q1 2026 Operational Performance, PDF p. 23, lines 717-727.. Access: Issuer-hosted PDF opened directly. - [Diamondback Energy Q4 2025 results / Barnett overview investor materials](https://ir.diamondbackenergy.com/static-files/69e4a842-30fe-4ce0-8b92-510ad5aff2e6) — Diamondback Energy; 2026 (issuer 2026 materials; exact presentation date to verify before external citation); Barnett overview / development economics: ~$1,000/ft delineation, expected ~$800/ft full development; source search-indexed in this pass, PDF body not directly opened.. Access: Issuer-hosted document surfaced by search; body not reopened in this pass. ## Diamondback Energy — Associated-gas microgrid and electric simul-frac TECH-05 · Water, emissions & energy · 2026-05-04; 2025 disclosure Electrified coverage is reported; no quantified realized cost savings per well/stage or diesel-comparable baseline is disclosed. Contract capacity is not an outcome. Status: Q1 2026 letter says about 80% of frac fleet was electrified. Separately, sustainability/CDP disclosure describes a microgrid serving a producing field and drilling/completions, and a December 2024 agreement for four electric simul-frac fleets with approximately 200 MW of capacity. Economics: Not disclosed Baseline: Not disclosed AI role: Non-AI electrification/power infrastructure. Do not infer AI contribution from vendor platform descriptions. Limits: Separate achieved coverage from agreement/capacity; the claimed economic benefit is directional and not measured against a disclosed diesel baseline. Next question: What diesel gallons, power/fuel costs, stage/day utilization and total hookup/capex support the ~80% electrified fleet claim and savings after gas/fuel logistics? - [Letter to Stockholders Issued by Diamondback Energy, Inc. (Q1 2026); 2025 Sustainability/CDP disclosure](https://ir.diamondbackenergy.com/news-releases/news-release-details/letter-stockholders-issued-diamondback-energy-inc-11) — Diamondback Energy; 2026-05-04; 2025 disclosure; Q1 letter, operating performance; sustainability/CDP PDF pp. 84, 87. Access: See research audit. - [Source reopening in Round 2](https://ir.diamondbackenergy.com/news-releases/news-release-details/letter-stockholders-issued-diamondback-energy-inc-11) — See original source; 2026-05-04; Q1 2026 stockholder letter, electrified frac-fleet statement and directional rationale. Access: Diamondback issuer page opened; text accessible. - [Source reopening in Round 2](https://www.diamondbackenergy.com/static-files/faf5ab25-5ab5-4404-8c04-c7bd387ae418) — See original source; 2025 report; agreement disclosed 2024-12; Microgrid powers producing field and drilling/completions; four-fleet/approximately 200-MW contract scope. Access: Diamondback 2025 sustainability report PDF opened. - [Source reopening in Round 2](https://voltagrid.com/diamondback-energy-halliburton-and-voltagrid-sign-agreement-for-four-electric-simul-frac-fleets-in-the-permian-basin) — See original source; 2024-12-13; Agreement to deploy four electric simul-frac fleets, approx. 200 MW. Access: VoltaGrid partner announcement opened. ## Chord Energy — Chord’s first seven four-mile wells come in below their initial budget T20-R12-US-01 · Drilling & completions · 2025 operating results, reported 2026-02-25. Chord reported seven four-mile wells turned in line in 2025. Initial performance was meeting or exceeding expectations, average costs were below budget, and tracer data for all seven indicated contribution from the full lateral. The company said four-mile well costs were reduced by more than 10% versus its 2025 initial budget designs, supported by strong performance and efficient execution including single-run cleanouts. Status: Operating wells with early performance reported; first four-mile DSU development on the Toonie pad had begun fracturing and was scheduled to come online in Q1 2026 at disclosure. Economics: Reported >10% reduction versus initial 2025 budget designs; no absolute cost per well or capital amount disclosed in the cited release. This is a company-reported budget comparison, not audited realized savings or a comparison with two-mile wells. The 2025 capital figure and production result are broader portfolio data and should not be assigned to this technology. Baseline: Chord's initial 2025 budget designs for four-mile wells. The release does not specify design assumptions, sample selection, actual well count in the average-cost calculation, lateral-length convention or realized cost boundary. AI role: No AI contribution is disclosed for the extended-lateral design, tracer evaluation or single-run cleanout. A separate AI/ML rod-lift result in the same release is already recorded as AI-097 and is not part of this finding. Limits: Early well results may be affected by reservoir quality, spacing, working interest, service prices and design changes. Tracer indication supports full-lateral contribution but is not proof of uniform production, EUR uplift or economic superiority. The source does not report a matched shorter-lateral comparator, full-cycle recovery, payback, or eligible inventory count. Next question: For additional eligible four-mile DSUs, what are realized D&C cost per lateral foot, cycle time, production/EUR versus matched two- or three-mile wells, and the cost of tracer/cleanout changes after normalizing for location and service prices? - [Chord Energy Reports Fourth Quarter and Full-Year 2025 Financial and Operating Results, Issues 2026 Outlook and Declares Base Dividend](https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend) — Chord Energy; 2026-02-25; Operations Update, lines 210–214; reported deployment, cost and tracer scope. Lines 84–85 summarize seven wells and program outlook.. Access: Directly opened primary issuer HTML; full relevant operating update inspected.. ## Chord Energy — Chord expands its four-mile lateral program to 26 executed wells T20-R19-US-01 · Drilling & completions · 2025 pilot/development and early 2026 scale-up; issuer updates published 2025-11-04 and 2026-08-05. Longitudinal reports describe distinct vintages: November 2025—four four-mile laterals TIL YTD, seven expected for FY2025, with the Rystedt well matching two two-mile analog production after six months and named tracer data showing full-lateral contribution; February 2026—seven four-mile wells TIL in 2025, full-lateral tracer indication across all seven, and average cost >10% below initial 2025 budget design; August 2026—26 four-mile wells executed as of early Q3, costs/early performance in line with expectations, plus the Basin's first trimulfrac completion and selective continued use planned. These measures have different cohorts and baselines; no absolute cost, normalized savings or trimulfrac-specific benefit is disclosed. Status: As of early Q3 2026, Chord reported 26 total four-mile wells executed; the release separately says early program performance and costs were in line with expectations. It does not establish that all 26 were TIL or producing. The Basin's first trimulfrac completion had been executed; the source does not give a distinct well count or trimulfrac cost/savings result. Economics: The only quantified cost statement in the reviewed history is >10% below Chord's initial 2025 design budgets for the seven-well 2025 program. Absolute dollars, cost per lateral foot, included D&C/facilities/land/water costs, budget rebasing, the 26-well cohort cost and trimulfrac-specific savings are not disclosed. The six-month production comparison for Rystedt is early-time only and lacks matched capital and EUR data. Baseline: The 2025 cost claim compares average four-mile well cost against initial internal 2025 design budgets; exact well count in the average, scope and budget assumptions are not given. The Rystedt comparison uses two two-mile analog wells, but matching criteria and production normalization are not disclosed. The August 2026 'in line with expectations' baseline is unspecified. AI role: No AI role is stated for four-mile lateral execution or trimulfrac. Chord separately mentions continuing AI artificial-lift optimization in the same August release; that is a distinct workflow and is not attributed to the four-mile program. Limits: Population and method scope differ across disclosures: seven 2025 wells versus 26 executed by early Q3 2026, and the later release does not identify how many used trimulfrac. The releases do not state whether all executed wells are TIL, their working interest, eligible future inventory, or whether every pad can accommodate four-mile geometry. Reported early performance and budget variance are issuer claims, not independently validated causal savings. Do not compare Chord's >10% below initial budget directly with another operator's cost/foot or equipment cost/well metric. Next question: For the next eligible four-mile DSUs, what is actual full D&C cost per completed lateral foot versus matched shorter laterals after normalizing geology, service prices, facilities, working interest and budget vintage; and what incremental cost and cycle-time effect does trimulfrac show against the prior completion design? - [Chord Energy Reports Second Quarter 2026 Financial and Operating Results, Declares Base Dividend and Updates 2026 Outlook](https://ir.chordenergy.com/2026-08-05-Chord-Energy-Reports-Second-Quarter-2026-Financial-and-Operating-Results%2C-Declares-Base-Dividend-and-Updates-2026-Outlook) — Chord Energy; 2026-08-05; Lines 156–161: 26 four-mile wells as of early 3Q26; capital cost and early-performance status; first basin trimulfrac and selective continuation; AI lift optimization separately noted.. Access: Directly opened primary issuer HTML; exact operations-update passage inspected.. - [Chord Energy Reports Third Quarter 2025 Financial and Operating Results, Declares Base Dividend and Issues Updated Outlook](https://ir.chordenergy.com/2025-11-04-Chord-Energy-Reports-Third-Quarter-2025-Financial-and-Operating-Results%2C-Declares-Base-Dividend-and-Issues-Updated-Outlook) — Chord Energy; 2025-11-04; Lines 159–164: four four-mile wells YTD, seven expected FY25; Rystedt six-month production versus two two-mile analogs; named examples and execution methods.. Access: Directly opened primary issuer HTML; exact operations-update passage inspected.. - [Chord Energy Reports Fourth Quarter and Full-Year 2025 Financial and Operating Results, Issues 2026 Outlook and Declares Base Dividend](https://ir.chordenergy.com/2026-02-25-Chord-Energy-Reports-Fourth-Quarter-and-Full-Year-2025-Financial-and-Operating-Results%2C-Issues-2026-Outlook-and-Declares-Base-Dividend) — Chord Energy; 2026-02-25; Operations update inspected previously in round 12, lines 210–216: seven four-mile 2025 wells; average costs >10% below initial 2025 budget designs.. Access: Primary issuer page indexed and directly opened in round 12; not freshly reopened for this pass. Earlier directly inspected text is retained in round-12-us.json.. ## ConocoPhillips — Temporary fiber-optic sensing with adaptive frac-stage control R1-US-01 · Drilling & completions · Recent trial; exact date not stated Surface equipment adjusted stage volumes by as much as -5% to +25%. Executive says this approach could lower completion cost by $5-$15 per foot while improving recovery and reducing frac hits. Status: Trial disclosed as recently completed; no broader deployment status established. Economics: Potential $5-$15 per lateral foot reduction; not booked/measured outcome in the disclosure. Baseline: No well/stage denominator, comparator, absolute volumes, or baseline cost specified. The stage-volume change is reported output; $5-$15/ft is potential benefit, not a realized saving. AI role: Unclear/hybrid. Automation is explicit; AI is neither attributed nor ruled out. Do not call this an AI result. Limits: Issuer executive interview; trial size, duration, control algorithm, eligible-well share, economics and independent verification are absent. Do not infer sustained uplift or commercial rollout. Next question: On how many stages/wells was fiber deployed, and what matched-well $/ft, recovery and frac-hit changes were realized after cable/control costs? - [Q&A / Nick Olds on unlocking Lower 48 results through scale, inventory and tech](https://www.conocophillips.com/spiritnow/story/qa-nick-olds-on-unlocking-lower-48-results-through-scale-inventory-and-tech/) — ConocoPhillips; 2026-03-24; Lines 99-107: advanced diagnostics, first-of-its-kind fiber trial, autonomous stage-volume adjustments, potential cost range and distinct AI gas-lift statement. Access: Issuer page opened; full relevant section reviewed. - [Source reopening in Round 2](https://www.conocophillips.com/spiritnow/story/qa-nick-olds-on-unlocking-lower-48-results-through-scale-inventory-and-tech/) — See original source; 2026-03-24; Published 2026-03-24; lines 99-107, especially fiber-optic trial and potential benefits. Access: Issuer page opened; full text reviewed. ## ConocoPhillips — Canada facility digital twins for maintenance and work planning R1-US-05 · Maintenance & reliability · Use cases observed/reported by May 2025 COP says workers used the 3D digital twin to locate a hidden spool, isolate/blind it without extensive physical searching or disruptions, and avoid unnecessary downtime/production halts. This is one qualitative case; no time, cost, uptime or production amount is given. Status: Digital twin adopted for Surmont CPF2 and Montney C-11-K by May 2025. Montney then had limited functionality; SAP/document/data integration upgrade was still planned for summer 2025. Full build-out of main Canadian facilities was estimated to take until end-2026. Economics: Potentially avoided search/maintenance delay and crane/scaffold waste; no measured amount. Baseline: Anecdotal task comparison to physical searches and 2D plot plans; no measured baseline or controlled comparison. AI role: AI role not stated. Digital twin/data visualization is established; do not infer no-AI architecture from silence. Limits: Issuer story with qualitative user examples; adoption and economic benefit not quantified. Montney model had less functionality and integrations were still being built. Next question: What downtime hours/barrels deferred and field labor/search cost were avoided per twin-enabled work order versus standard document search? - [Visualizing success / Canada rolls out digital twins for Surmont and Montney](https://www.conocophillips.com/spiritnow/story/visualizing-success-canada-rolls-out-digital-twins-for-surmont-and-montney/) — ConocoPhillips; 2025-05-27; Canada deployment, work use cases, hidden-spool example and build-out status. Access: Issuer page opened; full relevant passage reviewed. - [Source reopening in Round 2](https://www.conocophillips.com/spiritnow/story/visualizing-success-canada-rolls-out-digital-twins-for-surmont-and-montney/) — See original source; 2025-05-27; Published 2025-05-27; lines 41-79, current tasks, spool example, limitations, future integration/build-out. Access: Issuer page opened; full article reviewed. ## ConocoPhillips — Eagle Ford horseshoe wells combine two laterals in one drilling event T10-R02-US-01 · Drilling & completions · Recent Eagle Ford projects described in an issuer article published 2026-05-18; exact spud/completion dates for all four wells are not specified. ConocoPhillips describes an Eagle Ford horseshoe-well design that joins two laterals in one drilling event. The issuer estimates $10-$15/BOE lower cost of supply versus traditional designs; test-well tracer evidence indicated production contribution along the wellbore. Broader application remains under evaluation. Status: Completed small-scale application. Company says production was in line with expectations and comparable to standard long laterals; broader applicability was still being evaluated. Economics: Label $10-$15/BOE as an issuer estimate, not realized project cost savings or a disclosed AFE amount; single-project test results do not establish repeatability. Baseline: Issuer comparison is against traditional well design / standard long laterals, but does not provide matched-well characteristics, absolute capex, date range for savings, or a detailed cost-of-supply model. Production is described as comparable and in line with expectations, not as uplifted recovery. AI role: No AI role disclosed; the disclosed intervention is physical well geometry and trajectory execution. This does not establish that no AI tools were used in planning or operations. Limits: Only four wells across two projects are described; source gives no individual well results, full eligible-well denominator, cost breakdown, uncertainty range or operator working-interest basis. Benefits are company-reported estimates. Deployment beyond Eagle Ford was under evaluation at publication; this is not evidence of basin-wide standardization. Next question: For which lease geometries and reservoir conditions does the design beat a matched two-well/standard-long-lateral alternative after incremental trajectory, completion, intervention and production-risk costs, and has the four-well method been repeated at scale? - [Value curve: Horseshoe wells expand development options](https://www.conocophillips.com/spiritnow/story/value-curve-horseshoe-wells-expand-development-options/) — ConocoPhillips (SpiritNow); 2026-05-18; Design and rationale paragraphs; paragraphs describing two projects/four wells, capital savings, $10-$15/BOE cost-of-supply estimate, production/tracer result, and evaluation outside Eagle Ford.. Access: Issuer HTML opened directly. ## ConocoPhillips — Lower 48 productivity per foot improved in 2025 as lateral length and completion-design optimization advanced T20-R13-US-01 · Drilling & completions · Calendar year 2025; disclosed during Q4 2025 earnings call on 2026-02-05. Preserve the year-over-year per-foot results as basin-level portfolio observations: Delaware oil productivity per foot +about 8% while average lateral length +9%; Eagle Ford oil productivity per foot +7%. Management cited Delaware spacing/stacking optimization and Eagle Ford diverter use, but also emphasized acreage quality, basin mix and inventory; the sources do not isolate a causal technology contribution or give dollar/recovery results. Status: Operating portfolio results for 2025 reported on 2026-02-05; ongoing design optimization. Economics: No dollar savings, well cost, incremental recovery volume or technology-specific return was disclosed. The 2026 Lower 48 capital reduction outlook is a forward plan and cannot be assigned to diverters or the productivity figures alone. Baseline: Management compares 2025 to 2024 basin oil productivity per foot; the transcript does not provide cohort construction, lateral-length adjustment methodology, well count, geological mix, or a matched non-diverter comparison. AI role: No AI contribution is identified. Management describes ongoing basin development optimization and the use of diverters in Eagle Ford completion designs; no model or automation is named. Limits: The Delaware result includes a 9% increase in average lateral length and management explicitly cites inventory quality, spacing/stacking and geographic mix as drivers. Eagle Ford productivity is a basin metric; diverter use is mentioned but not quantified or isolated. Do not infer a causal effect or convert productivity-per-foot gains into EUR or value without well-level data. Next question: Can ConocoPhillips provide basin-level matched cohorts that separate lateral-length, rock/location, diverter and spacing effects and connect each to D&C cost and full-cycle recovery? - [Q4 2025 ConocoPhillips Earnings Call Transcript](https://static.conocophillips.com/files/resources/cop-usq_transcript_2026-02-05.pdf) — ConocoPhillips / LSEG transcript hosted on ConocoPhillips domain; 2026-02-05; PDF pp. 7–8; transcript lines 240–262. Management's answer on basin performance, lateral lengths, spacing/stacking and diverter completion designs.. Access: Directly opened issuer-hosted PDF; relevant Q&A passage inspected.. ## ConocoPhillips — PD 868 modular robotic pipe handling in Montney TECH-02 · Drilling & completions · 2025-06-18 Source confirms 105,000 m drilled without manual connections, 400,000 m tubulars handled hands-free, and ~8,400 rig-floor personnel hours eliminated. Status: One named land rig, PD 868, with use across several recent pads under a three-year Precision Drilling collaboration; not fleet-wide. Coordinates are currently manually programmed. Economics: Not disclosed Baseline: Not disclosed AI role: non-AI robotics and programmed automation; image-based spatial recognition is described as a future goal Limits: Issuer case story; metrics are company-reported and the footprint is one named rig / several pads. No independent validation or cost conversion disclosed. Next question: How many tasks/wells/rig-days are represented, what time-at-risk or injury rate changed, and what incremental rig cost offsets ~8,400 personnel hours removed? - [Enhancing safety and efficiency / Montney’s automated drilling rig](https://www.conocophillips.com/spiritnow/story/enhancing-safety-and-efficiency-montneys-automated-drilling-rig/) — ConocoPhillips; 2025-06-18; Fast facts and technology/deployment description. Access: See research audit. - [Source reopening in Round 2](https://www.conocophillips.com/spiritnow/story/enhancing-safety-and-efficiency-montneys-automated-drilling-rig/) — See original source; 2025-06-18; Published 2025-06-18; lines 39-50, PD 868 scope, outcome counts, programming approach and future vision goal. Access: Issuer page opened; fast facts and deployment text reviewed. ## ExxonMobil — Baytown low-carbon hydrogen plant with CCS paused; separate methane-pyrolysis demonstration remains planned R1-LIM-04 · Water, emissions & energy · Pause announced in 2025; source is ExxonMobil Advancing Climate Solutions report, published May 2026. Pre-pause figures sourced from ExxonMobil’s 2024-12-11 2030 plan; current report clarifies status. The main Baytown project did not proceed to investment/deployment because management judged market and policy insufficient to justify continued investment. No project capex, expected return, or sunk cost was disclosed in the source reviewed. ExxonMobil continues to engage customers/policymakers. Separate technology tests/plans should not be treated as substitutes already scaled. Status: Main low-carbon hydrogen development paused in 2025 pending market/policy conditions. Different technology, methane-pyrolysis demonstration is planned, not operating in the source. Baytown hydrogen burners have been installed/tested as a separate operating technology; this does not unpause the low-carbon hydrogen plant. Economics: Not disclosed Baseline: Not disclosed AI role: No AI role. The case is relevant to AI energy demand only if surfaced as a separate adjacent opportunity; it should not be counted as an AI deployment or AI-powered energy asset. Limits: Issuer explanation; no third-party economic analysis or customer offtake status disclosed in the source. The report uses corporate promotional language for remaining technologies. The potential 90% burner emissions reduction is a reported test result with no test duration, baseline conditions, or scale detail visible in the cited passage. This is a market/policy pause, not proof of technical failure of hydrogen or CCS. Next question: What hydrogen offtake price/utilization, CCS capex and policy support would clear Exxon’s investment hurdle, and are burner/pyrolysis economics distinct? - [2026 Advancing Climate Solutions Report](https://corporate.exxonmobil.com/-/media/global/files/advancing-climate-solutions/2026/2026-advancing-climate-solutions-report.pdf) — ExxonMobil; 2026-05; PDF p. 19 / printed page 19: Baytown hydrogen pause and policy/market explanation; p. 19–20: burner test and separate methane-pyrolysis demo plan.. Access: Opened; issuer report. - [ExxonMobil announces plans to 2030 that build on its unique advantages](https://corporate.exxonmobil.com/news/news-releases/2024/1211_exxonmobil-announces-plans-to-2030-that-build-on-its-unique-advantages) — ExxonMobil; 2024-12-11; Baytown project pre-pause scale, 98% CO2 capture ambition and prior expected FID/startup timing.. Access: Opened via indexed issuer material; context source. - [Source reopening in Round 2](https://corporate.exxonmobil.com/-/media/global/files/advancing-climate-solutions/2026/2026-advancing-climate-solutions-report.pdf) — See original source; 2026-05; PDF pp.5 and 18; lines 248–254 and 1168–1207.. Access: issuer PDF opened. - [Source reopening in Round 2](https://corporate.exxonmobil.com/news/news-releases/2024/1211_exxonmobil-announces-plans-to-2030-that-build-on-its-unique-advantages) — See original source; 2024-12-11; Earlier plan for Baytown low-carbon hydrogen/CCS.. Access: issuer source referenced in original record; not needed to establish current pause. ## ExxonMobil — Pneumatic-device replacement at acquired Pioneer Permian assets T10-R04-US-01 · Water, emissions & energy · 2025 deployment, disclosed 2026-05-05. ExxonMobil says it replaced more than 6,000 pneumatic devices in acquired Pioneer Permian assets during 2025; this is part of a broader effort to remove natural-gas-driven pneumatic devices, with integrated Permian completion targeted by 2030. Status: More than 6,000 devices reported replaced during 2025; company goal is elimination of pneumatic devices across integrated Permian operations by 2030. Economics: No methane volume avoided, retrofit cost, maintenance effect, or payback is reported here; do not imply measured emissions reduction for this subset. Baseline: Natural-gas-driven pneumatic controls are the qualitative baseline. The source does not state the Pioneer device inventory at the start of 2025, fraction of eligible devices replaced, pre/post measured methane emissions, or a matched-site comparison. The >10,000 companywide since-2023 figure has a broader scope than the 6,000 Pioneer-only 2025 figure. AI role: Physical equipment/process retrofit: replacing natural-gas-driven pneumatic controls with alternatives such as air compressors, mechanical valves, electricity or nitrogen depending on site conditions. The source does not describe AI as part of the device-replacement action. Limits: Issuer sustainability reporting. Replacement count is a deployment output, not measured methane reduction or a completed integrated-Permian program. Retrofit scope and solution mix vary by site; infrastructure replacement may have ancillary reliability benefits, but no quantified uptime/repair result is given. Do not infer compliance, economics or methane reduction from count alone. Next question: For the 6,000 Pioneer replacements, what were eligible inventory, replacement mix and capex, and what did direct before/after measurement show for methane mass, vented-gas sales, reliability and maintenance by device class? - [Driving reductions in methane emissions](https://corporate.exxonmobil.com/publications/advancing-climate-solutions/driving-reductions-in-methane-emissions) — ExxonMobil, Advancing Climate Solutions 2026; 2026-05-05; Sections 'What we've done / What we're doing' lines 78-87 and 'Replacing pneumatic devices' lines 111-120.. Access: Issuer HTML opened directly. ## ExxonMobil — Methane monitoring and Vantage operating outcomes update: coverage, flaring and response denominators remain incomplete T20-R15-US-01 · Water, emissions & energy · Company-reported results cover 2024–2025; disclosure dated 2026-05-05. Exxon says Vantage/COMET provides continuous methane monitoring and real-time monitoring/production management. Its May 2026 page reports more than 400 continuous-monitoring sites on heritage ExxonMobil-operated Permian assets, and separately says more than 6,000 pneumatic devices were replaced in 2025 at acquired Pioneer Permian assets. Vantage’s 2024–2025 operating highlights include >34 MMcf of flaring eliminated and >18,000 driving hours avoided; source gives no project-level counterfactual or cost attribution. The >60% methane-intensity reduction is company-wide operated-asset performance from 2016–2025, not a Vantage result. Status: Vantage was unveiled in October 2025 and extends the COMET platform launched in 2022. Heritage operated Permian continuous-monitoring coverage exceeds 400 key sites; deployment to acquired Pioneer key sites is targeted by end-2026. More than 6,000 pneumatic-device replacements on acquired Pioneer Permian assets were reported for 2025. Economics: No monitoring-system cost, device-replacement cost, value per recovered gas volume, alert response labor, repair cost, avoided loss or payback disclosed. Flared-gas volume and travel hours are operational outputs and should not be monetized without additional assumptions. Baseline: Keep separate: company-wide methane-intensity baseline is 2016; Vantage flaring/driving outputs cover 2024–2025; monitoring count covers heritage sites; Pioneer replacement count covers acquired Pioneer assets in 2025. AI role: Exxon describes continuous analysis and a centralized monitoring center but does not disclose whether AI/ML generates or prioritizes methane alerts, nor its role in repair decisions. Do not call the measurement/control loop AI without more detail. Limits: The source does not report alert count, source confirmation, false alarms, time-to-acknowledge, time-to-repair, closure evidence, leak volume before/after, repair success, or attribution of the 34 MMcf flaring figure specifically to Vantage alerts. It describes a monitoring layer and associated field operations but leaves the detection-to-repair chain unquantified. The 2026 report notes uncertainty in measured/estimated emissions and methodology. Next question: For continuous-monitoring sites, how many confirmed alerts resulted in a documented repair and emissions reduction, over what time window, versus a manual-LDAR counterfactual, including false-positive and monitoring/field-response costs? - [Driving reductions in methane emissions](https://corporate.exxonmobil.com/publications/advancing-climate-solutions/driving-reductions-in-methane-emissions) — ExxonMobil, Advancing Climate Solutions 2026; 2026-05-05; Lines 82–88 (monitoring/device scope), 124–146 (detection, Vantage and reported outputs), 249–259 (measurement boundary/uncertainty).. Access: Directly opened primary issuer page; relevant passages and footnotes inspected.. ## ExxonMobil — Vantage central operations and monitoring center at reported 22,000-site U.S. upstream scale T10-R05-US-01 · Production & process operations · The platform is described as operating as of 2025-10. Flaring reduction and driving-avoidance figures are explicitly for calendar 2024; uplift period is not specified. Vantage is reported to monitor >22,000 sites; it manages/optimizes >1 million b/d oil and >3 million b/d water, monitors/directs >3 Bcf/d gas, and reports >29,000 b/d oil uplift from field and remote production-optimization activities. Separate 2024 metrics include >20 MMcf flaring eliminated and >6,000 driving hours avoided. Status: Vantage publicly unveiled 2025-10-10; operating 24/7 by announcement. COMET pre-existed, launched in 2022; Vantage is an expanded/renamed operating-center initiative rather than a platform developed wholly in 2025. Economics: The company-reported production-uplift figure could be material, but the page gives no duration, gross/net basis, persistence, decline convention, commodity mix, intervention/capital cost or cash margin. The 2024 flaring/driving figures are distinct operational outputs and should not be monetized without fuel, emissions and response-cost baselines. Baseline: No pre-Vantage/control baseline or counterfactual is specified for the >29,000-b/d uplift. Vantage was unveiled in 2025 but the uplift is attributed to field and remote production optimization activities, not clearly solely to the center. Page says 22,000 sites; a later 2026 report uses 'wells' for a similar-scale coverage figure, so retain the 2025 source's unit rather than silently converting sites to wells. AI role: Vantage integrates real-time monitoring and production management; the 2025 launch article does not disclose whether AI/ML determines alerts or actions, nor the degree of automated control. COMET is named as a monitoring platform; do not infer autonomous closed-loop operation from the article's future-state language. Limits: Issuer disclosure does not specify site-versus-well denominator, eligible population, matched baseline, attribution to AI versus remote/field operations, or whether the stated uplift is sustained; keep these as issuer-reported scale/outcome claims. Next question: How much of the 29,000-b/d uplift is durable incremental production attributable to Vantage-enabled interventions, over what measured period and well population, and what staffing, sensor, integration and field-response costs were required? - [Introducing Vantage: A new lens on upstream operations](https://corporate.exxonmobil.com/who-we-are/our-global-organization/business-divisions/upstream/a-new-lens-on-upstream-operations) — ExxonMobil; 2025-10-10; Vantage overview and 'Big numbers, bigger impact,' lines 34-64.. Access: Issuer HTML opened directly. - [Driving reductions in methane emissions](https://corporate.exxonmobil.com/publications/advancing-climate-solutions/driving-reductions-in-methane-emissions) — ExxonMobil, Advancing Climate Solutions 2026; 2026-05-05; Monitoring and detection section lines 124-146; near-continuous monitoring roll-out boundary lines 127-130.. Access: Issuer HTML opened directly. ## ExxonMobil — Proprietary lightweight proppant TECH-01 · Drilling & completions · 2025-12-09 Issuer reports early results showing about 20% recovery improvement from proprietary lightweight proppant technology alone. Status: Technology is in use in the Permian; issuer reports early results but gives no well count or deployment coverage. Economics: Not disclosed Baseline: No well count, deployment share, comparator/baseline definition, measurement period, cost, or independent validation disclosed. No denominator or dollar bridge should be inferred. AI role: Physical completion-material technology; no AI contribution is disclosed. Limits: Issuer-reported early result; no denominator, baseline definition, time period, independent comparison, or deployment count disclosed. Next question: What is the recovery uplift denominator, control population, production period, cost/ton proppant premium and net PV per well behind the early ~20% claim? - [ExxonMobil raises its 2030 Plan – transformation delivering higher earnings, stronger cash flow, and greater returns](https://corporate.exxonmobil.com/news/news-releases/2025/1209-exxonmobil-raises-2030-plan-transformation) — ExxonMobil; 2025-12-09; Upstream section, lines 64-69. Access: See research audit. - [Source reopening in Round 2](https://corporate.exxonmobil.com/news/news-releases/2025/1209-exxonmobil-raises-2030-plan-transformation) — See original source; 2025-12-09; Upstream paragraph, lines 64–69.. Access: issuer HTML opened. ## ExxonMobil — Open Subsurface Data Universe (OSDU) Data Platform TECH-06 · Data, people & purchasing · 2026-07-10 Evaluation times were said to be dramatically shortened, but no delta, denominator, or independent validation was provided. Status: Issuer describes current platform work and a qualitative reported evaluation-time benefit; broader/full operational capability remains in progress. Economics: Not disclosed Baseline: Not disclosed AI role: OSDU is an open subsurface data foundation. AI is mentioned generically as a potential/associated application, not as the cause of the qualitative evaluation-time claim. Limits: Issuer-reported qualitative outcome, without before/after times, asset counts or independent validation. Do not attribute it to AI. Next question: What evaluation-time delta, assets/projects evaluated, data-engineering cost and resulting decision quality/value are attributable to OSDU? - [Unleashing the power of open data](https://corporate.exxonmobil.com/locations/asia-pacific-region/ap-newsroom/2026/unleashing-the-power-of-open-data) — ExxonMobil; 2026-07-10; Guyana link and ‘Real-world wins’. Access: See research audit. - [Source reopening in Round 2](https://corporate.exxonmobil.com/locations/asia-pacific-region/ap-newsroom/2026/unleashing-the-power-of-open-data) — See original source; 2026-07-10; OSDU platform and Guyana examples; qualitative evaluation-time statement; future fully-operational language. Access: ExxonMobil issuer page opened; text accessible. ## Occidental Petroleum (Oxy) — Horn Mountain gas-lift and caisson ESP projects implemented under Gulf of America 2.0 T10-R03-US-01 · Production & process operations · Implemented during 2025; disclosed in Oxy FY2025 Form 10-K filed in 2026. Oxy says the Horn Mountain gas-lift and caisson ESP projects delivered some of the highest-margin production in its portfolio. In the same discussion it reports Gulf of America production and major-equipment uptime, but does not attribute either metric to the artificial-lift projects. Status: Implemented at Horn Mountain during 2025; water-injection projects elsewhere in the GOA 2.0 program were planned for 2026/2027 and are separate. Economics: No project capital or standalone uplift quantified. Phrase 'some of the highest-margin production' as management's qualitative portfolio comparison, not a measured lift-project margin or return. Baseline: No pre-project production, lift method, cost or margin baseline disclosed. Portfolio production (132 Mboe/d) and equipment uptime (>99%) are context only and should not be treated as project-level uplift. AI role: No AI role disclosed. The named interventions are down-hole gas-lift and caisson electric submersible pumps; no-AI disclosure is not proof that no digital tools were used. Limits: The source names multiple projects but does not count individual installations/wells, define 'highest margin' comparator, or isolate project contribution. Reported 96 gross wells are across eight platforms and 14 active fields, not an eligible lift-project denominator. Water injection dates are plans at filing, not completed outcomes. Next question: What were the project-level incremental oil volumes, margin, lift/intervention capex, operating cost and pre-project decline profile for each Horn Mountain installation, and have Marlin/Horn Mountain injection milestones actually started? - [Oxy 2025 Form 10-K / Annual Report](https://www.oxy.com/siteassets/documents/investors/2025-annual-report.pdf) — Occidental Petroleum; 2026-02-19; Management's Discussion and Analysis, Gulf of America, PDF pp. 33-34 (printed pp. 26-27), lines 1473-1497.. Access: Issuer-hosted PDF opened directly. ## Occidental Petroleum (Oxy) — Oman Block 53 artificial-lift improvement associated with lower workover-rig cost per barrel T10-R03-US-02 · Maintenance & reliability · The three years preceding 2025-05-08; exact start/end dates and annual trajectory not provided. Oxy management reported that Block 53 drilling cost per foot and workover-rig cost per barrel each fell 50% over the prior three years, while saying artificial-lift performance had improved significantly. This is block-level operating context; the source does not isolate the lift technology's contribution to the workover-rig cost change. Status: Operating performance trend over a three-year period; disclosed in May 2025. Economics: No AI/hardware-specific cost attribution or absolute $/bbl disclosed. Baseline: Three years before the May 8, 2025 call; exact start/end dates and absolute costs undisclosed. AI role: No AI or specific lift technology disclosed; issuer management describes performance improvement, not its technical mechanism. Limits: Management-call attribution only; no specific lift equipment/process change or operational dataset is cited. Cost per barrel may change with barrel denominator as well as rig cost. Do not conflate with adjacent drilling-cost-per-foot claim or claim improved recovery without production data. Next question: Was the 50% change driven by longer pump run-life, fewer interventions, lower rig day rates, more barrels per workover, or another factor—and what were annual absolute costs and production-adjusted denominators? - [Q1 2025 Occidental Petroleum Corp Earnings Call Transcript](https://www.oxy.com/siteassets/documents/investors/quarterly-earnings/oxy1q25transcript.pdf) — Occidental Petroleum / LSEG Refinitiv transcript hosted by Oxy; 2025-05-08; PDF p. 9, remarks by Kenneth Dillon, SVP and President, International Oil and Gas Operations, lines 317-330.. Access: Issuer-hosted transcript PDF opened directly. ## Occidental (Oxy) — DJ Basin Aggregate Recycling Facility TECH-04 · Water, emissions & energy · 2025 Official report indexed text reports 2024 facility saved over $18 million compared with traditional commercial disposal; recycled approximately 142 megaliters (142 million liters) of water; avoided 400,000 truck-transport miles; recycled over 30,000 m3 of material. Status: Operating DJ Basin facility and 2024 result, not a future target. Economics: Not disclosed Baseline: $18m+ is issuer-reported avoided disposal cost against traditional commercial disposal; it is not revenue. Cost boundary, denominator and independent assurance were not available in directly opened report text. AI role: Physical process/materials-recovery and water-recycling operation; no AI role disclosed. Limits: Issuer-reported facility result and comparison; the cited report summary does not give full cost boundary, detailed denominator, or independent assurance for the savings estimate. Next question: How does $18m+ avoided disposal cost reconcile to facility throughput, alternative disposal tariffs, capex/opex, water treatment and quality, and avoided-mile denominator? - [2025 Sustainability Report](https://calciumchloride-prod.oxy.com/siteassets/documents/publications/2025-sustainability-report-web.pdf) — Occidental Petroleum; 2025; Aggregate Recycling Facility section. Access: See research audit. - [Source reopening in Round 2](https://calciumchloride-prod.oxy.com/siteassets/documents/publications/2025-sustainability-report-web.pdf) — See original source; 2025; 2025 Sustainability Report, p.63, Aggregate Recycling Facility (index result).. Access: official issuer-hosted PDF; direct browser open returned internal error/oversize, indexed extract available; content not fully directly inspected. ## Saudi Aramco — Upstream water reuse and zero-liquid-discharge technology portfolio T10-R01-GLOBAL-02 · Water, emissions & energy · 2025; source is Aramco Sustainability Report 2025, issued in 2026. Around 8.3 million gallons of wastewater were recycled at Uthmaniyah and Shedgum gas plants for stimulation/drilling in 2025. The separately described ZLD technologies/trial are not stated as the source of that volume; 99% purity is a process claim, not independent water-quality verification. Status: Issuer describes a commercialized portfolio of water-conservation technologies and reports 2025 operating reuse; it also reports 2025 advancement of ZLD treatment/salt-recovery methods and a field trial using residual ZLD waste in drilling operations. Economics: No named-technology capex/opex, water-treatment tariff, groundwater value, disposal savings, reuse yield or field-level unit cost disclosed. Portfolio's 'up to 250 million gallons per year per plant' groundwater-conservation potential for one contaminated-water fracturing-fluid technology is a stated potential, not the measured result at Uthmaniyah/Shedgum. Baseline: No plant-specific before/after volume, freshwater displacement denominator, water quality specification for the 8.3 million gallons, or alternative-disposal comparator reported. AI role: The water treatment/reuse mechanisms are physical/process technologies. The report separately discusses digital discharge monitoring, but does not attribute AI to the water-reuse volume or ZLD trial; absence of disclosure is not proof no AI is used. Limits: This is a bounded portfolio update rather than one isolated project: separate the 8.3 million gallons of wastewater reuse, ZLD purity claim, ZLD residual-waste trial and potential per-plant groundwater savings. Reported activity is useful upstream water/process context; do not convert it to financial ROI or attribute outcomes to ZLD without source proof. Next question: Which treatment process produced the reported site volume, what is the input/output water-quality and volume balance, and how do full treatment/disposal costs compare with freshwater and disposal alternatives? - [Saudi Aramco Sustainability Report 2025](https://www.aramco.com/-/media/publications/corporate-reports/sustainability-reports/report-2025/english/2025-saudi-aramco-sustainability-report-full-en.pdf) — Saudi Aramco; 2026 (2025 reporting year); PDF printed pp.36-37 / opened text lines 5077-5136: 8.3 million gallons at Uthmaniyah/Shedgum; portfolio process claims; 99% purity and residual-ZLD field trial.. Access: Official PDF opened; relevant pages and text directly reviewed.. ## Saudi Aramco — Jubail CCS hub phase-one operating target moved from 2027 to 2028 T10-R08-GLOBAL-02 · Water, emissions & energy · Schedule amended 2025-09-02; latest scope/schedule confirmed in 2025 issuer update. Aramco’s July 2025 article (amended 2 Sep) moves Jubail CCS phase-one planned operational date from 2027 to 2028; up-to-9 Mt CO₂/year remains capacity ambition. The article does not give the schedule-change cause or evidence of a technical failure. Use 2024 report only as older stage context. Status: Under development; issuer amended schedule to 2028, one year later than previous 2027 plan; annual capacity target up to 9 Mt CO2 (6 Mt Aramco facilities, 3 Mt external industrial facilities per 2024 sustainability report). Economics: No current project cost, delay cost or expected unit storage tariff disclosed in the reviewed article/report. Baseline: Schedule baseline revised from planned 2027 to 2028; capacity is a planned maximum, not installed/operating volume. AI role: No AI role disclosed; capture/transport/storage hub is industrial process and infrastructure. Limits: Reason for schedule revision not disclosed in the opened article. Do not infer construction problems or technology failure from a date change. CCS hub is industrial infrastructure and not a named upstream production deployment. Next question: Does the 2028 timeline change upstream-to-hub sequencing or vendor/transport utilization assumptions, and what is the revised cost-to-complete and offtake commitment? - [Closing the loop: how Aramco is progressing towards a circular economy](https://www.aramco.com/en/news-media/elements-magazine/2025/closing-the-loop-how-aramco-is-progressing-towards-a-circular-economy) — Aramco; 2025-07-15; amended 2025-09-02; Lines 136-140, 163. Access: Direct issuer article opened; date and amendment found. - [Saudi Aramco Sustainability Report 2024](https://www.aramco.com/-/media/publications/corporate-reports/sustainability-reports/report-2024/english/2024-saudi-aramco-sustainability-report-full-en.pdf/) — Aramco; 2025 publication; reports 2024; PDF pp. 39, 2024 progress toward planned initiatives. Access: Indexed primary report excerpt. ## Saudi Aramco — 2025 portfolio cost per boe is a scale benchmark, not a technology return T10-R09-GLOBAL-02 · Data, people & purchasing · Year ended 2025; 2024 comparator. Aramco reports portfolio average upstream lifting cost $3.51/boe (2025) vs $3.53 (2024), and upstream CAPEX $8.0/boe vs $8.3. These are portfolio benchmarks influenced by geology, locations, infrastructure, depletion, fiscal discipline and technology; they do not isolate a technology contribution or compare like-for-like asset economics. Status: Reported actual FY2025 outcome in annual report. Economics: Actual portfolio averages, not marginal cost, project break-even, technology spend, or technology ROI. Reported CAPEX per boe is not same as lifting cost and should be kept separate. Baseline: 2024 comparator: $3.53/boe lifting and $8.3/boe upstream capex; 2025 actual: $3.51/boe and $8.0/boe. AI role: Report references scaled technology application as one of several drivers of low costs, alongside geology, favorable operating locations, and infrastructure/logistics. No isolated technology effect disclosed. Limits: Aramco's reservoir/location/geology and giant integrated scale make a raw $/boe comparison with smaller offshore or unconventional projects misleading. No technology counterfactual, asset-level eligible population, or technology spending allocation is reported. Next question: Which unit-cost dimensions can be normalized across operators (lifting versus total cost versus CAPEX per boe), and which structural advantages must be controlled before attributing differences to technology? - [Annual Report 2025, Upstream operations](https://www.aramco.com/-/media/publications/corporate-reports/reports-and-presentations/2025/fy/sections/ara-2025-results-english.pdf) — Saudi Aramco; 2026 publication; FY ended 2025; PDF pp. 48-49 (printed); lines 1479-1498, 1572-1607. Access: Direct annual-report PDF opened. ## Saudi Aramco / Yokogawa — Fadhili moves from an AI pilot to autonomous gas-treatment control T20-R11-GLOBAL-01 · Production & process operations · Commissioning reported 29 October 2025; benefit measurement period not disclosed. Yokogawa reports initial Fadhili results of 10%-15% lower amine and steam use, about 5% lower power use, improved process stability, and significantly less operator manual intervention despite ambient-condition changes. The same release says detailed evaluation was ongoing. It describes three deployment phases, simulator training/validation, then integration with CENTUM VP before autonomous control of the AGR core process. Status: Yokogawa reported commissioning/implementation of multiple autonomous-control agents on 2025-10-29. The implementation was still undergoing detailed evaluation when the supplier disclosed initial results. Aramco's current undated page later says permanently adopted, but that page could not be directly fetched in this pass; do not claim a verified adoption date from it. Economics: No absolute chemical/steam/power quantities, prices, annualized savings, project cost, baseline duration, or return/payback are disclosed. Percentage reductions are supplier-reported initial results and not independently validated in the release. Do not multiply percentages by plant-wide utility/cost totals without system boundary and measurement data. Baseline: Implied pre-deployment operating usage for the Fadhili AGR system; the comparison period and normalization for throughput, feed composition, ambient conditions, or gas quality are not specified. AI role: Multiple coordinated agents use Yokogawa's Factorial Kernel Dynamic Policy Programming (FKDPP), a reinforcement-learning algorithm, to directly and autonomously control acid gas removal (AGR). A simulator trained agents; they were evaluated for reliability/validity and integrated into Yokogawa CENTUM VP to retain existing process-control safety functions. Limits: This is an integrated gas-processing unit rather than upstream well/field production, but it is a consequential operator process-control analogue. Direct process control and safety-system integration are meaningful maturity details; commissioning is not proof of sustained performance. No measurement/error design, guardrails, intervention rate, independent validation, or duration supplied. Exact page now confirms AI-141 is same initiative; do not count as a new deployment. Next question: What were the baseline and post-commissioning test windows, absolute normalized amine/steam/power use per gas throughput, operator interventions and constraint violations, and what independent evaluation supports sustained autonomous control? - [Aramco and Yokogawa Achieve a Major Milestone with Commissioning of Multiple Autonomous Control AI Agents at Major Gas Facility](https://www.yokogawa.com/news/press-releases/2025/2025-10-29/) — Yokogawa Electric Corporation; 2025-10-29; Release dated 29 October 2025; paragraphs on phased AGR control implementation and preliminary reductions.. Access: Official supplier release found through its indexed full text and 2025 press-release index; direct root open returned bot challenge. Earlier agent read a mirrored URL; official URL replaces that citation.. - [AI and Big Data - Oil & Gas Industry](https://www.aramco.com/en/what-we-do/energy-innovation/digitalization/ai-and-big-data) — Saudi Aramco; Undated current page; operating period stated October 2024-April 2025; Fadhili Gas Plant section; indexed primary search excerpt.. Access: Search-indexed primary excerpt; direct page open timed out. ## ADNOC — Taurob heavy-duty inspector robot at Taweelah Gas Compression Plant R1-GL-05 · Maintenance & reliability · ADNOC release dated 2026-05-21; operator-robot completion is a target by end-2026, not completed evidence. Routine inspection deployment is described, with no quantified robot-attributed outcome or cost. Keep the separately planned operator/manipulator robot and the 30% HSE Cockpit.ai incident reduction separate; neither is an outcome of the deployed Taurob robot. Status: Inspector robot deployed and conducting routine on-site autonomous inspections in May 2026. A separate heavy-duty operator robot that lifts/grips equipment and turns valves remained planned for end-2026. Economics: Robot capex, integration, maintenance, and cost per inspection not disclosed. Baseline: Hazardous human inspection is the implicit alternative; no before/after baseline disclosed. AI role: Taurob inspection system uses LiDAR/thermal imaging; the source does not attribute its function or outcomes to AI. Limits: Single-plant deployment and capability description do not establish adoption or performance. Keep inspector robot separate from future manipulation robot and from separate AI HSE metrics. Gas compression plant is not an upstream producing asset. Next question: How many inspection hours/routes, callouts or exposure-hours does the deployed robot remove per year, net of integration and maintenance cost? - [ADNOC Deploys Industry-First Heavy-Duty Robot to Strengthen Safety, Reliability and Performance](https://adnoc.ae/en/news-and-media/press-releases/2026/adnoc-deploys-industry-first-heavy-duty-robot-to-strengthen-safety-reliability-and-performance) — ADNOC; 2026-05-21; Deployment, sensors, future operator robot, and separate HSE Cockpit.ai claim, lines 527-543. Access: Issuer page opened; full text accessible. - [Source reopening in Round 2](https://adnoc.ae/en/news-and-media/press-releases/2026/adnoc-deploys-industry-first-heavy-duty-robot-to-strengthen-safety-reliability-and-performance) — See original source; 2026-05-21; Taurob robot at Taweelah Gas Compression Plant; separate heavy-duty operator robot planned for end-2026; HSE Cockpit statistic. Access: ADNOC issuer page opened; text accessible. ## ADNOC — Produced-water reinjection reported at Bu Hasa and Asab T10-R03-GLOBAL-02 · Water, emissions & energy · 2025 report; operating period for 100% achievement not separately specified. ADNOC states 100% produced-water reinjection at Bu Hasa and Asab. The report does not define the water-volume denominator, reporting period or cost, nor isolate digital tools as the cause. Status: Operating achievement claimed by issuer in 2025 sustainability report; broader mature-field PWRI roadmap remains in progress. Economics: No named-site volumes, energy, water-treatment cost, avoided disposal cost, or production uplift disclosed in the reviewed chapter. Baseline: No named-site baseline or definition of the 100% denominator is provided in the accessible report text. Company-wide table reports 446 million m3 water reinjected for reservoir management in 2025 but does not link that volume specifically to PWRI at the two sites. AI role: Report says performance indicators are supported by digital surveillance and optimization tools, but does not attribute the 100% PWRI achievement to AI or name a model. Digital support and reservoir process should remain separate. Limits: Do not interpret 100% as all produced water across ADNOC or as 100% water reuse company-wide; denominator, measurement boundary, timing, and site volumes need clarification. Site achievement should not be conflated with the separate Sustainable Water Supply Project (treated seawater, still under development). Next question: Does full produced-water reinjection at two mature sites reduce make-up water and disposal costs without a countervailing injection-energy, injectivity, or water-handling penalty? - [Sustainability Report 2025: Delivering Responsibly, Upstream water management](https://reports.adnoc.ae/e/sustainability-report-2025/mjoOSdaQ33) — ADNOC; 2026 publication; reports 2025; Chapter page 33 / printed pp. 66-67; lines 2, 4-8. Access: Direct issuer report chapter opened. ## ADNOC — Sustainable Water Supply Project for Bab and Bu Hasa injection T10-R04-GLOBAL-01 · Water, emissions & energy · Planned commissioning 2026; forecast capacity and savings are prospective. ADNOC describes a designed capacity >115 million imperial gal/day (about 3.2 million bbl/day) and a project for seawater treatment/supply to Bab and Bu Hasa. Capacity and up-to-30% energy-use benefit are design/company claims; do not report as commissioned throughput or measured savings. Current source status/commissioning date should be sourced before presenting as current. Status: Under development/current issuer status says due to be commissioned in 2026; not verified commissioned as of this research pass. Economics: No project CAPEX, water tariff, OPEX, pump energy baseline, savings measurement protocol, or realized energy reduction disclosed on the project page. Baseline: No baseline or denominator behind up-to-30% reduction disclosed; planned rather than measured. AI role: No AI contribution disclosed; conventional water treatment, pumping and pipeline infrastructure. Limits: Issuer project page indexed, not directly opened (request timed out). Current reported schedule is 2026; treat as due/expected, not commissioning evidence. It is distinct from produced-water reinjection at Bu Hasa and Asab reported in R03. Next question: Before sanction/replication, what metered energy-per-barrel and aquifer-water displacement will the network deliver after start-up, and what tariff/availability risk sits in the TAQA partnership? - [Sustainable Water Supply Project](https://www.adnoc.ae/en/our-projects/sustainable-water-supply) — ADNOC; Undated current issuer project page; accessed 2026-10-07; Project overview and fast facts. Access: Search-indexed primary page; direct open timed out. ## ADNOC / SLB — SLB DrillOps-enabled Real-Time Operations Center deployed across >120 rigs T10-R07-GLOBAL-01 · Drilling & completions · Deployment stated as of 2026-08-04; performance window and ramp dates not given. ADNOC reports >120 rigs and claims 30–40% lower engineering effort, 2–3x rig coverage per engineer, 4–12 hours shorter incident response, and potential avoidance of 1–2 downtime days. Measurement windows, baselines, frequency and independent validation are not disclosed; no cost/payback provided. Status: ADNOC announced deployed fleet-wide platform on 2026-08-04; production deployment according to issuer release. Economics: No contract value, rig-level deployment cost, gross/net savings, uptime history, or savings attribution shared; performance figures are issuer claims without methodology or evaluation period. Baseline: Unspecified prior staffing/rig coverage and performance windows; reference to prior reporting cycles taking several days/full day of analysis is not quantified by use case. AI role: Explicit: AI-powered performance insights combined with real-time operational data, analytics, dashboards and DrillOps. Reported impact should be regarded as issuer/vendor claim pending defined baselines and audit. Limits: Fleet coverage is meaningful scale, but the listed benefit ranges need denominators and measurement definition. The 1–2 days is potential avoided downtime, not observed downtime avoided per rig. No external validation disclosed. Next question: What per-rig staffing, reporting, incident and downtime logs substantiate the savings claims, and what portion is caused by software/AI versus process redesign and centralization? - [ADNOC and SLB Deploy AI Platform Across Over 120 Drilling Rigs to Strengthen Upstream Performance](https://www.adnoc.ae/en/news-and-media/press-releases/2026/adnoc-and-slb-deploy-ai-platform-across-over-120-drilling-rigs-to-strengthen-upstream-performance) — ADNOC; 2026-08-04; Lines 527-543; repeated release date 548-560. Access: Direct issuer release opened. ## ADNOC Onshore / Jereh Oil & Gas Engineering Corporation — ADNOC’s remote-well contract shows the infrastructure around AI T20-R17-LIMITS-01 · Data, people & purchasing · Award announced 2024-11-05; completion target 2027. ADNOC announced an EPC award worth up to $920m (AED 3.38bn) to Jereh for remote sensing and well-operating equipment at Bab, Bu Hasa and Southeast. The >2,000-well figure is the broader program target, not an explicitly itemized contract population. The announcement says data from equipment and a private 5G network under construction would enable autonomous RoboWell control. This is a material physical-infrastructure procurement signal; the award cap is not AI-only spend or realized value. Status: Up-to-$920m EPC award announced 2024-11-05 for installation at three named fields. Overall digitalization program completion was targeted for 2027. Release identifies prior sensor/control-hardware deployments at Mender, Qusahwira and NEB; it says the private 5G network was being built and that it would enable AI tools such as RoboWell. It does not say RoboWell was already operating across the awarded fields. Economics: Up to $920m announced EPC contract value (not necessarily fully spent); no separation of software, network, hardware, EPC or support cost; no AI-specific return. Baseline: Release identifies earlier digitalization deployment fields but gives no before/after well performance or comparable cost baseline. AI role: ADNOC says real-time field data will enable tools including AIQ RoboWell to control wells autonomously; award's stated scope is remote sensing and well-operating equipment, not an itemized AI contract. Limits: Issuer press release; total contract value includes remote sensing and well-operating equipment and EPC scope. It does not separate AI software, implementation, hardware, telecom, maintenance or other costs. The 2027 rollout is a target; fieldwide live scope and measured outcomes are not reported. Next question: Which parts of the award's total contract value are addressable to software, integration, sensors, networks and service, and what evidence will qualify suppliers for the next asset tranche? - [ADNOC Awards $920 Million Contract to Extend AI-Powered Well Digitalization Program to Cover over 2,000 Wells](https://www.adnoc.ae/en/news-and-media/press-releases/2024/adnoc-awards-920-million-contract-to-extend-aipowered) — ADNOC; 2024-11-05; Lines 527–541. Access: directly opened issuer announcement. ## Aker BP — HIPlog measures flow inside a multilateral well at Alvheim T10-R02-GLOBAL-01 · Subsurface & recovery · Aker BP's official news index dates the page 2026-01-01; the article says the first campaign was in November but does not state the year. Retain as 'November, year not stated in article' rather than infer a date. One three-branch Frosk Attic well, nine stations, initial campaign: signals and branch contributions measurable; two branches had good quality and the third had greater uncertainty associated with geometry/low rates. Repeated monitoring is planned, not yet a multi-well reliability record. Status: First offshore pilot installed and measurement campaign completed; Aker BP reports signals from all branches, good measurement quality in two, and greater uncertainty in one. Repeated measurements planned over several years; it is not yet a proven production-uplift or scaled deployment. Economics: No technology cost, intervention avoided, deferred production avoided, incremental recovery, or cash impact disclosed. The claimed mechanism may reduce information gaps without production interruption, but no economic result is yet measured. Baseline: No conventional production-log comparator or matched before/after branch allocation; the source positions prior measurement as difficult/impossible, without quantifying measurement performance or intervention costs. AI role: Physical measurement mechanism uses permanent small heat sources to send controlled thermal signals into production flow, read by existing temperature sensors; data calculate branch/zone production contributions. No AI contribution is disclosed; the data could support later analytics, but no such connection is reported. Limits: Single pilot well; early in well life; one of three branches has greater uncertainty; subsequent repeated measurements are needed for trend and allocation confidence. Aker BP and service provider statements are issuer/vendor claims; no independent validation located. Next question: Can repeated HIPlog measurements resolve branch-level allocation at low rates and awkward geometry reliably enough to change intervention or production-management decisions, and at what installed/recurring cost compared with conventional production logging? - [New technology provides better insight into wells at Alvheim](https://akerbp.com/en/new-technology-provides-better-insight-into-wells-at-alvheim/) — Aker BP; 2026-01-01 (official news-index date; article body omits date); Lines 85-105 method and non-interference; lines 125-139 named well, station count, first measurement and uncertainty; lines 153-157 next steps.. Access: Official issuer page opened; full relevant text reviewed. Publication date checked in Aker BP official news index, where the item is listed as 01.01.2026.. - [Aker BP news archive, page 2](https://akerbp.com/en/news/page/2/?filter=news) — Aker BP; Current archive; retrieved 2026-10-07; Entry 'New technology provides better insight into wells at Alvheim' dated 01.01.2026.. Access: Official archive search-index text inspected.. ## Aker BP / production-license partners — Aker BP extends reservoir contact with multilateral exploration wells T10-R02-GLOBAL-02 · Subsurface & recovery · Drilling began early May 2025 and ran about three months; campaign completion announced 2025-08-21/22. Issuer/regulator report a 45-km multi-target campaign over about three months, 40 km in reservoir, discoveries in four wellbores and one dry target (Alfa Sør); reported recoverable resource estimate is 96–134 mmboe. This is a campaign outcome, not a normalized drilling-rate/cost comparison. Status: Campaign completed in August 2025; discovery volumes remain to be assessed as part of Yggdrasil development. This is an exploration result, not a producing-well performance test. Economics: No campaign cost, well-cost comparator, days-saved benchmark, cost per target or economic value attributable to multilateral architecture disclosed. 96–134 mmboe is an issuer/regulator-reported discovery estimate, not reserves, sanction or technology-caused value. Baseline: No matched conventional-well campaign, reservoir exposure baseline or cost/time per prospect disclosed. The evidence demonstrates technical reach and exploration access, not a comparable unit-cost advantage. AI role: Aker BP describes a horizontal multilateral drilling method and high-quality subsurface data. The source mentions other AI tools in the broader Yggdrasil program but does not attribute the well architecture, drilling result or discovery to AI. Limits: The campaign comprised one main well plus sidetracks A–G, not one literal borehole. Results combine geometry, geology, rig, planning and alliance execution. One target was dry and one sidetrack did not reach planned TD per regulator details. Do not convert the estimated discovery into realized production or claim the drilling method alone caused the resource outcome. Next question: Can long horizontal multilateral exploration reduce cost per target or accelerate appraisal/tie-back choice versus separate wells, after accounting for sidetrack complexity, failures and data value? What execution cost/time data are needed before replicating the architecture? - [Significant oil discovery in the Yggdrasil area](https://akerbp.com/en/borsmelding/significant-oil-discovery-in-the-yggdrasil-area-2/) — Aker BP; 2025-08-21; Lines 84-106; campaign targets, multilateral approach, drilling length/period, longest branch, resource estimate, partner/asset context.. Access: Official issuer page opened; full relevant text reviewed.. - [Oil discovery in the Yggdrasil area (25/1-14, 25/1-14 A, B, C, D, E, F and G)](https://www.sodir.no/en/whats-new/news/exploration-drilling-results/2025/oil-discovery-in-the-yggdrasil-area-251-14-251-14-a-b-c-d-e-f-and-g/) — Norwegian Offshore Directorate; 2025-08-22; Lines 302-323; campaign/wellbores, discovery count and volume, five-prospect description; lines 324-348 wellbore-specific results, dry target and planned-TD limit.. Access: Official regulator page opened; relevant details directly reviewed.. ## Aker BP — Land-controlled ROV stabilizes Idun North well in one workday T10-R05-GLOBAL-01 · Maintenance & reliability · Published 2026-06-15; actual operation date not specified. One Idun North well stabilization by gravel placement was controlled from Haugesund using ROV/vessel coordination. “12 hours” is the work duration compared with a traditional full offshore rotation of up to 14 days—not a matched 14-day job-duration baseline. Aker BP says this was first-of-kind and expresses qualitative lower time/cost benefits. Status: Single successful operation; Aker BP calls it the first of its kind and says it aims to repeat the operating model more often. Economics: Company claims lower time and cost but provides no vessel day rate, intervention cost, personnel count or net realized savings. Baseline: Comparison is issuer's typical full offshore rotation (up to 14 days), not a matched prior intervention; task completion reported around 12 hours. AI role: No AI contribution disclosed. The evidence concerns remote operating model, communications and a piloted ROV; do not tag as AI adoption. Limits: This is a single operation and no repeatability, intervention success rate, well integrity outcome, safety comparison, or measured cost baseline is disclosed. The vessel still had an offshore crew; key personnel and ROV operator were onshore. Next question: Can this remote ROV operating model be repeated on routine well-intervention scopes and reduce vessel-days while preserving intervention quality and control? - [Styrte undervannsoperasjon fra land – kuttet 14 dager til 12 timer](https://akerbp.com/styrte-undervannsoperasjon-fra-land-kuttet-14-dager-til-12-timer/) — Aker BP; 2026-06-15; Lines 86-114; Aker BP archive dates item 15.06.2026. Access: Direct issuer page opened (Norwegian); archive listing directly/search opened. ## Aker BP — Aker BP tests a resident inspection drone at Edvard Grieg T10-R05-GLOBAL-02 · Maintenance & reliability · Published 2025-06-02. One remotely controlled offshore drone inspection at Edvard Grieg is reported as a first/demonstration and a learning milestone. No scaled operating coverage, avoided mobilizations, reliability or cost result disclosed; “AI” appears as broad milestone language without a task-specific model/function. Status: First full resident-drone operation on the Norwegian shelf; evidence gathered to develop use cases toward full operational deployment. Economics: No actual inspection time, offshore hours removed, failure rate, CAPEX, operating cost, or realized saving disclosed. Supplier estimate of up to 70% inspection-cost reduction appears in SPE trade press, not adopted here as realized operator value. Baseline: No quantified manual inspection or prior mission baseline in issuer release. AI role: Issuer describes the solution as a combined robotics, AI and digital milestone but identifies autonomous navigation and sensor collection, not a specific AI model or AI-dependent function. Keep AI role as mentioned but unspecified. Limits: A successful demonstration is not continuous fleet deployment. No inspection task list, coverage/completeness, image-to-work-order accuracy, weather availability, aviation safety event rate, or specific AI inference is reported. Next question: What recurring inspection task and validated defect-detection coverage are needed to move from a successful flight to an operating service with auditable economics? - [Aker BP and Aker Solutions complete first remote-controlled offshore drone flight](https://akerbp.com/en/aker-bp-and-aker-solutions-complete-first-remote-controlled-offshore-drone-flight/) — Aker BP; 2025-06-02; Lines 81-93; archive listing dated 02.06.2025. Access: Direct issuer page and archive/search listing opened. ## Aker BP — Taurob inspection robot pilot on pre-production Fenris unmanned platform T10-R05-GLOBAL-03 · Maintenance & reliability · Pilot initiated by May 27, 2026; production target Q3 2027. Fenris robot is in pre-production pilot; platform production is expected Q3 2027. Aker BP says scale-up depends on pilot results and explicitly plans evaluation of reliability/availability, usability, HSE and cost-benefit. Do not call it an in-service operating deployment or report savings. Status: Pre-production pilot begun after platform installation; continuation to operating phase conditional on success. Economics: No mobilization avoided, costs, availability result or cost-benefit yet reported; issuer says access constraints make visits costly but gives no baseline. Baseline: No quantified traditional inspection time/cost or pilot metrics. Pre-production low-risk learning environment. AI role: AI not identified in the pilot description; robot collects observations, inspection data, and supports remote operations. Lack of AI mention is not evidence that none is used. Limits: Pilot only, not production performance. Future transfer to Munin/Yggdrasil or brownfields is an ambition, not evidence of replication. Pilot scope/ambition remains adjustable. Next question: Will Fenris robot availability and workflow integration reduce vessel-supported platform visits without creating missed-anomaly or maintenance-response risk? - [Robot inspections at Fenris: building tomorrow’s operations step by step](https://akerbp.com/en/robot-inspections-at-fenris-building-tomorrows-operations-step-by-step/) — Aker BP; 2026-05-27; Lines 85-119; Aker BP archive listing dated 27.05.2026. Access: Direct issuer page opened; archive listing search opened. ## Aker BP — Yggdrasil is preparing for commissioning; production remains planned for 2027 T20-R18-GLOBAL-01 · Data, people & purchasing · Aker BP Q4 2025 results-call transcript dated 2026-02-11; later control-center tour page is within the 2026 source window but does not print a publication date in its opened text (indexed as approximately May 2026). Planned operating start remains 2027. In the results-call transcript, Aker BP described the field's continued construction and commissioning path, including subsea spool and pipeline work, power-from-shore cable installation and preparing facilities for start-up. It says the area will start production in 2027 and be remotely operated from the Stavanger integrated operations center. The later issuer page confirms the center had been toured and repeats 2027 as the operating date. Aker BP describes thousands of sensors, Cognite Data Fusion and supplier/AI-agent data sharing; condition-based maintenance and planning agents are planned from start. This confirms stage progression toward commissioning, not realized remote-operation or AI performance. Status: Aker BP's Feb 11 2026 call described ongoing subsea hookup, cable/pipeline installation, topside outfitting and commissioning, with start-up in 2027. A subsequent issuer report described a tour of the Stavanger integrated operations center and again said Yggdrasil would be remotely operated from there in 2027. No first production or live remote operations was reported in these sources. Economics: No Yggdrasil-specific operating-center, platform integration, sensor/data-platform, software, cybersecurity or AI cost is disclosed. Aker BP discussed investment-estimate increases for major projects, but the reviewed sources do not attribute those increases to the operating model or AI stack; no technology cost bridge or benefit is supported. Baseline: Prior baseline states the same program was under construction/commissioning preparation with 2027 production target. The new call gives detailed construction work and confirms the target, while the later tour establishes that the center is available for demonstration. No actual production, crew reduction, vessel-days, maintenance change or control-room operating window is given. AI role: Aker BP says AI agents will be used from first production for condition-based maintenance and planning and scaled as solutions mature. As of the sources reviewed, that is a stated start-up plan; no production use or AI performance result is disclosed. Thousands of installed/expected facility sensors feed Cognite Data Fusion, with visualization and control-room workflow as enabling infrastructure, distinct from AI itself. Limits: Operator statements and planning disclosures; production is not yet established in these sources. Do not call center tour proof of offshore operating capability or count planned AI use as deployed. The 2027 schedule is a target and may move. Project-wide capex movement cannot be used as the cost of digital/AI components. Next question: When first production begins, what share of operations is actually executed remotely, with what offshore crew/vessel and reliability baseline, and which condition-based maintenance or planning agents are active with measured decision quality and economics? - [Aker BP Q4 2025 results call transcript](https://akerbp.com/wp-content/uploads/2026/02/aker-bp-2025-q4-transcript.pdf) — Aker BP; 2026-02-11; PDF pages 7–8: 2027 start, commissioning activity, remote-operation center and day-one AI plans; pages 12–13 for project cost commentary context.. Access: Issuer PDF opened directly; relevant pages reviewed. - [Demonstrated the Future of Oil and Gas Operations for the Minister of Digitalisation](https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/) — Aker BP; 2026 (page text does not print exact date; indexed as approximately May 2026); Lines 85–100: center tour, 2027 operating target, data infrastructure and prospective AI-agent scope.. Access: Issuer HTML opened directly; full relevant text available. ## Aker BP — Yggdrasil prepares its data foundation for remote operations T10-R06-GLOBAL-01 · Data, people & purchasing · As of 2026-04-15, planned for first production in 2027. Sensors/Cognite Data Fusion/integrated operations infrastructure is being prepared for future Yggdrasil production; AI agents are described as a first-production plan, not an already operating field result. Thousands of sensors are planned/being integrated; distinguish installed hardware from completed integrated operations. Status: Construction/commissioning preparation. Yggdrasil first production/centralized operation stated for 2027; digital operations center toured in April 2026. Economics: No Yggdrasil-specific data-platform or integration CAPEX, software fees, cybersecurity/IT cost, staffing baseline or realized production/maintenance benefit disclosed in this source. Baseline: No comparable conventional operating-model baseline; production not expected until 2027 per this source. AI role: AI is layered on top: Aker BP says agents will be used from first production for condition-based maintenance and planning and suppliers may share AI agents. Underlying foundation is sensors, Cognite Data Fusion, visualization and control/operating workflow, not itself an AI model. Limits: Architecture and intent described by operator, not production evidence. “Without need for manual offshore operations” applies to stated operating capability/ambition and does not establish no offshore personnel or intervention. The AI start date is forward-looking and contingent on production start and solution readiness. Next question: What is the all-in cost and operating model to integrate sensor, historian, platform, control system, visualization and supplier access—and how will go-live prove incremental value over a staffed offshore/control-room model? - [Demonstrated the Future of Oil and Gas Operations for the Minister of Digitalisation](https://akerbp.com/en/demonstrated-the-future-of-oil-and-gas-operations-for-the-minister-of-digitalisation/) — Aker BP; 2026-04-15 (newsroom archive listing); Lines 85-100; archive listing dated 15.04.2026. Access: Direct issuer page and issuer archive listing opened. ## Independent academic authors; public SPE E-Challenge benchmark — Predicting pump life becomes harder on a different well T20-R12-LIMITS-01 · Maintenance & reliability · Published to IEEE Xplore 2026-08-17; conference held 2026-08-02 to 2026-08-05. The IEEE abstract says no evaluated remaining-useful-life regression model exceeded R² = 0.24 under leave-one-well-out validation, a test intended to better reflect generalization to previously unseen wells. The best reported classification configuration achieved AUC 0.93 and F1 0.72 on the stated benchmark. These metrics apply to different tasks; they do not prove field performance or contradict other cases without matched data and evaluation design. Status: 2026 research paper / benchmark evaluation only; no live operator deployment claimed. Economics: No deployment cost or economic result; research benchmark only. Baseline: Authors contrast leave-one-well-out validation with same-well random splits/cross-validation, which they say can overstate performance for unseen-well deployment. The indexed abstract does not disclose the corresponding same-well metric values. AI role: Gradient-boosting machine-learning models classify ESP failure and estimate remaining useful life; authors use leave-one-well-out testing to assess transfer to unseen wells. Limits: The study is independently verified only to the extent of the IEEE publisher-indexed abstract. Do not imply full paper, methods, feature engineering, confidence intervals or class-specific metrics were audited. No deployment claim. Next question: Can a supplier show performance on held-out wells from the buyer's asset, with a prospective test and a decision threshold tied to avoided deferment and false-alarm/workover cost? - [Cross-Well Validated Binary Classification for ESP Failure Prediction Using Gradient Boosting](https://ieeexplore.ieee.org/abstract/document/11647491/) — IEEE Xplore; 2026-08-17; IEEE-indexed abstract and metadata; direct publisher-page open returns HTTP 418. Access: indexed primary abstract only; direct open failed again in R20. ## Acceleware Ltd. / Emissions Reduction Alberta-supported project — RF XL’s demonstration needed more power and further qualification T20-R18-LIMITS-01 · Subsurface & recovery · Demonstration result reported as completed in 2023, outside the 2024-10-07 to 2026-10-07 source window; ERA page accessed 2026-10-07 but update date is not stated. ERA reports that the downhole RF XL system delivered significantly less than planned power because of mechanical and operational implementation issues with the drilled well and did not fully demonstrate the technology. It separately reports full-power operation of the CTI and heat transfer using a commercial-length well pair. More subsurface upgrades and wells were needed. This is incomplete system qualification and added-cost exposure, not proof RF heating is inherently infeasible or that the project was permanently canceled. Status: ERA's page header labels the funded project 'Completed' and says the initial demonstration phase concluded in 2023. Its body separately says the ERA-supported pilot was paused, upgrades remained in progress, and restart options were being investigated. The page has no precise last-updated date, so current post-2023 operating status cannot be established from it. Economics: ERA page lists project value $28.2m and funding $5m; the page does not specify currency. Additional wells are said to add significant cost, without a revised budget. Do not label the values CAD or treat the project value as loss/write-off. Baseline: Planned high-power system performance and commercial-scale pilot objective; no numeric planned-versus-achieved kW or production baseline is given on the opened page. AI role: No AI role; this is radio-frequency subsurface heating and electric recovery technology, included as a physical-technology comparison. Limits: Project-funder summary, not an independent engineering report. It gives no detailed root-cause data, test duration, reservoir response, production or cost-per-barrel result. Demonstration concluded in 2023, outside the current evidence window. The page's current status is insufficient to establish whether upgrades/restart later occurred. Next question: What full-scale, multiwell subsurface performance and levelized oil/recovery cost must be demonstrated before an operator accepts the technology, and what additional capital is required after the inverter subsystem is qualified? - [Radio Frequency XL (RF XL) Enhanced Oil Recovery](https://www.eralberta.ca/projects/details/radio-frequency-xl-rf-xl-enhanced-oil-recovery/) — Emissions Reduction Alberta; Current project page; accessed 2026-10-07; project completion reported 2023; Lines 7–46. Access: directly opened project-funder page. ## Airbus — Skywise aviation data platform shows large fleet reach, with limited public outcome attribution T20-R19-GLOBAL-02 · Maintenance & reliability · FY2025 adoption and reach; published in Airbus 2025 Board Report in February 2026. Airbus reports Skywise user growth of approximately 16% in 2025 to approximately 55,000 and support for more than 11,900 commercial aircraft (Airbus and other OEMs). It says the platform enables significant annual airline savings by reducing flight delays and cancellations and improves aircraft operational availability, without disclosing dollars, avoided-event denominator, platform-only versus services attribution, or a matched baseline. About 54% of Airbus's in-service fleet was supported by Skywise and related digital solutions, including Health Monitoring, Predictive Maintenance and Fleet Performance. Status: Industrialized platform with 55,000 users across Airbus functions and divisions in 2025; external-customer platform usage is described as supporting aircraft operations and maintenance. Economics: No customer-specific implementation/hosting cost, annual dollar value, number of avoided delays/cancellations, benefit attribution to platform versus OEM services, or payback is disclosed in the annual report section. These claims cannot be carried over to upstream assets. Baseline: No flight-delay/cancellation baseline, aircraft cohort comparison or before/after window is specified. Adoption figures are scale metrics, not operating value metrics. AI role: Skywise includes data-platform and digital solutions such as predictive maintenance; the annual report does not attribute its reach or qualitative delay/cancellation savings specifically to AI or to one predictive model. Limits: Aviation is a cross-industry analogy for connected-asset platforms and adoption reach, not a comparable upstream technology deployment. Fleet count/users do not measure predictive model quality or avoided operating cost. Keep broad platform foundations distinct from AI use cases and from product-specific outcome evidence. Next question: For an upstream connected-asset platform, what field/asset denominator, active-user measure, operational workflow, cost base and validated before/after outcome can be disclosed beyond platform reach? - [Airbus SE Report of the Board of Directors 2025](https://www.airbus.com/sites/g/files/jlcbta136/files/2026-02/airbus_se_report_of_the_board_of_directors_fy_2025_1.pdf) — Airbus; 2026-02 (FY2025 annual report); PDF page 42, lines 1860-1876: platform adoption, aircraft reach, fleet share and stated qualitative savings.. Access: Issuer PDF opened directly; full relevant section available. ## Azule Energy / SLB — Delfi reservoir platform renewal and gradual enterprise-data integration T10-R06-LIMITS-01 · Subsurface & recovery · Prior Delfi use for two years by 2026-04-24 release; extension agreement reported 2026-04-24. Exact initial go-live date and expansion schedule not provided. SLB's 2026-04-24 release reports two years of Delfi use inside Azule's reservoir organization, followed by a three-year agreement to continue/expand. Workflow scope is reservoir studies, modeling, simulation, well planning; larger operational data integration is described as progressive ('over time'). No AI use, enterprise completion, migration cost or customer outcome is disclosed. Status: Three-year agreement to continue and expand use. It builds on two years of existing reservoir-organization use; enterprise-scale connection to broader operational-data environments is described as occurring over time. Economics: No contract value, migration/implementation cost, license spend, staffing requirement, NPV, or quantified operational or subsurface performance outcome disclosed. Baseline: The source establishes two years of prior use in the reservoir organization before the new agreement. It does not establish how many users, assets, datasets, workflows or systems were included during those two years, nor what was replaced or retained. AI role: Release frames platform use as digital/software. It does not say Azule deployed AI in the described workflows or attribute a result to AI. Treat as a data/workflow foundation unless more precise evidence emerges. Limits: Supplier/issuer announcement, not customer-authored implementation audit. Renewal and usage duration are not proof of enterprise-wide rollout, workflow adoption depth or operational value. The phrasing 'over time' signals progressive data integration but gives no delivery milestones, sources, standards, migration burden, data-quality controls, owner/accountability, or interoperability test. No inference that the platform operates field equipment or that generative/agentic AI is in use at Azule. Next question: What systems and datasets are actually connected, which workflows are routinely used by asset teams, who owns data quality and change management, and what measured value justified renewal after the first two years? - [SLB Announces First-Quarter 2026 Results](https://www.sec.gov/Archives/edgar/data/87347/000119312526174940/d111579dex99.htm) — SLB / U.S. SEC; 2026-04-24; Lines 9–36 (release date); 486–503, specifically 491–493 (Azule duration, workflows, contract and integration wording); 971–979 (forward-looking statement caveat). Access: Primary SEC-hosted issuer release opened directly. ## Baker Hughes / unnamed major Americas operator — An ESP model missed half the failed pumps in its test T20-R11-LIMITS-01 · Maintenance & reliability · Paper presented October 2024; underlying pump records are historical. A Baker Hughes paper reports 50% sensitivity, 78% specificity and 64% balanced accuracy. Nine failed pumps were missed; five involved short circuits. Nine misses at 50% sensitivity imply 18 failed-pump test cases. The authors identify insufficient SCADA resolution as a constraint. Status: The authors report client-approved deployment with daily predictions for currently running ESPs after the retrospective evaluation. The paper is supplier-authored and does not provide independent operator corroboration. Economics: One supplier-reported workover case claims $4.3m of reduced deferred production. It is not fleet-wide, recurring or net of full deployment costs. Baseline: The retrospective test uses the paper’s own alarm definitions. Maintenance or optimization after an alarm can change how a false alarm is classified. AI role: Ensemble combines XGBoost failure classifiers with reliability models and engineering damage rules; engineers review daily remaining-life rankings and choose interventions. Limits: Anonymous operator, small test and supplier authorship. The paper reports subsequent deployment; this is a measured model limitation rather than evidence of abandonment. There is no prospective controlled comparison or complete cost disclosure. Next question: What prospective sensitivity and false-alarm rate can the vendor demonstrate on a new asset using its actual installed SCADA resolution, and what incremental sensor or edge-compute cost is needed to address the short-circuit misses? - [An Artificial Intelligent Cutting-Edge Technology for Automated Field Production Optimization & ESP Failures Prediction (SPE-221571-MS)](https://dam.bakerhughes.com/asset/2145a967-01f1-4ffa-8ee1-37b085e6fff1/spe-221571-ms-bkr-licensed.pdf) — Society of Petroleum Engineers; paper by Baker Hughes authors; 2024-10-29/30; PDF pp. 1–2 (scope and prior methods); pp. 4–6 (population, split, metrics, confusion matrix and performance); p. 6 (SCADA-resolution limitation); pp. 8–10 (operationalization and individual examples).. Access: directly opened primary PDF. ## CNOOC — Typhoon-mode controls help keep offshore production running AR-CNOOC-01 · Production & process operations · Cumulative operating period not specified; report published 2025-09-09. CNOOC reports that upgraded controls allow production during typhoons and that typhoon mode cumulatively recovered production losses exceeding 600,000 tonnes of oil equivalent. Status: Operating system described in the 2025 interim report Economics: No project cost, cash benefit or AI-specific return disclosed. Baseline: The loss-estimation method and comparison events are not disclosed. AI role: AI supports the wider emergency-response system; the reported recovery is attributed to typhoon mode and upgraded controls, not to AI alone. Limits: Cumulative start date, event set and independent validation are not supplied. This is a broader control-system result; it is neither annual production nor an isolated AI gain. The sentence does not explicitly restrict that cumulative figure to Shenhai-1. Next question: How much shutdown time did the upgraded controls avoid, and what part required AI rather than remote instrumentation and operating procedures? - [CNOOC 2025 Interim Report](https://www.cnoocltd.com/english/investorrelations/reports/interimreport/202509/P020250909591355146082.pdf) — CNOOC Limited; 2025-09-09; Printed p. 15 / PDF page index 15, Scientific and Technological Innovation. Access: Full operator-hosted report opened by researcher and editor; relevant passage reviewed.. ## CNOOC Limited — Enping 15-1 adds offshore CO₂ reinjection for oil recovery T20-R13-GLOBAL-01 · Water, emissions & energy · 2025 commissioning; interim report published 9 September 2025. CNOOC’s 2025 interim report confirms that Enping 15-1’s offshore CCUS project entered operation. It combines reservoir carbon storage with oil recovery. The report does not quantify actual injection, incremental oil or operating costs. Status: Commissioned in 2025. The earlier CCS demonstration started in 2023; this note covers the later CCUS phase. Economics: No project cost, cost per tonne captured or measured oil increment is disclosed in the reviewed passage. Baseline: No Enping-specific pre-CCUS emissions/production baseline or matched counterfactual disclosed in the directly opened interim report passage. AI role: No AI is attributed to the capture, separation or injection mechanism. The disclosed technology is physical process equipment and reservoir injection; AI involvement in other Enping workflows is unknown. Limits: The available passage confirms commissioning and the broad mechanism. It does not establish the exact startup day or performance. An earlier research source referred to Bozhong 26-6; its capacity and production forecasts belong to that other field and were removed. Next question: What are measured annual capture/injection volumes, purity and uptime, incremental oil attributable to injection, monitoring/permanence requirements, and project capex/opex per net stored tonne versus the pre-injection baseline? - [Interim Report 2025](https://www.cnoocltd.com/english/investorrelations/reports/interimreport/202509/P020250909591355146082.pdf) — CNOOC Limited; 2025-09-09; PDF p.9 / lines 249–255 confirms commissioning of the offshore CCUS project in Enping 15-1; PDF p.15 / lines 395–433 per originating record for broader technical wording.. Access: Directly opened CNOOC primary PDF; p.9 claim passage inspected.. - [CNOOC Limited Brings On-stream Bozhong 26-6 Oilfield Development Project (Phase I)](https://www.cnoocltd.com/english/presscenter/pressreleases/2025/202502/t20250206_112957.html) — CNOOC Limited; 2025-02-07; Lines 13–26 identifies Bozhong 26-6 and its Phase I metrics.. Access: Directly opened CNOOC primary HTML; asset identity and metric passage inspected.. ## Canadian Natural Resources (CNRL) — Commercial-scale solvent-assisted SAGD at Kirby North R1-GL-02 · Subsurface & recovery · Evidence source: April 2025 CNRL corporate presentation, p. 54; Kirby North operation began 2024-06. Older background: Kirby South pilot reported in an undated company case study; do not treat it as the commercial-pad outcome. >80% is the fraction of solvent recovered, not energy savings, SOR reduction, GHG reduction, or cash savings. The 40-50% SOR and approximately $1/bbl figures are listed as expected benefits, not demonstrated Kirby North outcomes. Status: Commercial-scale Kirby North solvent pad continues operating in Q1 2026 per issuer May7 2026 update, reopened by root after Round3 review. Later status not checked. The separate Kirby South pilot was targeting Q2 2026 injection; actual start is not established. Economics: Commercial pad capex, solvent purchase/recycle cost, incremental facility cost, and actual per-barrel savings not disclosed. Baseline: No Kirby North matched-well before/after SOR or realized cost baseline is disclosed. Keep older Kirby South pilot results separate. AI role: None disclosed. Limits: Solvent recovery fraction alone does not demonstrate net energy, emissions, production, or cash benefit. Commercial-pad monitoring was still in progress. Do not transfer the pilot outcome or stated target to Kirby North or the full asset base. Next question: What is net solvent cost and cash margin per incremental barrel after solvent make-up, recovery, energy, throughput, and matched-pad SOR are included? - [Corporate Presentation, April 2025](https://www.cnrl.com/content/uploads/2025/04/WView_Corp-Pres_04.pdf) — Canadian Natural Resources; 2025-04 (exact publication day not stated); Printed p. 54, 'Technology & Innovation with Solvents: SAGD & CSS'. Access: Issuer-hosted PDF; indexed text accessible. - [Case Study: Solvent Enhanced Oil Recovery](https://www.cnrl.com/case-studies/case-study-solvent-enhanced-oil-recovery/) — Canadian Natural Resources; Undated web page; accessed 2026-10-07; Kirby South pilot; Kirby North plan; Primrose pilot description. Access: Issuer page opened via indexed search text; publication date not established. - [Source reopening in Round 2](https://www.cnrl.com/content/uploads/2025/04/WView_Corp-Pres_04.pdf) — See original source; 2025-04 (day not stated); Printed p.54, Kirby North solvent recovery and stated benefits. Access: issuer-hosted PDF direct retrieval unavailable; indexed issuer text reviewed. - [Source reopening in Round 2](https://www.cnrl.com/content/uploads/2025/05/Q125-Interim-Report.pdf) — See original source; 2025-05-08; Printed p.6, commercial-scale Kirby North pad, injection start, >80% solvent recoveries, workovers and monitoring. Access: issuer-hosted PDF opened. - [Source reopening in Round 2](https://www.cnrl.com/content/uploads/2025/08/WV_Corp-Pres_Aug.pdf) — See original source; 2025-08; Printed p.56; >80% solvent recovery; 40-50% SOR and ~$1/bbl listed as benefits. Access: issuer-hosted presentation indexed text reviewed. - [Additional root primary-source check](https://www.sec.gov/Archives/edgar/data/1017413/000101741325000043/a06302025q2pressrelease.htm) — Canadian Natural Resources; Date not established; Printed p.8, thermal in-situ section, lines229-230. Access: Issuer Q2 2025 results on SEC opened; relevant full text reviewed.. - [Additional root primary-source check](https://www.cnrl.com/content/uploads/2026/05/0507-Q126-Front-End.pdf) — Canadian Natural Resources; Date not established; Printedp5 lines166-172; May7datelineonp1. Access: Primary issuer PDF opened by root after Round3 draft review; printedp5 relevanttext reviewed.. ## Canadian Natural Resources (CNRL) — Kirby North solvent SAGD remains reported operating through Q1 2026; Q2 reporting is silent on its status T10-R01-LIMITS-02 · Subsurface & recovery · Kirby North first solvent injection June 2024. Q2/25 workovers executed on two well pairs; Q1/26 operation explicitly reconfirmed. Latest Q2/26 update reviewed omits current Kirby North status. The latest direct evidence found for Kirby North is “continues to operate” in Q1 2026 issuer report, not a Q2 update. Do not infer ceased operation from omission in Q2 report, or assert it continued in Q2. The Q2/26 release reports pilot-stage Kirby South target shifted to Q1/27; production performance cited is from new wells, not solvent-pilot results. The well/pad labels and solvent versus diluent terms must remain separate. Status: Q1 2026 is latest explicit confirmation found for Kirby North commercial pad operation. Q2 2026 leaves Kirby North unaddressed; separate Kirby South pilot has diluent injection target Q1 2027 following strong new-well production. No current Kirby North status/performance or solvent economics disclosed. Economics: No Kirby North realized SOR reduction, solvent recovery update, net cost savings, or GHG performance in Q1/26/Q2/26 updates reviewed. Historical >80% solvent recovery is from 2025 disclosures. Benefits of 40-50% SOR and ~$1/bbl remain targets/benefits, not demonstrated current commercial-pad outcomes. Baseline: No current matched-pair/control-pad baseline or full solvent purchase/recycle/steam cost disclosed. AI role: No AI role disclosed; physical reservoir/process technology. Limits: Do not call this a failure, successful economic scale-up, or current Q2 operating status without a newer explicit source. Preserve chronology: Q1/26 active; Q2/26 report does not restate status; Kirby South Q1/27 remains a target. Root-source-checks.json contains root’s explicit May 7 Q1 2026 source check; this task reopened that same issuer PDF and then checked the newer Q2 issuer/SEC release. Next question: Will CNRL disclose current pad SOR, production, solvent-recovery fraction, solvent losses/recycle cost, realized cash/steam savings and matched baseline—and does the separate Kirby South diluent pilot proceed in Q1/27? - [2026 First Quarter Results, front-end release](https://www.cnrl.com/content/uploads/2026/05/0507-Q126-Front-End.pdf) — Canadian Natural Resources; 2026-05-07; Printed p.5, thermal in-situ section, lines 166-172; report date p.1. Access: Official issuer PDF opened directly. - [Canadian Natural Resources Limited Announces 2026 Second Quarter Results](https://www.sec.gov/Archives/edgar/data/1017413/000101741326000043/a06302026q2pressrelease.htm) — Canadian Natural Resources / SEC EDGAR Form 6-K exhibit; 2026-08-06; Printed p.5 / webpage lines 192-193; six months ended June 30, 2026. Access: SEC issuer filing opened directly. - [Canadian Natural Q2 2025 issuer results](https://www.sec.gov/Archives/edgar/data/1017413/000101741325000043/a06302025q2pressrelease.htm) — Canadian Natural Resources / SEC EDGAR; 2025-08-07; Printed p.8, thermal in-situ section. Access: Primary source available via explicit prior source check and official search result; previous Round 3 review documents Q2 workover wording. ## Compañía General de Combustibles (CGC) — CGC standardizes post-acquisition enterprise systems and process governance T20-R16-GLOBAL-01 · Data, people & purchasing · Undated case-story implementation; the page itself prints no publication or go-live date. Its URL contains a 2025/09 asset path, which does not establish publication date. In-window timing is therefore unverified. CGC consolidated a fragmented post-acquisition application/process landscape around SAP Cloud ERP Private, documented integrations and mapped more than 100 critical processes. It reports fewer than 70 customizations, compared with a stated industry average of 550 for ERP programs of this size, and says aligned procurement workflows replaced inconsistent legacy practices. This supports system/process standardization and acquisition-integration capability, not an upstream production outcome. Status: SAP customer story describes a transformation implemented and in use, with next-phase real-time process control-tower capabilities still planned. No start, go-live or completion date is disclosed in the article; its URL places it in SAP's 2025/09 asset path. Economics: No implementation cost, recurring license/hosting spend, procurement savings, labor reduction, cycle-time change or financial return is disclosed. The planned real-time control tower is a future phase, not an achieved result. Baseline: Before program, acquired businesses had overlapping applications, undocumented integrations, different procurement tools/workflows and extensive customization. No quantified before count of CGC customizations, application inventory, process cycle time or normalized ERP peer denominator is supplied. AI role: No AI role is disclosed. SAP Cloud ERP Private is the central ERP; SAP LeanIX inventories applications/integrations, SAP Signavio maps process workflows, SAP Enable Now supports user training, and SAP UEM tracks application adoption. These are digital foundations and change-management tools, not evidence of an AI deployment. Limits: Vendor-authored case study and executive quotation; result denominator for the 550 customization comparison is not detailed. Company-wide scope is not equivalent to field deployment. Keep in a digital-foundations register separate from AI initiative counts; a future AI story would need a distinct use case and evidence. Next question: What was the actual go-live and recurring cost, and did harmonized procurement, close, maintenance or field-support workflows measurably improve after the acquired E&P systems were integrated? - [CGC: Harmonizing fragmented systems to drive a company culture of clarity and efficiency](https://www.sap.com/asset/dynamic/2025/09/28aaac43-227f-0010-bca6-c68f7e60039b.html) — SAP; Undated (URL path includes 2025/09; publication date not printed or verified); Lines 11–25 and 41–64; process counts, customization claim, acquired-system fragmentation and vendor/customer narrative.. Access: Directly opened SAP primary customer story.. ## Continental Resources — Continental’s recent asset-level operating detail remains a research gap T10-R01-LIMITS-03 · Production & process operations · No qualifying deployment period found in official source searches for 2024-10-07 through 2026-10-07. Retain only as bounded search gap: official pages reviewed are broad/legacy capability pages with no dated 2024-10-07–2026-10-07 named technology outcome. No conclusion about innovation level or adoption. Status: Search gap. Current official pages describe ECO-Pad and horizontal drilling in broad/legacy terms; no recent named technology deployment with date, scale and result located. Economics: No recent source with technology-level capex, unit cost, savings or performance denominator found. Baseline: None; no recent specific initiative matched against a comparator. AI role: Unclear; no recent dated technology deployment located. Legacy corporate materials do not establish current AI or non-AI deployment. Limits: Bounded search of company site and targeted query phrases; no exhaustive web/regulator/technical-literature search. The operational pages may be undated and describe established capability; current page retrieval date is not a launch date. Next question: Can current company, state regulator, supplier contract or technical-conference evidence identify a 2024-2026 Continental asset deployment with method, date, scale and measured outcome? - [Oil & Natural Gas Operations](https://www.clr.com/operations/) — Continental Resources; Undated current page; accessed 2026-10-07; ECO-Pad/horizontal drilling overview; page has no recent dated deployment or output denominator. Access: Official company page opened. - [Horizontal Drilling](https://www.clr.com/operations/horizontal-drilling/) — Continental Resources; Undated current page; accessed 2026-10-07; General explanation of horizontal drilling; no recent specific project/result. Access: Official company page opened. ## Continental Resources — Continental’s current public materials leave deployment questions unanswered T10-R01-US-03 · Production & process operations · Latest located SEC filing is FY2025 Form 10-K; company operations pages describe long-running historical methods. Continental’s FY2025 10-K describes its technical toolkit at a high level—3D seismic, horizontal drilling, geosteering, advanced completions, pad/row development and enhanced-recovery technologies—and states it is a private company that continues required SEC filings. Company operations pages describe ECO-Pad as a decades-old concept and advanced completions, without a recent named implementation, denominator, measured KPI or cost outcome. Status: No recent named technology stage verified in targeted public sources. Economics: No recent initiative-specific economic result or deployment-cost denominator identified. The 10-K offers portfolio/financial disclosures, but generic technology categories cannot support a comparative technology ROI claim. Baseline: Not disclosed for any newly deployed initiative in sources inspected. Historical ECO-Pad and early Bakken technical milestones are outside the current publication window and are not represented as launches. AI role: Unclear; recent filings do not specify AI role for current operating programs. Limits: Targeted search/access gap, not proof Continental has no current technology deployment. SEC filing page was surfaced in search index but direct open/search-within failed in this browser; CLR pages are issuer material and largely historical/generic. Private status narrows disclosure, but its debt-indentured SEC reporting remains a source to monitor. Next question: Can debt-issuer filings, state permits, technical conference papers or operator/vendor technical case studies identify a named 2024–2026 Continental deployment with start date, eligible asset denominator and actual unit-cost/recovery/safety outcome? - [Continental Resources, Inc. Form 10-K for year ended December 31, 2025](https://www.sec.gov/Archives/edgar/data/732834/000119312526064403/ck0000732834-20251231.htm) — Continental Resources / SEC EDGAR; 2026-02-23; SEC filing-detail page confirms filing date 2026-02-23 and links FY2025 10-K; search-indexed filing text gives private-company status and general technical capabilities. Main 10-K body direct open/search was unavailable.. Access: SEC filing-detail index opened; 10-K body itself only search-indexed. - [Oil & Natural Gas Operations](https://www.clr.com/operations/) — Continental Resources; Undated current website page; Operations page: ECO-Pad history and up to 12 wells per pad; historical/general capability language.. Access: Issuer HTML opened; lines 17–32 show historical ECO-Pad account and current high-level capabilities. ## EQT Corporation / EOG Resources / Diamondback Energy / Devon Energy — Company well-cost figures need different comparisons T10-R09-US-02 · Data, people & purchasing · FY2025 actual versus FY2024 for EOG/EQT; Q1 2026 individual-well record and Q2 2026 equipment/performance observations for Diamondback; Devon plan/2025 bundled efficiency report. Preserve distinct measures: EOG 7% portfolio average well-cost decline; EQT 13% average cost-per-foot decline; Diamondback Q2 ~14% equipment-cost-per-well QoQ decline, and separate one-well Barnett record; Devon's $1bn plan targets and bundled metrics. No source provides a shared cost boundary and matched eligible population. Status: Reported FY2025 outcomes, Q1/Q2 2026 operating observations, and 2026 guidance/targets; these stages and date bases must remain labeled. Economics: Do not put the percentages into one savings leaderboard. The KPI numerators differ (whole average-well cost, cost/foot, equipment-only cost/well, speed/throughput and program-level capex), while periods, basin mix and design populations differ. A usable future cross-company panel should require actual USD/gross completed lateral foot with the same cost boundary; retain completion cost and drilling cost separately, identify WI/net-gross basis, and pair with a defined eligible-well population and production/recovery outcome. Keep throughput and uptime in separate panels. Baseline: EOG FY2025 versus FY2024 portfolio average; EQT FY2025 versus FY2024 cost per foot; Diamondback equipment cost/well Q2 versus Q1 2026 and one-well Barnett record; Devon metrics compare year over year and broad plan versus initial guidance. These are not matched-control or common-price/current-dollar baselines. AI role: Not applicable to the arithmetic; the reported cost results are company-level outcomes and are not attributable to one AI or hardware intervention in the source evidence reviewed. Limits: Issuer definitions and cost scopes do not reconcile; no same-basin matched well cohorts across operators; EOG/EQT disclosures do not offer cost numerators; well length, lateral-foot denominator, completion intensity, facilities, water/sand, workovers and service-price/mix effects are not fully controlled. Devon >$250m program capital reduction is not technology savings. Do not infer owner returns or IRR without project capex and production/cost response. Next question: Can each issuer provide a same-basin/well-class bridge from actual well capex to cost per completed lateral foot, with drilling, completion, facilities and logistics separated and matched to the production/recovery response and eligible well count? - [EOG Resources Reports Fourth Quarter and Full-Year 2025 Results; Announces 2026 Capital Plan](https://filecache.investorroom.com/mr5ir_eogresources2/444/4Q_2025_Earnings_Press_Release.pdf) — EOG Resources; 2026-02-24; FY2025 highlights and 2026 Outlook, PDF p. 1-2, lines 7-19 and 53-62.. Access: Issuer-hosted PDF opened directly. - [EQT Reports Fourth Quarter and Full Year 2025 Results and Provides 2026 Guidance](https://ir.eqt.com/newsroom/news-releases/news-release-details/2026/EQT-Reports-Fourth-Quarter-and-Full-Year-2025-Results-and-Provides-2026-Guidance/default.aspx) — EQT Corporation; 2026-02-17; Fourth-quarter and recent highlights, lines 132-134.. Access: Issuer HTML opened directly. - [Form 8-K for Diamondback Energy Inc. filed Aug. 3, 2026 (Q2 2026 results)](https://ir.diamondbackenergy.com/static-files/3edc9cca-d7ee-45e5-b420-1b2b8fab639f) — Diamondback Energy; 2026-08-03; Q2 2026 Operational Performance, PDF p. 22, lines 719-725.. Access: Issuer-hosted PDF opened directly. - [Devon Energy Unveils Value Enhancing Business Optimization Plan](https://investors.devonenergy.com/investors/press-releases/press-release-details/2025/Devon-Energy-Unveils-Value-Enhancing-Business-Optimization-Plan/default.aspx) — Devon Energy; 2025-04-22; Capital efficiency and production optimization target buckets, lines 46-58.. Access: Issuer HTML opened directly. ## Ecopetrol S.A. — Produced-water reinjection for recovery and pressure support totals 160.6 million m3 across Colombian upstream assets T20-R14-GLOBAL-01 · Water, emissions & energy · Calendar year 2025 vs 2024. The issuer's 2026-04-24 current water-management page reports the figures. Ecopetrol reports 161.0 million m3 of produced water reused in upstream in 2025, up 12% from 2024 and equivalent to 30.3% of produced water generated. It says the increase was mainly caused by better performance in production-water reinjection for recovery and/or pressure maintenance in named assets. Of the 161.0 Mm3 reused, 160.6 Mm3 was reinjected for those purposes. Status: Operating 2025 water-management practice; issuer's 2026-04-24 report describes an annual result and year-over-year change. Economics: No field-specific incremental oil/gas, injection cost, treatment energy, water-quality threshold, or payback disclosed. This measures reinjection volume, not incremental recovered barrels; pressure maintenance and recovery are combined purposes. Baseline: 2024 produced-water reuse volume is the percentage-growth comparison, but absolute 2024 volume and field-level comparison are not given in the cited passage. The 30.3% denominator is total produced water generated in 2025. AI role: No AI is attributed. The described mechanism is produced-water separation and reservoir reinjection for pressure maintenance and/or hydrocarbon recovery; analytics involvement is not specified. Limits: Issuer aggregate for multiple assets and mature water-management mechanisms; no unique project name, field-by-field split, matched control or causally measured production uplift. Keep the 181 Mm3 total-water reuse figure separate (includes non-upstream streams) and 160.6 Mm3 recovery/pressure reinjection separate from the 283.3 Mm3 disposal injection total. Next question: For each named field, what produced-water volume and quality entered reinjection, what share supported pressure versus incremental recovery, and what incremental production, decline change, treatment/injection cost and water-disposal savings were measured against a matched baseline? - [Operational efficiency in water management](https://www.ecopetrol.com.co/wps/portal/Home/tesg/environmental/towards-water-neutrality/water%20management%20efficiency) — Ecopetrol; 2026-04-24; Page lines 227-261: date, operating totals, upstream produced-water reuse and reinjection purpose/asset list; lines 273-276 distinguish disposal injection.. Access: Issuer HTML opened directly. ## Equinor — Hydrone R resident subsea intervention drone at Njord R1-GL-01 · Maintenance & reliability · Issuer article dated 2025-06-25; 240-day operating period ended 2025-01-24. Duration/residency milestone. Vessel/support-boat reduction is qualitative; no avoided vessel-days, uptime delta, cost or emissions result disclosed. Status: Operational trial on an operating field; 240-day continuous submersion milestone completed. This is not evidence of fleet-scale commercial rollout. Economics: Development/testing investment and per-day operating cost not disclosed. Baseline: Conventional comparison described qualitatively as support vessels traveling to/remaining at field and daily ROV deployment/recovery. No matched vessel-day or cost baseline disclosed. AI role: No AI contribution disclosed. Limits: A record-duration trial proves subsea residency and field operation, not economics or replacement of crewed vessels across a field. Equinor still reports divers working year-round. Inspection/light maintenance scope is narrower than heavy intervention or repair. Next question: What vessel-days, interventions, maintenance tasks, and all-in vessel/subsea support cost did this single 240-day residency displace versus a crewed-ROV baseline? - [A deep dive into emission cuts](https://www.equinor.com/magazine/a-deep-dive-into-emission-cuts) — Equinor; 2025-06-25; Hydrone R sections; operating period and 240-day milestone, lines 46-65. Access: Issuer page opened; full text accessible. - [Source reopening in Round 2](https://www.equinor.com/magazine/a-deep-dive-into-emission-cuts) — See original source; 2025-06-25; Hydrone R section; continuous residency, 330 m depth, May 20 2024-Jan 24 2025 dates. Access: official indexed text reviewed; direct page fetch timed out. ## Equinor / Saipem — Hydrone-R adds a longer operating history at Njord T10-R05-LIMITS-01 · Maintenance & reliability · Current Saipem update dated 2026-02-06 reports >500 total residency days since start and >240-day continuous stay; it describes a well commissioning activity at end-2025. Prior specific stay was 2024-05-20 to 2025-01-24 (240 days) in Equinor's June 2025 retrospective. Saipem reports >500 cumulative residency days, a 240-day continuous seabed stay, hundreds of Njord missions, and its claimed contribution to late-2025 well commissioning during 12.5m waves. Residency is not uptime; support to one commissioning activity is not proof all conventional ROV tasks are displaced. No availability/failure/success/cost denominator. Status: Operating under the service contract. Supplier says it has performed hundreds of missions; in late 2025 it enabled activities supporting commissioning of a Njord well. It completed a 240-day continuous seabed stay, not a year-round uptime test. Economics: No contract value, per-mission cost, avoided-vessel-day count, comparative cost, availability or verified emissions avoided disclosed. Benefits are described qualitatively. Baseline: A previous Equinor source reports 240 continuous days at 330 m depth before recovery. The 2026 Saipem update calls this a continuous stay/residency record; it does not state continuous active operation or uptime. Project page says both Hydrone-R and Hydrone-W are remotely controlled from Njord-A or onshore; do not conflate support from an autonomous vehicle with fully autonomous operations. AI role: AI-based navigation is a supplier-declared design capability in project context; do not attribute each mission/outcome to AI. Limits: Supplier claims only. No data on failed missions, interventions, uptime/availability denominator, faults, sensor detection/false alarms, maintenance hours, cable/docking availability, operator safety incidents or missions unsuitable for weather. The 12.5 m wave example is a bounded mission/commissioning support claim, not proof that all conventional ROV work can be displaced. Existing research already covers the 240-day milestone; this is only a later 2026 operational update. Next question: Before scaling a resident subsea robot, what independently reported mission-success, availability, intervention, maintenance, false-detection and avoided-vessel-day metrics would establish safety and full-cost advantage over conventional ROV support? - [Saipem’s underwater drone Hydrone-R performs autonomous mission in the Arctic coral area in Norway](https://www.saipem.com/en/media/news/Saipem-s-underwater-drone-Hydrone-R-performs-autonomous-mission-in-the%20Arctic-coral-area-in-Norway) — Saipem; 2026-02-06; Lines 116–126. Access: Primary supplier release opened directly. - [The Hydrone project in the Njord Field](https://www.saipem.com/en/saipem-worldwide-projects/hydrone-submarine-development-njord-field) — Saipem; Project page; accessed 2026-10-07; Lines 115–185, especially 130–176. Access: Primary supplier project page opened directly. - [From pioneer diving to autonomous drones](https://web-equinor-web-sites-prod.c2.radix.equinor.com/magazine/a-deep-dive-into-emission-cuts) — Equinor; 2025-06-25; Lines 40–65 and 89–96. Access: Primary operator article opened directly. ## Equinor and Wisting license partners — Wisting field power concept: power from shore assessed and ruled out; gas-turbine power generation retained for concept maturation R1-LIM-01 · Production & process operations · Source published 2026-06-25. Equinor references the earlier project postponement in 2022; this round uses the 2026 primary source for current status. Power from shore was ruled out after detailed assessment on technical complexity and cost. Equinor continues concept work around an energy-efficient gas turbine. CCS remains conditional on technology, regulatory requirements, and acceptable cost/time. Partners: Equinor 42.5%, Aker BP 27.5%, Petoro 20%, INPEX Idemitsu Norge 10%, subject to a pending partner transaction approval. No cost baseline or realized project outcome disclosed. Status: Keep as early concept maturation, no FID; source indicates concept selection possible by end-2026 and FID planned end-2027. Economics: Not disclosed Baseline: Not disclosed AI role: No AI evidence. This is a non-AI infrastructure/concept choice that sets the boundary for where a later AI workflow could be deployed and powered. Limits: Operator announcement; no cost comparison, engineering study, emissions comparison, quantified schedule impact, or independent review published in this source. Wisting remains a possible development, not a sanctioned project. Do not describe electrification or CCS as failed technologies: only this project’s specific power-from-shore concept was screened out. Next question: What lifecycle cost/schedule/reliability/emissions study supported rejecting shore power versus gas turbine, and what cost/technical threshold could reverse the decision? - [Proposal for environmental impact assessment programme for Wisting submitted for public consultation](https://www.equinor.com/news/20260625-wisting-proposal-environmental-impact-assessment-programme) — Equinor; 2026-06-25; Project status, power-from-shore decision, gas-turbine concept, CCS conditions, decision timeline and licensee interests; paragraphs 2–7 and facts.. Access: Opened; issuer source. - [Source reopening in Round 2](https://www.equinor.com/news/20260625-wisting-proposal-environmental-impact-assessment-programme) — See original source; 2026-06-25; Project status and concept options; published EIA-programme notice.. Access: issuer HTML opened. ## Equinor and Snøhvit partnership — Snøhvit Future onshore compression and Hammerfest LNG electrification R1-LIM-02 · Water, emissions & energy · Key status/cost sources published 2025-12-22 and 2026-10-07. Original PDO estimate submitted 2022; included for baseline only. Current estimate NOK 26.5bn (2026 value) versus NOK 13.2bn in the 2022 PDO, which Equinor restates as NOK 15.2bn in current-value terms; thus approximately NOK 11.3bn / 74% above the inflation-adjusted baseline. Cost and schedule were affected by complexity integrating into an operating LNG plant, operating challenges, weather, difficult rock in the cable tunnel, shutdowns/turnaround, and equipment inflation. 2026 update says 60% complete and timetable unchanged. Owners: Equinor 36.79%, Petoro 30%, TotalEnergies EP Norge 18.4%, Vår Energi 12%, Harbour Energy Norge 2.81%. Status: Issuer-indexed Oct. 7 update says ~60% complete, compressor module installed in August 2026, schedule unchanged; compression planned 2029 and electrification 2030. Economics: Current issuer-indexed estimate NOK 26.5bn in 2026 value vs NOK 15.2bn restated 2026 value for the 2022 PDO (original nominal PDO NOK 13.2bn): +NOK 11.3bn, about +74% versus the restated baseline. Label estimate, not realized cost. Baseline: Not disclosed AI role: No AI role. It is an adjacent upstream asset reliability / emissions-capex case; any AI system would need to work around live-plant construction, turnarounds, and commissioning constraints. Limits: The emissions reduction is a forecast, not a measured result. Cost attribution is qualitative; the source does not break out the increase by cause or supplier package. Equinor says the project remains profitable, but publishes no current NPV, return threshold, or sensitivity in the cited update. Do not call this a technical failure: it is a costly, difficult live-plant integration project still progressing. Next question: How much of NOK 26.5bn vs NOK 15.2bn restated PDO is attributable to integration, civil works, inflation, shutdowns and scope, and what is remaining cost-to-complete / contingency? - [Updated cost estimate for the Snøhvit Future project](https://www.equinor.com/news/20261007-updated-cost-estimate-snohvit-future) — Equinor; 2026-10-07; Cost estimate and baseline; causes; supplier work; 60% complete; purpose, dates and ownership; lines 15–45.. Access: Opened; issuer source. - [Status update for the Snøhvit Future project](https://www.equinor.com/news/20251222-status-update-snohvit-future-project) — Equinor; 2025-12-22; Original change from plan, NOK 20bn+ estimate in 2025 values, one-year compression delay and operating-plant complexity; lines 19–43.. Access: Opened; issuer source. - [Source reopening in Round 2](https://www.equinor.com/news/20261007-updated-cost-estimate-snohvit-future) — See original source; 2026-10-07; Issuer news index identifies current estimate, timetable and Oct. 7 publication.. Access: direct URL inaccessible in browser; issuer newsroom index/search snippet surfaced source text, so detailed claims not fully reopened. - [Source reopening in Round 2](https://www.equinor.com/news/20251222-status-update-snohvit-future-project) — See original source; 2025-12-22; Intermediate cost/schedule update; operating facility, schedule delay, weather/tunnel/turnaround/equipment factors.. Access: issuer HTML opened in prior source audit. - [Additional root primary-source check](https://www.equinor.com/news/20261007-updated-cost-estimate-snohvit-future) — Equinor; Date not established; Cost/status/causes paragraphs and project facts. Access: Primary HTML opened 7 October 2026; relevant full text reviewed.. ## Equinor / Snøhvit partnership — Snøhvit Future’s cost rises as compression and electrification advance T10-R09-GLOBAL-01 · Water, emissions & energy · Cost updated 2026-10-07; emissions reduction is a future annual estimate after electrification. As of 7 Oct 2026, Equinor revised total Snøhvit Future project cost to NOK 26.5bn (2026 value) versus original nominal NOK 13.2bn (2022 PDO); inflation-adjusted comparison is NOK 15.2bn in 2026 value. Project reported 60% complete. Compression is planned for 2029; electrification for 2030. This is integrated project cost for production-maintaining compression plus electrification; not an abatement cost per tonne or evidence of technology failure. Emissions-cut estimate 850,000 tCO₂/year is a project forecast, not delivered reduction. Status: Construction about 60% complete as of 2026-10-07; compressor module installed Aug 2026; onshore compression planned 2029, electrification planned 2030. Economics: Total integrated compression/electrification project estimate NOK26.5bn in 2026 money versus issuer inflation-adjusted PDO equivalent NOK15.2bn: about74% increase (analyst arithmetic). No split by work package, lifetime, electricity expense, preserved gas cash flow or emissions denominator sufficient for lifecycle abatement cost or technology ROI. Omit the capital-per-annual-emissions ratio from comparative exhibits. Baseline: 2022 PDO estimate NOK 13.2bn, inflation-adjusted by Equinor to NOK 15.2bn in current value; current estimate NOK 26.5bn 2026-value. 850,000 tCO2/year is estimated future annual reduction from LNG electrification, not achieved and not full Snøhvit economics. AI role: No AI role in this evidence; project comprises compression at Hammerfest LNG and grid electrification infrastructure. Limits: Project combines production-maintenance compression and LNG plant electrification, so do not attribute total CAPEX solely to emissions abatement. Equinor names operational challenges at Hammerfest LNG, weather and difficult rock in cable tunnel; source gives project maturity (~60%) but no detailed schedule variance logic. Shareholder partnership economics differ from operator-only economics. Next question: How much of the revised NOK 26.5bn supports plateau production versus electrification, and what lifetime gas value plus avoided emissions underpin the partnership's profitability claim? - [Updated cost estimate for the Snøhvit Future project](https://www.equinor.com/news/20261007-updated-cost-estimate-snohvit-future) — Equinor; 2026-10-07; Lines 15-45. Access: Direct issuer page opened. ## Equinor — 2024 partial electrification across Troll B/C, Sleipner center and Gudrun T10-R04-GLOBAL-02 · Water, emissions & energy · During 2024; report published 20 March 2025. Partial electrification is reported at Troll B/C, Sleipner center and Gudrun; Equinor’s 5% year-on-year operated Scope 1+2 reduction is a portfolio result and cannot be allocated to these projects. Status: Partly electrified during 2024 according to Equinor's 2024 annual-report release. Economics: No asset-level electrification CAPEX, power cost, or payback disclosed in the reviewed annual-report release. Baseline: Portfolio-wide operated emissions baseline is 2015; year-on-year 2024 total. No asset-level baseline or counterfactual. AI role: No AI role disclosed; grid power and platform modifications are physical infrastructure. Limits: Partial electrification only; do not imply every platform load was electrified or assign the company-wide emissions decrease to this set of assets. Equinor separately stopped two early-phase electrification projects in October 2025 primarily due to high abatement costs (covered in R08/R09 context, not a same-project duplication). Next question: Which asset-level emissions/cost changes are attributable to each electrification project after grid and production-mix effects? - [Equinor presents 2024 Annual report](https://www.equinor.com/news/20250320-integrated-annual-report-2024) — Equinor; 2025-03-20; Lines 52-56. Access: Direct issuer release opened. ## Equinor — Early-phase Snorre and Halten electrification work stopped over economics T10-R08-GLOBAL-01 · Water, emissions & energy · Decision October 2025, disclosed in Equinor Q3 2025 results. Equinor stopped early-phase Snorre/Halten electrification work in October 2025 primarily because of high abatement cost. These were pre-operation capital-allocation decisions; no project-specific measured emissions/economics or evidence of technical failure is supplied. Status: Early-phase project plans discontinued in October 2025; not an operating technology failure. Economics: Decision explicitly economic; project-specific CAPEX, power tariffs, NOK/tCO2 decision threshold and sunk costs not disclosed in the Q3 release. Do not compare potential annual tonnes to realized savings. Baseline: No project-specific baseline emissions or achieved reduction is stated in the reviewed Q3 release; these were planned projects that did not proceed to deployment. AI role: No AI role; this was grid, cables and offshore platform modification engineering. Limits: Cancellation was due primarily to high abatement costs; this is not a failure of an operating electrification system. Do not label it as technical failure or infer economics beyond the stated reason. Next question: For proposed offshore power projects, which input costs (CAPEX, grid connection, electricity versus saved gas, policy support) pushed abatement cost over the partnerships' investment threshold? - [Equinor third quarter 2025 results](https://www.equinor.com/news/equinor-third-quarter-2025-results) — Equinor; 2025-10-23; Lines 26-35 and 87-89. Access: Direct issuer results page opened. ## Equinor — Equinor’s production-efficiency figure does not isolate AI’s contribution T20-R13-LIMITS-01 · Production & process operations · 2025 results; report signed 2026-03-09. Equinor's report asserts tangible AI value in 2025 and describes AI uses, but does not quantify an AI-specific production-efficiency contribution in the cited passage. It states that technology implementation and in-house expertise are central to sustaining ~92% NCS production efficiency; surrounding examples bundle AI/analytics with mature optimization, compressor washing, integrity logging, additive-manufactured components and online valve work. Åsgard A avoided-loss claim is attributed to a set of Next-Generation Maintenance interventions, not isolated AI. Status: 2025 annual-report summary of 2025 operational work and future ambitions. Economics: No AI-specific cost, savings, capital, or payback disclosed. Production-efficiency rate is a portfolio operating metric, not an AI ROI measure. Baseline: ~92% NCS PE is company-reported; method, exact population and comparison period are not specified in the cited innovation passage. AI role: The report names AI and advanced machine learning across well planning, predictive maintenance and operational efficiency, but does not link a measured field result to a specific model or workflow. Limits: Corporate self-report; portfolio-level denominator and exact calculation not specified in the cited passage; diverse non-AI interventions confound attribution. Do not infer AI caused the production efficiency or Åsgard A avoided-loss outcome. Next question: For an AI business case, can the operator identify a named workflow's eligible asset population, baseline, intervention date, production/deferment effect and incremental cost separately from routine maintenance and engineering work? - [Equinor 2025 Annual Report, section 2.4 Fuelling innovation](https://www.sec.gov/Archives/edgar/data/1140625/000114062526000013/exhibit154equinor2025ann.htm) — Equinor / SEC exhibit; 2026-03-09 (report approval date); Lines 548–556. Access: directly opened filed annual-report HTML. ## Equinor / Microsoft; OSDU Forum documentation — OSDU still leaves implementation and data-mapping work to do T20-R16-LIMITS-01 · Data, people & purchasing · Equinor customer story dated 2025-10-07; OSDU EDS documentation current as of 2025-09-09. Equinor/Microsoft report an OSDU-aligned Azure Data Manager configuration integrated into OMNIA and say hundreds of thousands of quality-controlled well logs are stored in cloud. The story says OSDU can be difficult to implement/maintain and on-prem instances may lack compute/scalability. OSDU EDS docs describe prior manual fetching/custom ETL and delay, and state reference mapping needs operator configuration. The evidence supports integration/implementation diligence; it does not show Equinor incurred a measured migration penalty, vendor lock-in or an exit failure. Status: Equinor reports platform use and hundreds of thousands of quality-controlled well logs available in cloud. OSDU EDS documentation (dated 2025-09-09) describes a standard service for metadata and external-data ingestion. Economics: No actual migration cost, egress fee, measured integration labor or migration ROI is disclosed. Equinor's quoted '250 years' is a vendor-published counterfactual and not an independently audited saving. Baseline: Prior state described by Equinor: data silos, manual search/transfer/reporting. OSDU EDS documentation describes pre-EDS manual fetching and separate custom ETL per provider; no shared numeric time/cost baseline. AI role: OSDU is a data foundation aligned with later analytics/AI use; the cited operator story reports workflow/data modernization, not a measured AI result. Limits: Customer story is vendor-authored; no independent cost or availability audit. Platform implementation is cloud-specific in the described Equinor configuration; OSDU alignment alone does not establish portability, data egress rights, model portability, or vendor-neutral total cost. The OSDU documentation describes capabilities and design intent, not comparative operator performance. No actual lock-in or exit failure is established. Next question: Does the procurement contract make data, metadata, mappings, APIs, derived features and model outputs exportable, and what are the measured transfer, egress, mapping-maintenance and replatforming costs under a vendor-exit test? - [Equinor targets data efficiency and a low-carbon future with Microsoft Azure Data Manager for Energy](https://www.microsoft.com/en/customers/story/25408-equinor-microsoft-azure-data-manager-for-energy) — Microsoft Customer Stories (with Equinor statements); 2025-10-07; Lines 28–38, 47–59, 62–67. Access: directly opened vendor customer story. - [Business Perspective of EDS](https://osdu.pages.opengroup.org/platform/data-flow/ingestion/osdu-airflow-lib/business-value/) — OSDU Data Platform documentation, The Open Group pages; 2025-09-09; Lines 65–87, 93–102. Access: directly opened OSDU documentation. ## Halliburton / Eni — Closed-loop surface, subsurface and managed-pressure drilling integration offshore Indonesia T10-R02-LIMITS-02 · Drilling & completions · Reported 2026-07-09; exact drilling dates and well identifier not provided. Halliburton says one deepwater exploration well with Eni offshore Indonesia; >15% efficiency is undefined. Preserve as supplier-reported and do not treat 'industry first' or 'scales' language as independent verification. Status: Halliburton-announced successful deployment on one deepwater exploration well; described as first full rig automation + MPD deployment in Asia-Pacific. Economics: No absolute rig hours, cost, NPT delta, pressure excursion statistics, or contract terms disclosed. Baseline: No metric definition, comparator well, time period, or efficiency denominator accompanies >15% in the press release. AI role: automation/closed-loop orchestration confirmed; this release does not specify a machine-learning method. Limits: Source is Halliburton’s supplier release, though Eni is named. One well does not establish general deepwater replicability, full automation boundary, safe fallback, or failure rates. “Industry first” is provider framing. No AI-specific method in this source. Next question: For an operator evaluating closed-loop drilling plus MPD, what were the actual controllable parameters, hold points, human override/failsafe rules, pressure excursions, uptime/NPT, and baseline behind the >15% efficiency figure? - [Eni and Halliburton Achieve Industry First With Closed-Loop Rig Automation in Deepwater Indonesia](https://ir.halliburton.com/node/36466/pdf) — Halliburton; 2026-07-09; One-page official release, lines 0-20; deployment, integrated control stack, >15% efficiency claim. Access: Official Halliburton investor PDF opened directly. ## Helmerich & Payne — Private-equity-backed independents are in the buyer pool, but H&P discloses no PE-specific technology adoption T10-R07-LIMITS-01 · Drilling & completions · Fiscal year ended 2025-09-30; point-in-time fleet data at 2025-09-30. H&P states PE-backed private operators exist in its U.S. customer segment. Its largest North America customer is unnamed and represents 18.2% ($429.4m) of North America Solutions revenue, not consolidated company revenue. 73 fixed-term and 71 well-to-well contracts are rig-count data, not buyer counts or technology procurement mix. Status: FY2025 operating/customer-mix disclosure; 144 of 223 marketed U.S. rigs were active at 2025-09-30. Of active rigs, 73 were on fixed-term contracts and 71 working well-to-well. Economics: Filing gives aggregate vendor revenue concentration, not technology-specific price, adoption, savings or margin. Largest North America customer revenue is not identified as PE-backed. No procurement timelines, software/equipment contracting terms, or PE-specific return data provided. Baseline: As of September 30, 2025, H&P had 144 active out of 223 marketed U.S. rigs, with 73 active under fixed-term contracts and 71 working well-to-well. These are rig counts, not technology adoption rates or customer counts. AI role: The filing describes H&P's drilling-technology portfolio broadly but does not identify an AI or non-AI technology purchase, feature, or result for PE-backed customers. Limits: Do not present the presence of PE-backed customers as evidence H&P sells FlexRobotics or another technology to PE-backed buyers. The largest customer is not named and its ownership is unspecified. Contract mix may matter to commercial durability but filing does not state technology revenue is tied to either contract type. This establishes a buyer-segment lead, not a GTM channel success case. Next question: Can a supplier or PE-backed operator source identify which technologies are procured at which assets, who owns the operating budget, whether contract term changes the procurement case, and what deployment economics are shared after service pricing? - [Helmerich & Payne, Inc. 2025 Form 10-K](https://www.sec.gov/Archives/edgar/data/46765/000004676525000071/hp-20250930.htm) — Helmerich & Payne / U.S. SEC; 2025-11-20; fiscal year ended 2025-09-30; Lines 368–392 (customer mix, activity and technology generalities); 398–403 (PE-backed customer segment and largest North America customer concentration). Access: Primary SEC filing opened directly. ## McKinsey survey of oilfield services and equipment leaders — Oilfield-service firms are scaling gen AI more slowly than they expected T20-R19-LIMITS-01 · Data, people & purchasing · 2023 expectation survey compared with May 2025 reported adoption; article published 2026-04-01. McKinsey says the May 2025 OFSE executive survey included 164 executives from providers, EPCs, equipment makers and services firms. Article text says less than 25% of companies progressed beyond pilot (lines 18–20); a separate comparison says the 2023 survey expected >70% to be piloting/scaling by 2025, whereas the 2025 report says <25% showed some use and 1% significant scale (line 52). Because definitions/wave denominators are not given in inspected text, present this as a directional reported gap, not a matched-cohort adoption decline. $12–20bn is modeled annual EBITDA potential; 65% is hypothetical allocation intent, not spend. Status: May 2025 survey results published April 1, 2026; adoption/self-report and modeled sector potential, not specific deployments. Economics: $12–20bn annual EBITDA is McKinsey's potential opportunity estimate with no realized company-level denominator; 65% of survey leaders said they would prioritize digital infrastructure and AI if allocating hypothetical $1bn, which is stated preference, not approved budget. Baseline: Reported 2023 forecast and 2025 stage statistics are not shown with comparable wave Ns or matched definitions in article text; retain as a reported directional comparison only. AI role: Survey concerns generative AI among oilfield services and equipment (OFSE) providers; report's modeled opportunity spans generative and analytical AI in their current operations. Limits: Consulting estimate and self-reported survey; 164 executive sample is not a population census, the article text does not disclose response rate or 2023 wave N, and the modeled sector opportunity is not realized value. OFSE supplier data do not establish upstream-operator adoption. Next question: For an upstream buyer/provider GTM case, which named operating workflow and eligible population underlie the business case, what implementation and integration costs are included, and what achieved financial metric is audited after scale? - [Gen AI in the OFSE industry: Progress lags behind intent](https://www.mckinsey.com/industries/oil-and-gas/our-insights/gen-ai-in-the-ofse-industry-progress-lags-behind-intent) — McKinsey & Company; 2026-04-01; Lines 4–25 (scope, sample, adoption and potential); lines 45–63 (investment intent and scaling barriers). Access: directly opened primary report page. ## Oil Sands Alliance members / Canadian Natural Resources as proponent — Pathways CO2 transportation network and storage hub serving oil-sands facilities R1-LIM-03 · Water, emissions & energy · 2025-08 project overview; 2026-05-15 implementation agreement; 2026-07-02 Canada-Alberta announcement. All are within the window. Observable result is change from conditional proposal to a policy-backed agreement to advance, not technology operation. The 2026 agreement makes pipeline and Pathways construction mutually dependent and specifies carbon pricing, credits, contracts-for-difference, and other fiscal mechanisms; the federal-provincial parties state current CCUS ITC rates are extended to 2035. The old 2025 schedule was conditional on approvals and sufficient fiscal support. No realized capture or emissions result is evidenced. Status: As of July 2026, describe as governments and Oil Sands Alliance agreeing to advance construction, with policy/fiscal framework; not as FID, completed permitting, construction completion, injection, or operating capture. May agreement itself said the trilateral MOU remained to be concluded; July announcement says an agreement was reached. Economics: No project capex or cost-per-tonne confirmed in sources checked. Baseline: Not disclosed AI role: No AI role. Potential AI relevance is secondary: measurement, monitoring, verification, maintenance and capture-plant operations after facilities exist; not evidence of AI demand or deployment today. Limits: The latest announcement is political/industry agreement, not FID, completed permitting, EPC contract, or measured CCS operation. The 2025 overview is proponent-prepared and frames project aims. The 2026 6 Mtpa minimum in-service-by-2035 target cannot be treated as achieved. No capital cost or cost per tonne was validated in the opened primary sources. July 2026 agreement ties the CCS project to a proposed oil export pipeline; commercial and policy dependency remains material. Next question: What is sanctioned project capex, capture-by-operator spend, $/tonne delivered, financing/tax-credit dependence and pipeline/permits critical path? - [Pathways Alliance: Project Overview](https://oilsandsalliance.ca/wp-content/uploads/2025/08/Project-Overview.pdf) — Pathways Alliance / Canadian Natural on behalf of Alliance members; 2025-08; Project is proposed; each operator manages facility capture; pipeline critical path; schedule conditional on regulatory/policy/fiscal certainty; source PDF pages 2 and 18.. Access: Opened; proponent source. - [Implementation Agreement for the Canada-Alberta Memorandum of Understanding of November 27, 2025](https://www.pm.gc.ca/en/news/backgrounders/2026/05/15/implementation-agreement-canada-alberta-memorandum-understanding) — Prime Minister of Canada / Government of Canada and Alberta agreement; 2026-05-15; Pathways sections 3.1–3.8: alliance participants; mutual dependence on oil pipeline; 6 Mtpa CCUS minimum by 2035; fiscal support and tax credit commitments.. Access: Opened; government primary document. - [Canada and Alberta advance west coast pipeline project proposal and Pathways Project Carbon Capture Initiative](https://www.pm.gc.ca/en/news/news-releases/2026/07/02/canada-and-alberta-advance-west-coast-pipeline-project-proposal-and) — Prime Minister of Canada / Government of Canada; 2026-07-02; Announcement that the governments and Oil Sands Alliance reached an agreement to advance construction; statement of combined 16 Mtpa objective.. Access: Opened; government primary source. - [Source reopening in Round 2](https://oilsandsalliance.ca/wp-content/uploads/2025/08/Project-Overview.pdf) — See original source; 2025-08; Project overview pp.1,6,18: proposed shared transport/storage network; schedule conditional on approvals/policy.. Access: issuer/alliance PDF opened in prior round; proponent-prepared. - [Source reopening in Round 2](https://www.pm.gc.ca/en/news/backgrounders/2026/05/15/implementation-agreement-canada-alberta-memorandum-understanding) — See original source; 2026-05-15; Sections 3.1–3.7, esp. 3.3–3.5: commitments, dependencies and emissions milestones.. Access: government HTML opened. - [Source reopening in Round 2](https://www.pm.gc.ca/en/news/news-releases/2026/07/02/canada-and-alberta-advance-west-coast-pipeline-project-proposal-and) — See original source; 2026-07-02; Announcement of agreement to advance Pathways; pipeline proposal referred to Major Projects Office.. Access: government HTML opened. ## PETRONAS Malaysia Petroleum Management (MPM) — Malaysia upstream introduces normally unattended facilities and remote/autonomous operations guideline T20-R17-GLOBAL-01 · Production & process operations · Published 2025-06-20; AuDi 2025 event held 2025-04-21. Forecast horizon to 2033. MPM introduced a Normally Unattended Facilities with Remote & Autonomous Operations guideline, describing potential to unlock marginal fields, extend asset life and improve safety. It states CAPEX and OPEX cost cutting of up to 25% as potential and links the initiative to MPM's aspiration of sustaining 2,000 kboe/d. It anticipates 30% of field operations will have such capabilities by 2033. These are stated potentials/forecast, not actual outcomes or evidence of 2,000 kboe/d causal support. Status: Guideline introduced in 2025; MPM forecasts that 30% of field operations will feature NUF-RAO capabilities by 2033. No implementation date, qualification criteria, commissioned facility or rollout cohort is disclosed. Economics: No realized CAPEX/OPEX savings, denominator, project cost, incremental production, emissions change or facility-level economics disclosed. Up to 25% is a potential ceiling, not an observed result. Baseline: No baseline operating model, staffing, visits, uptime, operating cost, emissions, or reference field cohort is supplied. The 2,000 kboe/d figure is an MPM sector aspiration, not the guideline's measured contribution. AI role: Guideline refers to remote and autonomous operations; it does not specify an AI method, model, control authority, vendor, or use of AI. Do not classify this as an AI deployment. Limits: Issuer/regulator-like sector management release describes direction and intended benefit. The source does not define technical requirements, minimum instrumentation, communications reliability, safety-integrity boundaries, human intervention model, or evidence required for classification. Treat as policy/operating-model context and a follow-up lead, not as an operating deployment or AI case. Next question: Which fields have been qualified and commissioned under this guideline, and what matched facility-level evidence supports savings, uptime, safety and emissions claims after accounting for communications, maintenance and emergency-response costs? - [Executive Annual Dialogue (AuDi) 2025: Driving Project Excellence, Technology, Innovation, and Collaboration in Malaysia’s Oil & Gas Sector](https://www.petronas.com/mpm/media/media-releases/executive-annual-dialogue-audi-2025-driving-project-excellence-technology) — Malaysia Petroleum Management, PETRONAS; 2025-06-20; Lines 62-75: event and publication dates, guideline, potential benefits and 2033 forecast.. Access: Issuer HTML opened directly; full text available. ## Petrobras / Libra consortium — Mero Permanent Reservoir Monitoring (PRM) seabed fiber-optic seismic network T10-R01-GLOBAL-01 · Subsurface & recovery · Phase-one installation completed by March 2026; reporting source dated 2026-04-14. Phase one was completed in March 2026 (>460 km over 222 km²), with first data collection scheduled for Q2 2026; phase two (>316 km over 140 km²) was under construction for completion in 2027. The approximately US$450m is total project investment by Petrobras and Libra partners—not savings, ROI, or a proven realized PRM-only cost-benefit. Mero field production above 680 kb/d in Jan 2026 is contextual, not PRM-attributed. AI with UFRJ is a stated data-use layer, distinct from physical sensing. Status: Phase one installation completed March 2026; first data collection was stated for Q2 2026. Phase two construction was ongoing in April 2026 and was expected to complete in 2027. The source does not confirm first-data receipt or reservoir-management result. Economics: Issuer-reported total project investment of about US$450 million. No phase-specific cost, operating cost, matched production baseline, realized incremental recovery or cash benefit is given. Petrobras cites Mero average monthly production above 680,000 b/d in January 2026 as field context, not a PRM-attributed result. Baseline: No pre-PRM controlled comparison or production/recovery outcome; initial data collection only planned for Q2 2026 in the dated source. AI role: The physical system is a seabed network of optical sensors/cables acquiring seismic data over time. Petrobras separately says it will use AI with UFRJ to capture information continuously from PRM; the model, task boundary, operating state and measured incremental AI value are not specified. Do not label the sensor network itself as AI. Limits: The scale and asset scope are unusually concrete, but performance is prospective. Company calls it the world's most extensive seismic-monitoring project; do not treat superlative as independently benchmarked. AI role is planned/described as continuous information capture with UFRJ, not an attributed production benefit. HISEP is a separate Mero subsea process technology and remains separately listed as GAP-HISEP-01 in the previous round. Next question: After first data arrive, does PRM change reservoir-management decisions enough to improve production/recovery or reduce intervention risk, and which measured benefit belongs to sensing, interpretation, or AI? What are full installation, maintenance and data-operations costs across the four FPSO areas? - [Petrobras investirá US$450 milhões no maior projeto de monitoramento sísmico mundial](https://agencia.petrobras.com.br/w/negocio/petrobras-investir%C3%A1-us-450-milh%C3%B5es-no-maior-projeto-de-monitoramento-s%C3%ADsmico-mundial) — Agência Petrobras / Petrobras; 2026-04-14; Lines 45-50 and 61-72; phase-one completion, cable lengths/areas, asset coverage, project investment, data timing, AI/UFRJ statement.. Access: Official page opened; relevant text directly reviewed (Portuguese).. - [Annual Report and Form 20-F 2025](https://petrobras.com.br/documents/2677942/17808296/FORM%2B20F%2B2025%2BEN.pdf/3b6371b3-900d-ae05-294d-f9acd9d8b26e?download=true&t=1777312960000&version=1.0) — Petrobras; 2026 (2025 filing); Printed p.83, PRM paragraph; first seismic acquisition scheduled for 2026; network connected to Sepetiba and Marechal Duque de Caxias.. Access: Official PDF indexed text reviewed; direct PDF open returned internal error.. ## Petrobras / Libra consortium — Petrobras plans its Mero 3 HISEP pilot for 2028 T10-R01-LIMITS-01 · Production & process operations · Mero 3 FPSO first production began 2024-10-30. HISEP pilot connection/operation is scheduled for 2028 per Petrobras 2025 Form 20-F and release/planning materials. HISEP is a high-pressure subsea separation concept designed to separate high-CO2 gas at the seabed and reinject it, relieving topside gas-processing bottlenecks; benefits remain intended/projected. Petrobras primary project-specific sources identify HISEP as planned at Mero 3 from 2028. This conflicts with generic current English Petrobras page saying a growing number of FPSOs already use “CCUS and Hisep.” The source evidence does not establish broad HISEP deployment; likely conflation with existing CO2 reinjection/CCUS across FPSOs. Status: Mero 3 FPSO began production 2024-10-30; specific Petrobras release says HISEP will be in operation from 2028. Keep HISEP as a future pilot, not broad current deployment. The 2025 20-F/business-plan passages remain indexed-only in this audit; direct PDF access was not achieved. Economics: No HISEP-specific project cost, vendor contract value, realized production uplift, avoided topside capex, emissions outcome, or unit economics established in reviewed sources. Supplier relevance is prospective: subsea separator, pumps, controls, subsea interfaces and FPSO tie-in. Baseline: No comparable Mero operating case without HISEP, baseline GOR/throughput constraint, or before/after metric disclosed in reviewed materials. AI role: Physical subsea separation and reinjection; no AI role disclosed. Do not infer from data systems or general automation. Limits: Do not describe HISEP as broadly operating, already delivering emissions/production benefits, or failed. Direct PDF opening for 20-F and Business Plan failed; relevant official indexed passages were available and corroborated by an opened Petrobras newsroom release and generic official page. Preserve source-specific stage and date. Next question: Before treating HISEP as a scalable supplier or operating opportunity, what are the qualification gates, design envelope, offshore uptime, GOR/CO2 baseline, topside capacity displaced, reinjection energy, emissions boundary, capex, and date for witnessed commissioning at Mero 3? - [Annual Report and Form 20-F 2025](https://petrobras.com.br/documents/2677942/17808296/FORM%2B20F%2B2025%2BEN.pdf/3b6371b3-900d-ae05-294d-f9acd9d8b26e?download=true&t=1777312960000&version=1.0) — Petrobras; 2026 (reporting year 2025); Printed p.83, Mero Field / New Technologies: pilot will connect to FPSO Marechal Duque de Caxias in 2028. Access: Official source search-index text reviewed; direct open returned internal error. - [Petrobras 2026-2030 Business Plan (PN 2026-30)](https://canalfornecedor.petrobras.com.br/documents/d/canal-do-fornecedor/1-apresentacao_business-plan?download=true) — Petrobras; 2025-11 (plan edition; file/query metadata; exact PDF publication date not exposed in opened page); Indexed slide 62: “HISEP Tecnológico” at Mero 3 and first oil expected 2028. Access: Official plan search-index text reviewed; direct open failed due large PDF. - [Marechal Duque de Caxias FPSO starts producing in pre-salt](https://agencia.petrobras.com.br/pt/w/negocio/navio-plataforma-marechal-duque-de-caxias-comeca-a-produzir-no-pre-sal) — Petrobras / Agência Petrobras; 2024-10-30; Lines 81-85 per prior explicit saved primary-source check in root-source-checks.json; release says Mero 3 started oil production and HISEP planned from 2028. Access: Official issuer page and indexed source reopened; date/locator cross-checked with saved root-source-checks.json. - [Exploration and production: going deep to innovate](https://petrobras.com.br/en/quem-somos/exploracao-e-producao) — Petrobras; Undated current web page; retrieved 2026-10-07; Lines 168-171 and 239-242: broad claim that a growing number of FPSOs use CCUS and HISEP; HISEP glossary statement. Access: Official web page opened. ## Petrobras — Santos Basin pre-salt CO2 reinjection program scales to 24 FPSOs T10-R03-GLOBAL-01 · Water, emissions & energy · 2025 result; comparison with 2024; cumulative since program start (start year not stated in this source). Petrobras reports 19.6 Mt CO₂ reinjected in 2025 versus 14 Mt in 2024 and cumulative >80 Mt across systems on 24 FPSOs. This is portfolio-scale physical CCUS-EOR activity; not isolated incremental recovery, emissions avoided net of lifecycle, or technology ROI. Status: Operating fleet-wide program; 24 operating FPSOs equipped with capture and reinjection systems as reported for 2025. Economics: No per-field incremental production, cost per tonne, injection CAPEX/OPEX, or net recovery value disclosed in reviewed source. Baseline: Annual reinjection comparator: 14 million tonnes in 2024; 19.6 million tonnes in 2025. This is throughput growth, not an incremental oil-recovery baseline. AI role: AI contribution to the injection program or reported outcome is not disclosed in the reviewed sources; this is a physical/process technology finding, not an AI case. Limits: Issuer-reported aggregate only; no attribution of additional barrels to CO2 vs depletion, water injection, field mix, or other interventions. CO2 reinjection is also used to avoid venting reservoir CO2. Do not treat volume growth as proof of EOR economics or net emissions benefit. Next question: What field-level evidence and cost data would be required before treating high-throughput CO2 reinjection as a repeatable, economic recovery advantage rather than emissions management plus pressure support? - [Climate Change and Energy Transition Supplement 2025](https://petrobras.com.br/documents/2677942/9192383/Caderno%2Bdo%2BClima%2B2025%2Bingl%C3%AAs%2Bfinal_rev1.pdf/3809f8ca-4def-7df7-7e6e-0ee811d49a90?download=true&t=1779219386000&version=1.0) — Petrobras; 2026 publication; reports 2025 results; PDF pp. 100, 114 (lines 3614-3617, 3979-3987). Access: Direct PDF opened. - [CCUS](https://agencia.petrobras.com.br/pt/w/ccus) — Petrobras Agência; Undated current issuer explainer; accessed 2026-10-07; Section on CO2 capture and reinjection; indexed issuer page. Access: Issuer page indexed; direct text not opened. ## Petrobras — Ativo 360 and operational-data integration described in 2025 Form 20-F T10-R06-GLOBAL-02 · Data, people & purchasing · Fiscal 2025, SEC filing date 2026-04-09. Ativo 360 is described in the 2025 Form 20-F as an operational-data integration platform associated with predictive insights/recommendations. The filing separately describes AI use cases but does not establish that all Ativo 360 functions are AI-based or quantify platform outcomes. Status: Described as a platform in use; extent and business-unit deployment not stated. Economics: No Ativo 360 investment, license/integration cost, deployment count, uptime/maintenance effect, or realized value disclosed in excerpt. Baseline: No adoption denominator or before/after comparison stated. AI role: The filing explicitly positions Ativo 360 alongside predictive insights/recommendations and separately describes AI usage. Do not count the platform itself as a single AI deployment or infer which functions are AI-based absent use-case detail. Limits: SEC filing is an issuer representation, but platform-level scope, underlying product architecture, specific assets, deployment date and division between conventional analytics and AI are unspecified. Treat as context, not measurable performance. Next question: Can the company disclose the platform's deployment footprint, AI/non-AI boundary, workflow owner and audited operating outcomes without treating a shared data layer as value in itself? - [Petrobras Annual Report and Form 20-F 2025](https://www.sec.gov/Archives/edgar/data/1119639/000129281426002168/pbrform20f_2025.htm) — Petrobras / U.S. Securities and Exchange Commission; Filed 2026-04-09; Business section, Research, Development and Innovation; SEC-indexed filing excerpt (page 149 of annual report). Access: SEC filing and accession index searched/opened; filing content returned as indexed primary excerpt. ## Petrobras — Business Plan 2025–29 publishes prospective equilibrium price and total-cost assumptions T10-R09-GLOBAL-03 · Data, people & purchasing · Plan horizon 2025–2029; announced Nov 2024 (source accessed in 2026). Petrobras Business Plan 2025–29 (published 21 Nov 2024) stated prospective average Brent equilibrium of $28/bbl and forecast average total cost of produced oil $36.5/boe, including lifting, government participation, depreciation and depletion. These are plan assumptions/targets, not realized current costs or technology outcomes. Status: Prospective plan assumptions, not realized results. Economics: Planning denominator only, no realized production/cost outcome and no technology-specific break-even. $28/bbl “equilibrium Brent” and $36.5/boe “total cost of produced oil” are differently defined measures and should not be subtracted or treated as a simple breakeven price without further reconciliation. Baseline: Prospective plan average for 2025–2029; not a historical cost baseline. E&P planned CAPEX $77.3bn over five years is a portfolio allocation, not technology-specific spend. AI role: Plan describes technology-enabled new production systems generically; no specific technology return is included in these denominators. Limits: Issuer's public plan summary does not expose field-level assumptions, fiscal treatment by asset, cost distribution, scenario sensitivities, or technology attribution. Terms differ from Aramco's lifting/capex unit-cost averages; do not make direct operator ranking from these numbers. Next question: What asset-level cost deck, fiscal terms, oil-price sensitivities and recovered-barrel assumptions allow a technology proposal to be assessed against the portfolio's published planning hurdle? - [Petrobras launches 2025–2029 Business Plan with investments of US$111 billion](https://agencia.petrobras.com.br/en/w/negocio/petrobras-lanca-plano-de-negocios-2025-2029-com-investimentos-de-us-111-bilhoes) — Petrobras Agência; 2024-11-21 (published before window; plan runs 2025–2029); Lines 63-80. Access: Direct issuer page opened. ## Precision Drilling — Alpha digital twin detects Haynesville mud-pump motor/bearing issue before maintenance window T10-R02-LIMITS-03 · Maintenance & reliability · Timeline in case study shows alerts on 2025-08-31; copyright 2025. Precision reports continuous mud-pump equipment monitoring detected abnormal vibration; an alert led the crew to stage a replacement motor and replace it during tripping. Maintenance confirmed excessive bearing end play. Precision says the action limited downtime hours by more than 50% and avoided what could have been a significant NPT event. The case establishes a specific alert-to-work-order response and confirmed component condition, not an observed catastrophic failure prevented. Status: Single rig event case published in 2025; deployed monitoring/alert and planned maintenance action. This supplements Alpha fleet coverage, not a new platform rollout. Economics: No actual hours, cost, avoided lost production or repair cost disclosed. >50% refers to downtime hours “limited” without defining denominator or counterfactual event duration. Baseline: Hypothetical unplanned failure/NPT event; no measured no-alert baseline or comparable rig event. AI role: Precision explicitly says its Digital Twin Solution uses machine-learning models to flag anomalies. This is an ML-enabled maintenance workflow, not evidence of autonomous equipment control. Limits: Vendor case study; operator and rig ID withheld. One selected success does not establish fleet false-alert rate, missed failures, model drift, incremental sensor cost or fleet-average availability impact. Bearing end play was confirmed; a full motor failure was not documented. Next question: What are the fleet-level lead-time distribution, precision/recall, false-alert and missed-failure rates, actual repair/NPT hours, sensor and support cost, and customer corroboration across the 78-rig reported footprint? - [Leveraging Predictive Analytics to Optimize Mud Pump Reliability with Our Digital Twin Solution](https://www.precisiondrilling.com/wp-content/uploads/2025/09/Precision-Drilling_Digital-Twin-Case-Study_Sept-2025-1.pdf) — Precision Drilling; 2025 (event 2025-08-31; copyright ©2025; exact publication day not printed); Printed pp.1-2; abnormal vibration, motor change during tripping, >50% downtime claim; alert timeline 2025-08-31. Access: Official Precision PDF opened directly. ## Quantum Capital Group — Quantum describes a shared data and AI approach; asset-level results remain unclear T20-R15-LIMITS-01 · Data, people & purchasing · Undated current page; not date-qualified to the research window. Quantum describes a Microsoft/Databricks-supported data platform and QIQ and lists specific upstream workflows across deal evaluation, acreage planning, production optimization and risk/compliance. These claims indicate the sponsor's stated digital operating model and possible buyer-side vocabulary. The page gives no implementation dates, participating portfolio-company names, user/well counts, spend, counterfactual or realized performance, so it cannot substantiate operational adoption or ROI. Status: First-party sponsor-described operating platform and workflow claims; source is undated, accessed 2026-10-07. No contract or portfolio-company implementation evidence on page. Economics: No technology budget, vendor contract value, implementation cost, savings or return disclosed. Baseline: No baseline, outcome denominator or external comparator disclosed. AI role: Sponsor says its Quantum Energy Cloud unifies portfolio data and that proprietary Quantum IQ generative AI supports information synthesis and workflow automation; operational models are described for completion design, artificial lift, base-production anomaly detection and downtime prediction. Limits: First-party marketing page; undated; sponsor-level rather than portfolio-company source; precise claims may describe aspiration or internal use at unknown scale. Do not count as a named operator deployment or treat PE sponsor as a single procurement decision-maker across portfolio companies. Next question: Can a supplier identify the operating-company decision owner, one named portfolio asset, spend authority, data access terms and a baseline-backed pilot case before treating the sponsor platform as an addressable rollout? - [Digital at Quantum](https://www.quantumcap.com/digital-mission/) — Quantum Capital Group; Undated; accessed 2026-10-07; Lines 53–77 (Quantum Energy Cloud); 82–109 (QIQ, operational workflow claims and outcome language). Access: directly opened first-party page. ## Rio Tinto — Autonomous haulage system provides a mature physical-automation scale and measurement analog T20-R19-GLOBAL-01 · Maintenance & reliability · Current fleet scale is undated on current issuer page. Per-truck estimate explicitly refers to 2018 operations. Rio Tinto says its Autonomous Haulage System trucks are managed by a supervisory system and central controller, follow predefined GPS courses, and track vehicle locations/speeds/directions. The company reports more than 130 trucks operating. It estimates that in 2018 each autonomous truck operated on average 700 hours more than a conventional truck and had 15% lower costs. The page does not provide the truck population, matched comparator, cost components, confidence range, or current replication of the 2018 estimate. Status: Mature operating fleet; current issuer page says more than 130 trucks are in operation. Cost/productivity estimate dates to 2018 and is not a current-period measured result. Economics: The 15% cost reduction is Rio Tinto's estimated 2018 per-truck comparison; no current period, absolute unit cost, implementation capex, fleet lifecycle cost, or independent validation is given. It cannot be transferred numerically to oil and gas, where wellsite, offshore and process-control constraints differ. Baseline: Company comparator is conventional haul trucks, with an estimated average 700 extra operating hours per autonomous truck in 2018. The exact conventional fleet, matching, denominator and period are not stated on the page. AI role: The cited haulage method is autonomous vehicle navigation/control using a supervisory system, central controller and GPS routes. Rio Tinto does not identify this as an AI-specific result. Do not count as AI evidence. Limits: Mining haulage has fixed routes, controlled operating areas and repetitive vehicle tasks; upstream wells, offshore facilities and subsurface decisions have different safety cases and workflow variability. Useful as a physical-automation qualification example and an evidence template (hours, unit costs, safety exposure), not as a forecast or ROI benchmark for oil and gas. Next question: For an upstream physical-automation case, can the operator disclose actual task hours, matched manual baseline, safety exposure, availability, lifecycle cost and repeatability across assets at a comparable level of clarity? - [Automation: Pioneering automation and robotics in mining](https://www.riotinto.com/en/mn/about/innovation/automation) — Rio Tinto; Current issuer page, no publication date shown; Lines 127-160; truck mechanism, fleet count and historical 2018 estimate at lines 148-151.. Access: Issuer HTML opened directly; full relevant text available. ## SLB / OQ Exploration & Production — Oman Neuro / PowerDrive well-construction results with named operator T10-R02-LIMITS-01 · Drilling & completions · SLB reported the results in its 2026-01-23 Q4/full-year 2025 earnings release; exact well dates are not stated. Keep three RSS wells at up to 50% ROP and one separate PowerDrive One + MWD well at 40% ROP with reduced tripping and casing time. Attribute only as SLB's report; no average, comparator, total-well/cost effect or exact well dates. Retina imaging is present, but no separate result is assigned to it. Status: Vendor-reported field use on four wells: three PowerDrive RSS wells and a separate PowerDrive One + MWD well. Economics: No absolute drilling time, total-well duration, dollars per well, contract value or incremental tool cost disclosed. ROP increase is drilling rate for selected well sections, not a 40-50% total cost reduction. Baseline: Company uses a qualitative performance comparison; no reference wells, interval lengths, geology match, averaging, or denominator beyond the four named deployment counts is disclosed in the filing. AI role: SLB says Neuro autonomous directional drilling and AI optimization were used. Physical equipment results are tied to PowerDrive RSS and MWD; do not assign the equipment-only effect to AI independently. Limits: Vendor-reported; no operator-side source or well-level data found in this pass. “Up to” is a maximum, not mean. Cannot isolate AI contribution from tool design, sensor quality, geology, BHA, execution or section selection. Do not recast Jan. 2026 reporting as launch date. Next question: Can OQ validate section-level offset comparisons and connect the tool/AI result to total well time, casing quality, cost per meter, and incremental service/tool cost across a repeated campaign? - [SLB Announces Fourth-Quarter and Full-Year 2025 Results](https://investorcenter.slb.com/static-files/f65b65ce-3607-438d-a980-cb2d1af044ae) — SLB; 2026-01-23; Printed p.18 (PDF p.18), lines 683-690, Oman/OQ paragraph. Access: Official investor PDF opened directly. ## SLB / Zubair Field operator (not named) — PCL digital twin and Lift IQ remote ESP power optimization T10-R03-LIMITS-01 · Production & process operations · Specific adjustment dates shown for one well (May 3 and May 28; year not stated). Case page publication date and overall measurement period not stated. SLB reports model-assisted remote voltage changes on two Zubair ESP wells and lower kVA by 28%/34% without production loss; expected run-life/TCO benefits remain projected. Test infrastructure limits to roughly annual well tests; model's independent flow validation error and project economics are not reported. Status: Field trial followed by operator decision to deploy the service to Zubair wells equipped with SLB ESPs; reported quantitative results are only for two analyzed wells. Economics: No monetary savings, power-cost baseline, implementation cost or measured run-life extension is reported. Longer ESP run life and lower total cost of ownership are stated as expected benefits, not measured outcomes. Baseline: Conventional tests were logistically constrained to about one per year per well. For one example well, SLB says a 25% motor-voltage reduction was estimated to reduce current from 38 A to 24 A without affecting production; change was applied in two steps and monitored. AI role: No AI claim in the opened source; hydraulic/electric digital twin plus engineer oversight. Limits: Evidence is an undated supplier-authored case. The measured power results cover two wells, not all Zubair wells. Model-derived real-time production estimates substitute for sparse tests; source does not disclose validation error against a contemporaneous independent flow measurement. Do not imply the expected run-life/TCO benefit was realized or that the intervention was AI. Next question: For operators with sparse well testing, what independent flow-rate validation and per-well energy/cost data are required before model-guided ESP changes can be underwritten at fleet scale? - [Remote ESP optimization reduces power consumption by 28% and 34% in two Iraq wells](https://www.slb.com/resource-library/case-study-with-navigation/al/pcl-and-lift-iq-iraq) — SLB; Undated page; accessed 2026-10-07; Lines 1891–1913 (asset and annual test constraint); 1916–1929 (model, two-well results, expected TCO); 1934 (two-step adjustment dates). Access: Primary page opened directly. ## SLB / unnamed South American operator — Lift IQ remote surveillance and artificial-lift intervention T10-R03-LIMITS-02 · Maintenance & reliability · 29 months; start/end calendar dates undisclosed. The page body says 1,195 VSD adjustments were remote (23% of total adjustments) over 29 months, with 26,141 fewer crew miles and at least 497 work-hours saved. Its headline/summary also frames 1,195 as avoided wellsite visits and 23% as fewer field visits; avoid equating remote adjustment count with avoided-trip count without clarification. No production/cost result; undated vendor page. Status: Operating remote-surveillance service; the case reports a 29-month period, but the page date is not stated. Economics: No production uplift, deferred-barrel reduction, service fee, cost avoidance or ROI is quantified. Work-hour and mileage results are operational measures, not a complete economic return. Baseline: Field had seven active rigs and >230 artificially lifted wells, difficult downhole conditions (including gas and solids), average interwell spacing 22 miles, and staff driving >87 miles per 12-hour shift. SLB describes shutdowns, restart delays and deferred production as the prior problem, without quantifying barrels. AI role: No AI disclosed. The source attributes intervention to experienced surveillance engineers using acquisition/transmission systems and control hardware; engineers remotely changed variable-speed-drive settings. Limits: Undated vendor case with unnamed customer. Counts and travel/work-hour savings are provider-reported; no operator corroboration or production/economic denominator is supplied. 23% refers to adjustments remotely implemented, not an asserted 23% reduction in visits, downtime or cost. The case shows remote human intervention and connectivity rather than AI or autonomous control. Next question: Can the operator link the travel and work-hour reductions to independently measured downtime, deferred production and total service cost before estimating economic value? - [Production life cycle management service eliminates 1,195 wellsite visits in South American field, saving a total of 26,141 mi](https://www.slb.com/resource-library/case-study-with-navigation/al/lift-iq-south-america-cs) — SLB; Undated page; accessed 2026-10-07; Lines 1891–1912 (scope and operating constraints); 1916–1921 (service mechanics and 29-month measured outcomes). Access: Primary page opened directly. ## Seplat Energy — Jisike associated-gas compression supplies gas-lift to producing OML 53 wells T20-R14-GLOBAL-02 · Production & process operations · Jisike station since 2025-08; source page currently available but has no explicit publication/update date. Seplat's 2025 results provide block production/operating context but do not isolate gas-lift response. Seplat says the Jisike station uses produced gas from the field for gas-lift in producing oil wells. The field also has a flow station with 15,000 bopd design capacity and the associated-gas compressor has 6 MMscfd design capacity. A separate Ohaji South compressor is projected to recover up to 10 MMscfd of associated gas for ANOH feedstock, but was only at mechanical completion in the current page; that target is not a Jisike result. Status: Station reported operational since August 2025. Seplat's current asset page says produced field gas is used for gas-lift of oil-producing wells. OML 53 Ohaji South AG compressor is a separate project at mechanical completion and must not be described as operating. Economics: No actual compressor throughput, incremental liquids per lifted well, avoided flaring volumes, fuel/power cost, uptime or payback is disclosed. Design capacity must not be represented as actual gas-lift volume. Baseline: No pre-gas-lift well production, pressure or downtime baseline is reported. The asset page gives block ownership and design capacities but no attributable operating delta. AI role: No AI is attributed. The reported mechanism is compression of produced associated gas for gas-lift of oil wells; digital/control methods are not described. Limits: The issuer page is undated, though it states operation since August 2025. Gas-lift is established production engineering; the useful evidence is asset-specific implementation and produced-gas source, not novelty. No AI or automation claim. Do not combine Jisike with not-yet-operating Ohaji South gas recovery project or ANOH ramp-up metrics. Next question: At what actual gas rate and well population is Jisike gas-lift operating, and what incremental oil, decline/downtime change, avoided flaring, compressor power/fuel cost, uptime and netback effect has the operator measured? - [OML 53 | Upstream | Our operations](https://www.seplatenergy.com/our-company/our-operations/upstream/oml-53/) — Seplat Energy; Undated current asset page; stated Jisike operating date August 2025; Oil Operations and AG Flares Out sections, lines 91-109; operator/working-interest context lines 116-130.. Access: Issuer HTML opened directly. - [FY 2025 Results Investor Presentation](https://www.seplatenergy.com/media/qo5pqfso/fy-2025-results-investor-presentation_260226-1.pdf) — Seplat Energy; 2026-02-26; Upstream performance and idle-well programme, PDF pp. 6 and 10; operating context and separate 49-well restoration metric.. Access: Issuer-hosted PDF opened directly. ## Seplat Energy / ANOH Gas Processing Company (Seplat and NGIC JV) — ANOH gas-processing and export chain starts supply, enabling four upstream wells and condensate production T20-R15-GLOBAL-02 · Production & process operations · First-gas release dated 2026-01-20; actual flows from 2026-01-16 onward. Early stabilized wet-gas production was 40–52 MMscfd and condensate production 2.0–2.5 kboepd. Seplat says four upstream wells that had been on standby since November 2025 were brought online to enable gas supply. The plant delivered 40–52 MMscfd processed gas directly to the Indorama petrochemical plant and produced 2.0–2.5 kboepd condensate while ramping. Design includes two 150 MMscfd trains, LPG recovery, condensate stabilization and a 16 MW power plant; designed for zero routine flares. Design capacity of 300 MMscfd is not actual throughput. Status: First gas/supply began 2026-01-16 after completion of 11 km Indorama export pipeline and regulator approval. The plant is ramping up; NLNG supply was not yet started in the cited release, and the OB3 export route remained under construction. Economics: No plant capex, gas-processing toll, well-level incremental production/cash flow, utilization, downstream sales price or return is disclosed. Seplat states the project creates distinct upstream gas-sales and JV-dividend income streams, but gives no amount in this release. Baseline: Four wells were on standby since Nov 2025; first gas enabled them to be brought online. No well-level pre/post production or counterfactual gas evacuation alternative is provided. Early plant throughput is far below 300 MMscfd design capacity. AI role: No AI or digital-control role disclosed. The reported deployment is a 300 MMscfd gas-processing/export system with two 150 MMscfd trains, LPG/condensate recovery and stabilization, and a 16 MW power plant. Limits: This is a large midstream process/logistics adjacency and infrastructure commissioning, not proof of a proprietary novel technology. Its upstream relevance is the practical dependency of gas development on processing and evacuation capacity. Avoid attributing all condensate/gas output solely to plant technology; reservoir production, pipeline, offtake, wells and ramp-up all matter. Next question: As ANOH ramps, what throughput/utilization, upstream well availability and netback, condensate/LPG yield, plant energy intensity, zero-routine-flare performance, and capital/processing economics are sustained? - [ANOH Gas Project Achieves First Gas](https://www.seplatenergy.com/news-insights/news/anoh-gas-project-achieves-first-gas/) — Seplat Energy; 2026-01-20; First-gas, throughput, facility configuration, upstream-well, pipeline and status details, lines 82-104.. Access: Issuer HTML opened directly. ## Shell — Gato do Mato / Orca FPSO and subsea development redesign R1-GL-03 · Production & process operations · Shell story dated 2025-04-02; FID March 2025. Shell 2025 Annual Report reports later execution status and Orca naming. The +20% capacity, nearly -16% project cost, -30% topsides weight and -20% GHG-intensity figures are design estimates versus Gato 2.0, not delivered operating results. The project reached FID in March 2025 and is in execution; installation was planned for 2028. Status: Redesigned concept reached FID in March 2025; project moved into execution. FPSO construction and offshore installation are future work; production expected 2029. Economics: Percentage estimate only; source does not disclose absolute cost, estimate class or scope normalization. Baseline: Prior Gato 2.0 design. The source provides percentage deltas but no absolute project cost, scope normalization, estimate class, or independent assurance. AI role: None disclosed. Limits: Projected capacity, cost, weight, and emissions changes depend on execution and production ramp-up. A single project does not validate general repeatability or prove the redesign caused future realized economics. Preserve ownership changes and expected start date as time-sensitive. Next question: What approved estimate class, absolute capex and normalized scope underpin the claimed ~16% cost reduction versus Gato 2.0, and what remains after FID? - [What does it take to design a vast and complex offshore oil and gas facility?](https://www.shell.com/news-and-insights/our-stories/what-does-it-take-to-design-a-vast-and-complex-offshore-oil-and-gas-facility.html) — Shell; 2025-04-02; Gato 3.0 redesign, co-design with Modec, and reported design deltas. Access: Issuer page opened; full text accessible. - [Shell Annual Report and Accounts 2025, Strategic Report](https://www.shell.com/investors/results-and-reporting/annual-report/_jcr_content/root/main/section_2113846431/link_list/links/item1.stream/1773292491081/1bb8ca77d4189274d0abd1c40bf1defe1f7b4a49/strategic-report-ar25.pdf) — Shell plc; 2026-03 (report publication month; exact day not captured); Printed p. 43, project renamed Orca; FID/execution/2029 expected start; ownership. Access: Issuer-hosted PDF search result accessible; direct open failed in this research call. - [Source reopening in Round 2](https://www.shell.com/news-and-insights/our-stories/what-does-it-take-to-design-a-vast-and-complex-offshore-oil-and-gas-facility.html) — See original source; 2025-04-02; Gato 3.0 redesign, projected design comparisons, FID/execution and 2028 installation plan. Access: Shell issuer page opened; text accessible. ## Shell — Sparta combines 20K equipment with managed-pressure drilling R1-GL-04 · Drilling & completions · Official story published June 3, 2026. No operating result. Up to 90,000 boe/d is planned peak production, not achieved production; no quantified cost delta is provided. Status: Under construction; first production expected 2028. Economics: Project capex, incremental cost of 20K equipment, electric compression premium, and expected breakeven/capital efficiency not stated in source. Baseline: Shell compares pressure capacity with most deepwater platforms and cites Appomattox as earlier 15,000-psi experience; no same-reservoir alternative or economic counterfactual disclosed. AI role: No AI role disclosed; pressure monitoring/control is not described as ML. Limits: Novel pressure envelope, salt drilling, equipment qualification, well control, construction and startup remain execution risks. Expected output is not proof of reserves or profitability. Need actual reliability and failure-mode data after commissioning. Next question: What incremental capex, cost per recoverable barrel, pressure-rated equipment qualification cost, and reliability penalty support the 20K design versus conventional development? - [Sparta: unlocking energy once thought out of reach](https://www.shell.com/news-and-insights/our-stories/sparta-unlocking-energy-once-thought-out-of-reach.html) — Shell; 2026 (web story; exact publication day not exposed in opened page); Design pressure and output lines 34-43; MPD and high-density fluid lines 63-77. Access: Issuer page opened; full text accessible. - [Source reopening in Round 2](https://www.shell.com/news-and-insights/our-stories/sparta-unlocking-energy-once-thought-out-of-reach.html) — See original source; 2026-06-03; Construction and planned 2028 start; 20,000-psi design, up-to-90,000 boe/d expected peak, managed-pressure drilling description. Access: Shell issuer page opened; text accessible. ## Texas Pacific Land Corporation / TPWR and Transmissive Water Services — Produced-water desalination test facility: 2025 pause followed by commissioning in 2026 T10-R04-LIMITS-01 · Water, emissions & energy · Development over several years; cumulative spend through 2025-12-31. Filing published 2026-02-25. TPL reported a temporary 2025 construction pause for equipment testing and potential integration. Its Q2 2026 filing says construction was complete and commissioning had commenced by June 30, 2026. Initial design capacity is 10,000 bbl/day; throughput and commercial economics remain unreported. Status: The 2025 pause is historical: TPL Q2 2026 says construction complete and commissioning commenced as of 2026-06-30. Do not describe facility as currently paused or as commissioned/operating. Cumulative spend moved from $45.5m to $55.8m; Q2 filing gives $48.1m capitalized. Economics: Cumulative process/equipment spend was $45.5m through 2025 and $55.8m through June 30, 2026, including $48.1m capitalized. This is cumulative development spend, not a standalone plant capital cost or unit-treatment cost. Baseline: TPL estimates the Permian produces >20 million bbl/d of produced water and describes injection as the incumbent disposal route. TPL's broader water segment is a separate, existing business; the desalination equipment itself is not yet identified as in commercial operation. AI role: No AI role disclosed; this is a physical treatment process and equipment development program. Limits: Do not call the program a failed deployment: the company characterizes construction as temporarily paused for testing/equipment integration. Do not treat lab success, stated 10,000-bbl/d design capacity or H1 2026 target as facility commissioning or delivered water volumes. No current post-H1-2026 status was established in this pass; no treatment energy intensity, recovery rate, product quality test results, disposal cost avoided, capex forecast or customer economics disclosed. Next question: What commissioning date, achieved throughput and quality, energy per barrel, full installed cost, and binding operator offtake would turn this development spend into a repeatable upstream alternative to disposal? - [Texas Pacific Land Corporation 2025 Form 10-K](https://www.sec.gov/Archives/edgar/data/1811074/000181107426000018/tpl-20251231.htm) — Texas Pacific Land Corporation / U.S. SEC; Filed 2026-02-25; reports through 2025-12-31; Lines 324–329 (process, status and spend); 1210–1213 (water service costs and R&D expense). Access: Primary SEC-hosted filing opened directly. ## Texas Pacific Land Corporation / TPWR and Transmissive Water Services — Produced-water desalination Phase 2B moves from construction pause to commissioning T10-R08-LIMITS-01 · Water, emissions & energy · Status as of 2026-06-30, in Q2 2026 Form 10-Q; filing date 2026-08-04. TPL's issuer filing reports construction complete and commissioning started for its 10,000-bbl/d initial-capacity facility. Cumulative spend through 2026-06-30 was $55.8m, of which $48.1m was capitalized; $10.2m was spent in H1 2026. Status: Construction complete and commissioning commenced as of 2026-06-30; not commissioning complete/commercial operation. Economics: Cumulative development spend increased $10.3m from the $45.5m disclosed through 2025-12-31. Spend/capitalized balance are not unit treatment cost, annualized operating cost or customer value. No price, offtake contract, revenue, cost per barrel, or avoided-disposal cost disclosed. Baseline: Prior status in TPL 2025 Form 10-K (through 2025-12-31): facility construction temporarily paused during 2025 for equipment testing and possible additions; H1 2026 completion was anticipated. New Q2 filing advances status to constructed/commissioning commenced. AI role: No AI disclosed; physical treatment process and equipment. Limits: Do not call the earlier pause a failed technology deployment: the stated reason was further testing/equipment integration, and the later filing says construction completed. Do not call commissioning successful or infer production throughput; current source says only commissioning commenced. No effluent quality, energy consumption, uptime, process recovery rate, operating expense, environmental permit status or customer use is reported. Next question: Will commissioning establish sustained output at the intended quality and energy intensity, and will named operator offtake/permits and full unit costs support commercial-scale reuse or discharge? - [Texas Pacific Land Corporation Q2 2026 Form 10-Q](https://www.texaspacific.com/investors/sec-filings/all-sec-filings/content/0001811074-26-000058/tpl-20260630.htm) — Texas Pacific Land Corporation / U.S. SEC; Quarter ended 2026-06-30; filed 2026-08-04; Lines 977–980 (printed page 20). Access: Primary issuer-hosted SEC filing opened directly. - [Texas Pacific Land Corporation 2025 Form 10-K](https://www.sec.gov/Archives/edgar/data/1811074/000181107426000018/tpl-20251231.htm) — Texas Pacific Land Corporation / U.S. SEC; Year ended 2025-12-31; filed 2026-02-25; Lines 324–329. Access: Primary SEC-hosted filing opened in prior Round 04. ## Texas Pacific Land Corporation / TPWR and Transmissive Water Services — Produced-water desalination capacity and existing water sales do not yet establish replicable unit economics T10-R09-LIMITS-01 · Water, emissions & energy · Q2 and six months ended 2026-06-30; filed 2026-08-04. Initial design capacity 10,000 bbl/day represents less than 0.05% of the issuer-stated Permian produced-water flow (>20 million bbl/day); commissioning had begun, so capacity is not treated volume. Existing Q2 water-sales gross revenue averages about $0.66/bbl; this is a separate service business, not desalination price, margin or return. Status: Construction reported complete and commissioning commenced in Q2 2026; no achieved operating capacity or customer/throughput is reported. Economics: TPL existing Q2 water sales: $39.733m revenue / 60.343m barrels = approximately $0.66 per barrel of gross segment revenue (calculated from the filing). Water service-related expenses were $11.6m for the quarter, but include treatment, transfer, purchases, repairs/maintenance, rentals and fuel across services; not allocable to the desalination project. No desalination throughput, price, energy cost, CAPEX recovery, unit operating cost, avoided injection fee or treatment margin disclosed. Baseline: Source separately reports current commercial water sales/volumes and new desalination design scale. Water sales volumes depended on customer demand; cost drivers and delivery-location/right-of-way expenses vary. These commercial segment values are not same-feedstock or same-process comparator data for produced-water desalination. AI role: None disclosed; physical treatment process and equipment. Limits: Do not call the $0.66/bbl revenue proxy a desalination price or margin. Do not divide whole-segment costs by commercial water-sale volume and present as process OPEX. Basin flow denominator is rounded/issuer-estimated, and design capacity is not actual output. The 0.05% ratio describes only the initial test facility relative to basin-scale flow; it says nothing about number or economics of eligible customers, produced-water quality, treatment recovery or disposal alternative cost. No post-commissioning result yet. Next question: What are commissioning throughput and yield, product-water quality, energy and consumables per barrel, full installed capital per sustained bbl/d, delivered treatment price, disposal alternative cost and contracted eligible volume at the same location? - [Texas Pacific Land Corporation Q2 2026 Form 10-Q](https://www.texaspacific.com/investors/sec-filings/all-sec-filings/content/0001811074-26-000058/tpl-20260630.htm) — Texas Pacific Land Corporation / U.S. SEC; Quarter ended 2026-06-30; filed 2026-08-04; Lines 977–980 (scale and project stage); 890 (not an oil/gas producer); 1149–1152 (sales/customer-demand and expense context); 1159–1183 (volumes, revenue and service expenses). Access: Primary issuer-hosted SEC filing opened directly. ## Texas Railroad Commission / Permian Basin operators — Produced-water disposal permitting now constrains pressure, volume and geologic eligibility T10-R04-LIMITS-02 · Water, emissions & energy · Announced 2025-05-16; effective 2025-06-01. The RRC requires expanded half-mile area-of-review, geology-based maximum surface injection pressure, and reservoir-pressure-based maximum daily injection volumes. Operators must assess old/unplugged wells and show injection will not fracture confining strata. Status: New regulator guidelines effective 2025-06-01. Economics: No quantified basin-wide cost, delay or alternative-treatment economics are supplied. The requirements can affect eligible disposal wells and available injection throughput, but no specific asset-level impact was established here. Baseline: RRC states area of review increased from quarter-mile to half-mile; maximum pressure and volume criteria apply to new/amended permits, while wells within 25 km of seismic events remain subject to separate seismicity review. AI role: None; regulator permitting requirements and subsurface pressure/geology constraints. Limits: This is regulator context, not evidence that a particular operator's disposal well was denied or curtailed. Avoid inferring that a specific desalination, recycling or electrification technology is economically preferred without asset-specific cost and permitting evidence. Next question: For a proposed water solution, what permitted injection volume, storage/takeaway capacity and treatment economics are available at the specific operating asset after applying the local rules? - [RRC Issues Enhanced Guidelines for Permian Basin Disposal Wells](https://www.rrc.texas.gov/news/05162025-permian-disposal-wells-guidance-release/) — Texas Railroad Commission; 2025-05-16; effective 2025-06-01; Lines 41–58. Access: Primary regulator page opened directly. ## TotalEnergies / Cognite — Three-year scale-up of industrial data and AI platform across operated upstream assets T10-R07-GLOBAL-02 · Data, people & purchasing · Announced 2025-09-26; implementation period three years. Three-year planned scale-up across operated upstream assets; release states data unification/AI readiness and intended visualization, equipment monitoring and AI acceleration. This is an agreement/objective, not proof rollout completed or a realized result. Status: Announced three-year scale-up agreement, Sep 2025; staged global rollout planned, not evidence of completed global deployment. Economics: No agreement value, license/integration cost, data-integration effort, operating savings, or benefit-sharing terms disclosed. Baseline: No starting coverage, number of sites/apps/data points, implementation milestones, or before/after operating metrics disclosed. AI role: Agreement bundles industrial-data foundation and AI platform. Operator specifically says the data layer unifies data and makes it AI-ready; deployment acceleration of AI is an objective, not evidence of individual AI workflows or realized value. Limits: Strategic supplier announcement; three-year program means full scope is future dated. “All operated upstream assets” is intended scope, not verified completion. No individual workflow result or vendor lock-in/data portability terms reported. Next question: What sites and workflows are actually onboarded at each tranche, what is the total platform/implementation cost, and can TotalEnergies exit or port contextualized asset data and workflows? - [Digital Transformation: TotalEnergies and Cognite Expand Their Partnership to Scale Industrial AI](https://totalenergies.com/newsroom/digital-transformation-totalenergies-and-cognite-expand-their-partnership-scale/?lang=eng) — TotalEnergies; 2025-09-26; Lines 31-46. Access: Direct issuer release opened. ## Woodside Energy / AWS — Woodside launches AI-assisted maintenance planning at North West Shelf T20-R12-GLOBAL-01 · Maintenance & reliability · North West Shelf launch reported 5 January 2026, following an Angel platform test. Woodside says Maint Intel recommends equipment-maintenance intervals using maintenance records, prior plans and reliability targets. Its Digital team and AWS collaborators reduced model-processing time from five days to under two hours. The issuer says this reduced unnecessary activities but supplies no count, work-hours, deferred maintenance, failures, availability or cost result. Status: Woodside reported launch at NWS and prior successful Angel test on 2026-01-05; global rollout to other assets was scheduled for later in 2026, not yet reported as completed in this source. Economics: No program cost, avoided maintenance spend, incremental downtime, cost per asset, failure rate or realized financial value disclosed. Processing-time reduction is a workflow productivity metric, not equipment reliability ROI. Baseline: Five-day prior model-processing cycle compared with under two hours after workflow development; the source does not define the comparison sample, data volume, model count or whether other workflow changes contributed. AI role: Woodside says AI analyzed maintenance history and equipment failure modes, then advanced statistical models recommended maintenance intervals against maintenance plans and reliability targets. Human maintenance teams remain the decision users; source does not say the tool autonomously changes work orders. Limits: Company feature story and reported statements, not independently audited. NWS launch and Angel test are distinct deployment phases. Do not infer increased uptime, fewer failures, or maintenance savings from faster model turnaround alone. No performance metric denominator or matched equipment cohort. Next question: How many assets and rotating/static equipment units use Maint Intel, what maintenance actions were changed, and what matched-cohort changes occurred in cost, unplanned downtime, failures, or availability after launch? - [Right fix at right time – AI delivers for Woodside](https://www.woodside.com/media-centre/news-stories/story/right-fix-at-right-time---ai-delivers-for-woodside) — Woodside Energy; 2026-01-05; Dated story lines 13-45: NWS launch, Angel test, workflow, rollout plan, and processing time statement.. Access: Issuer HTML opened directly. ## Woodside Energy — Pluto LNG flare upgrade pairs composition measurement with continuous combustion imaging T20-R15-GLOBAL-01 · Water, emissions & energy · Announcement 2026-03-31; no steady-state measurement window or before/after operating data provided. The upgrade includes a new gas chromatograph measuring specific flare-gas components, including methane and nitrogen, and advanced VISR cameras using a multispectral imager for continuous combustion-performance monitoring. Woodside says the cold/dry flare stream contains approximately 50–70% nitrogen, while prior Australian reporting had assumed flare gas was entirely hydrocarbons. The stated aim is better flare tuning, smoke management, safety, and emissions measurement; no measured improvement is quantified. Status: Woodside reported delivery of new flare-system technology on 2026-03-31 after more than five years of planning/design. The system was entering staged maintenance through 2026 to remain ready for Pluto Train 2 integration. Economics: No project cost, methane-slip reduction, combustion destruction efficiency, flare uptime, avoided reporting emissions, or payback disclosed. The system's direct value includes improved composition measurement and visibility; reported aspiration is not an achieved emissions metric. Baseline: Prior emissions reporting assumed the stream was entirely hydrocarbon, although Woodside reports 50–70% nitrogen in the cold-dry stream. There is no stated baseline for combustion efficiency, smoke events, methane slip, or downtime. AI role: No AI is disclosed. The deployed method uses a gas chromatograph for flare-stream methane/nitrogen composition and multispectral VISR cameras for live continuous combustion monitoring. Operators can use the measurement to adjust flare performance; autonomous control is not claimed. Limits: This is an LNG/process-facility adjacent case, not upstream well technology. It is a new measurement/monitoring package paired with flare hardware, and may be relevant to upstream operators with gas handling or flaring; avoid transferring Pluto's value without comparable flare composition, flow and safety constraints. The source does not quantify performance post-installation. Next question: After commissioning, what are observed flare destruction efficiency, methane-slip, smoke events, composition-adjusted reported emissions, safety/reliability and maintenance outcomes against a matched pre-upgrade baseline? - [Pluto’s flare upgrade redefines safety and emissions performance](https://www.woodside.com/media-centre/news-stories/story/pluto-s-flare-upgrade-redefines-safety-and-emissions-performance) — Woodside Energy; 2026-03-31; Dated story lines 13-40; five-year design/delivery and technologies/mechanism/limits.. Access: Issuer HTML opened directly. - [Scope 1 and 2 GHG emissions](https://www.woodside.com/sustainability/environment/climate/Scope-1-and-2-ghg-emissions) — Woodside Energy; Current issuer page; cites 2025 operating data; Pluto flare composition optimisation and VISR measurements, lines 98-121.. Access: Issuer HTML opened directly. ## October 14 targeted addition: TotalEnergies SLB announced a 15-year DrillPlan agreement for TotalEnergies’s global upstream portfolio. The work connects subsurface information with well design and engineering. Offset-well data, trajectories, casing, completion inputs and planning assumptions are brought into a common environment while retaining customer workflows and internal digital expertise. No rollout population, achieved operating benefit, contract value or return is disclosed. The release does not identify an AI component. It belongs in operating-technology context, not the AI-adoption count. A 15-year term does not prove the whole portfolio is already live. [SLB awarded digital drilling contract by TotalEnergies](https://investorcenter.slb.com/news-releases/news-release-details/slb-awarded-digital-drilling-contract-totalenergies) · 2026-10-08 ## October 14 evening: Eni and Halliburton Halliburton describes LOGIX coordinating drilling, tripping and managed-pressure drilling on an Eni deepwater exploration well offshore Indonesia. Rig equipment, automated well placement, downhole hydraulics and pressure controls operate in a connected closed loop. Human approval, handoff and safety-interlock arrangements are not detailed. Supplier reports over 15% better efficiency, without defining the efficiency measure, baseline or absolute time/cost effect. The release does not identify the AI/ML elements. Retained as operating technology, outside the AI-adoption and result counts. [Eni and Halliburton achieve industry first with closed-loop rig automation in deepwater Indonesia](https://www.halliburton.com/en/about-us/press-release/eni-halliburton-achieve-industry-first-closed-loop-rig-automation-indonesia) ## October 15 morning: integrated optimization and purchasing ### Suncor Suncor describes one linear-programming model connecting its oil-sands operations, Edmonton refinery, domestic markets and global export sales. During an upset, teams can change the crude slate, move diluent and route intermediate streams. The stated 10,000–15,000 barrels/day is intermediate-stream routing capability. It is not incremental AI production, savings or measured uplift. Record throughput and sales are not assigned to the optimizer. No architecture beyond linear programming, counterfactual, model-specific benefit, implementation cost or detailed control boundary. Keep it outside new AI-adoption and result counts. Suncor haulage and truck dispatch remain separate existing records. [Second Quarter 2026 Financial Results Call](https://www.suncor.com/-/media/project/suncor/files/investor-centre/transcripts/2026-q2-suncor-energy-financial-results-call-transcript-en.pdf?created=20260818150055&modified=20260828195420) · PDF page 10 (page index 9), Q&A lines 331–358, Dave Oldreive and Rich Kruger; call dated 5 August 2026 ### Petrobras Petrobras reports automating purchase requests to speed up purchasing and avoid costs. The same slide also describes AI projects, but this purchasing box does not identify AI. Petrobras reports 190,000 items automated per year and average avoided costs of R$111 million per year. Avoided cost is not the same as cash collected or net savings after implementation and support costs. Company-reported values without a disclosed baseline, cost method, implementation cost or independent validation. The slide does not establish whether the estimate is realized or forecast. Do not count this as AI adoption or add it to the separate AI benefits. [2026–2030 Business Plan](https://petrobras.com.br/documents/2677942/12513210/PN%2B2026-30_27nov_final_ing.pdf/3d500254-cbf1-c5ce-56b6-008b4a28a733?download=true&t=1765223843000&version=1.0) · Printed slide 158 / PDF page 158, purchase-request automation box; separate from the AI and digital-twin boxes ## October 15 follow-up: assistants, controls and foundations ### Chevron SPE reports that Chevron’s Peter Nesom described GEORGE, a production-engineering assistant used for analysis, optimization and automation tasks on more than 5,000 wells. More than 5,000 wells is reported use scope. No production uplift, time saving, cost or net return is supplied. Original conference reporting of named operator remarks, not an operator-authored deployment paper. The event program corroborates speaker and session only. Keep as reported use awaiting technical corroboration, outside additional adoption/result counts. Do not infer an LLM architecture, autonomous field control or a company-wide knowledge system. [Maximizing Every Barrel: Artificial Lift Embraces Automation](https://jpt.spe.org/maximizing-every-barrel-artificial-lift-embraces-automation) · 2026-09-04 · Chevron section: Peter Nesom remarks following the image caption; 25 August 2026 plenary, reported 4 September ### Petrobras A Petrobras-coauthored paper proposes a GPT-based assistant for reservoir-management systems, intended to let engineers use natural-language requests instead of navigating technical interfaces. The authors report favorable test behavior and training loss approaching zero. No held-out task-success denominator, active users, field result or economics is disclosed. Training loss is not proof of reliable operating decisions. Publisher-indexed original abstract reviewed; full paper not retrieved. Retained as research, outside operating adoption and outcome counts. No evidence of permission to change reservoir plans or execute downstream actions. [Virtual Assistant Using Generative Pre-Trained Transformers on Reservoir Management Context](https://onepetro.org/OTCBRASIL/proceedings-abstract/25OTCB/25OTCB/D011S005R007/792383) · 2025-10-21 · OTC-36185-MS; publisher-indexed author list and abstract, test-scenario paragraphs ### ExxonMobil ExxonMobil authors describe a closed-loop ESP pilot on 26 Midland wells. Commercial physics-based software recommends changes; internal checks test facility constraints and pump limits before execution. The authors describe less surveillance and execution work, without numerical time, production or financial results. No complete cost or net return is supplied. Original publisher-indexed abstract reviewed; full paper not retrieved. It establishes physics-based control and a validation gate, not an LLM controlling pumps. This is separate from AI-235’s gas-lift ML and its oil-uplift figures. [Closed-Loop ESP Optimization to Maximize Asset Value](https://onepetro.org/SPEESP/proceedings-abstract/25ESP/25ESP/D031S006R001/788064) · 2025-08-18 · SPE-225261-MS; abstract, system checks and 26-well pilot paragraphs ### Aethon Energy Aethon CIO Kang Chen reports a full cloud migration and technology overhaul, along with new data-quality, analytics and workforce-training initiatives. No quantitative operating or financial result is supplied. Named executive self-report in an advertising supplement, not independent technical validation. Cloud migration and upskilling do not prove AI readiness or deployment. Keep as company transformation context, outside AI adoption and outcome counts. [Dallas CIO ORBIE special advertising supplement](https://media.bizj.us/view/img/13143357/low-res-dallasorbiesciospecial-section2026-8951090d-b07f-4f22-bfba-90c5a075a999.pdf) · 2026-02-20 to 2026-02-26 · PDF page 10 / printed page 10, Kang Chen profile; supplement dated 20–26 February 2026 ## October 15 evening: control and engineering feedback ### Repsol Repsol says it completed the initial phase of its first Autonomous Optimizer with Control, developed internally for one section of a glycol plant. No measured plant result, cost or net return is disclosed. Completing an initial phase does not establish routine autonomous operation or a plant-wide rollout. Keep separate from Repsol’s multiagent, workforce and legal programs. [2025 Group Annual Financial Report](https://www.repsol.com/content/dam/repsol-corporate/en_gb/accionistas-e-inversores/ref/2026/ref19022026-2025-group-annual-financial-report.pdf) · 2026-02-19 · PDF page 195 / printed Management Report page 45, section 5.5: AI-based autonomous optimizers ### Corva / Oliasoft The vendors have developed an integration between Oliasoft WellDesign and Corva Real-Time AI and say they are preparing customer pilots. No pilot customer, well or rig count, measured benefit, implementation cost or ROI is disclosed. Model feedback is not evidence of automatic rig control. Two vendor versions of the same release do not provide independent validation; action permissions are unspecified. [Corva and Oliasoft Create a Continuous Learning Loop Between Well Design and Drilling](https://www.corva.ai/blog/corva-and-oliasoft-create-a-continuous-learning-loop-between-well-design-and-drilling) · 2026-10-12 · Joint release: engineering-to-operations, operations-to-engineering, and customer-pilot paragraphs ## October 16 morning: use-case fit and operating scope ### TotalEnergies Control Global quotes TotalEnergies engineer Thomas Klein saying a Port Arthur drum-level model achieved 75% prediction accuracy, but the drum-cycling use case was not retained. The report does not define the accuracy denominator or test set, or quantify cost or loss. Other pilot work continued; the entire program was not reported abandoned. Original conference reporting, labeled Sponsored. We could not retrieve the presentation or a later company statement. The earlier company release confirms the broader pilot, not the decision to drop this use case. Keep these figures separate from its five-event result. [Honeywell AI pilot aids coker unit operations at TotalEnergies refinery](https://www.controlglobal.com/show-coverage/honeywell-users-group/article/55383668/honeywell-ai-pilot-aids-coker-unit-operations-at-totalenergies-refinery) · 2026-06-11 · Drum cycling prediction section; Thomas Klein remarks at Honeywell Users Group Americas 2026 ### Helmerich & Payne H&P reports two FlexRobotics systems deployed on its rigs for ExxonMobil. It expects seven more over the following twelve months, for nine in total. Two is the deployed count; nine is the intended total. No quantified safety, time, cost or net return is disclosed. An ExxonMobil executive supports expansion intent; the joint account is not independent outcome validation. It does not establish installation of the separate AI camera add-on in AI-116. Keep as robotics context, outside AI-adoption counts. [H&P & ExxonMobil Expand Deployment of FlexRobotics Technology](https://ir.hpinc.com/news/press-releases/news-details/2026/HP--ExxonMobil-Expand-Deployment-of-FlexRobotics-Technology/default.aspx) · 2026-10-07 · Opening release paragraphs: first two systems, seven additional units over next 12 months; ExxonMobil quotation ### SLB SLB says Retina has imaged more than 50 wells across 18 locations during its first year of commercial deployment. Deployment and imaging specifications are not measured recovery, drilling-time savings or ROI. No comparative economics is disclosed. No AI/ML mechanism or automatic control is identified. Keep as instrumentation context; do not transfer OQ Oman’s separate drilling-tool results to Retina. [Retina At-Bit Imaging System Surpasses 50-Well Milestone in First Year of Deployment](https://www.slb.com/newsroom/features/2026/2026-1015-retina) · 2026-10-15 · Published date; Global adoption and Global deployment sections ## October 16 evening: PTTEP operating context At its March AGM, PTTEP described deployed AI and automation across subsurface, drilling and production work, including AI that monitors drilling data for early risk detection. The other examples are automated planning of well locations and wellhead platforms, and systems optimizing production across many wells. Their specific AI methods are not identified. Management says benefits are reaching financial performance, but gives no isolated AI result, baseline, cost or ROI. No fields, vendors, well count, model performance or approval/control boundaries are disclosed. The separate USD 328m new-business budget is not AI-only. This broad account does not establish the scope or success of the later offshore advisory pilot in AI-067. [Minutes of the 2026 Annual General Shareholders’ Meeting](https://www.pttep.com/storage/document/shareholder-meetings/agm2026/pttep-agm2026-minutes-en.pdf) · 2026-04-10 · PDF page 13 / printed page 12 of 25, response to question 1.7; cover letter on PDF page 1 confirms publication date ## October 17 morning: advisory use and data foundations ### Baker Hughes Baker Hughes describes an advisory-mode UAE carbonate-drilling case supported from its remote operations centre. AI assisted torque-and-drag work. Despite the product name, the case explicitly says advisory mode; who approved parameter changes is not stated. The vendor reports 22 alerts, 22 hours saved against a previous well, 5% higher penetration rate and no stuck pipe or fluid losses. The customer and well are unnamed. A prior-well comparison does not isolate AI from the wider automation service, crew or geology. Costs, accepted recommendations and repeatability are undisclosed; no ROI is established. The deployment date is unknown; the PDF revision does not establish a 2026 deployment. [Kantori drilling advisory case · UAE](https://dam.bakerhughes.com/m/6c9e1d4e04184778/original/i-Trak-service-maximized-ROP-saved-22hrs-UAE-cs.pdf#page=1) · PDF revised April 2026, with a 2023 copyright. Original publication and field-work dates are unknown; this may be older operating context. The September 2026 webpage date was not independently verified. ### Petroleum Development Oman PDO selected Kongsberg Digital’s Kognitwin to connect operational and enterprise data across operations and maintenance. The release calls it AI-integrated, but does not identify an AI task, model or authority to act. Safety, production and cost improvements are goals. No measured benefit, contract value or ROI is disclosed. No fields, facilities, users or workflow counts are identified. The company-description paragraph about gas operations does not define deployment scope. A passed target date alone does not establish delay or failure. [Kongsberg Digital and PDO digital-twin agreement](https://kommunikasjon.ntb.no/pressemelding/18645626/kongsberg-digital-and-petroleum-development-oman-sign-agreement-for-deployment-of-ai-powered-digital-twin?lang=en&publisherId=17849150) · Announced 8 September 2025, with an ambition to start deploying the twins that year. No later completion evidence was verified in this pass. ## October 17 evening: INPEX business-system foundation INPEX reports that SAP Cloud ERP Private began full operation in Japan in January 2026. With ABeam’s support, it aligned business processes to SAP standards instead of filling gaps with custom add-ons. The release says development was shortened, without giving elapsed time or a comparison. Reporting and workload improvements are objectives; no measured savings or ROI are disclosed. AI appears as compatibility and a future efficiency objective, not a deployed function. Users, departments, costs and approval rules are unspecified. No connection to the separate Copilot/management-meeting agents in AI-294 is established. [INPEX commences operations of SAP Cloud ERP](https://www.inpex.com/english/news/upload/20260624.pdf#page=2) · 2026-06-24 · PDF pages 1–2; Japan go-live and Fit-to-Standard approach on page 2 ## October 18 morning: energy data access and field foundations ### S&P Global Energy S&P Global Energy describes giving internal and customer agents conversational access to structured energy datasets. Domain experts curate narrow dataset agents. A reusable MCP interface combines them across commodities. The account says table permissions and audit controls carry through, and experts can test answers against benchmarks. A customer executive reports setup shrinking from development cycles measured in months to days. No sample, measured accuracy, active-user count, cost or ROI is supplied. This is a vendor-published customer account, not an independent quality or security assessment. Benchmarks are described without scores. Access to research data does not establish authority to trade, approve investment or control field equipment. [From Data to Dialogue: S&P Global Energy’s Genie and MCP architecture](https://www.databricks.com/blog/data-dialogue-how-sp-global-energy-made-its-structured-data-estate-conversational-databricks) · 2026-09-25 · Date; per-dataset curation; Unity Catalog permissions; FastMCP composition; customer quote and benchmark loop ### PetroReconcavo PetroReconcavo reports implementing SAP and activating six LTE towers in Bahia and Rio Grande do Norte. The ERP work connected production data and material tracking. The field network supports monitoring and prepares for IoT and automation. A separate data-lake passage describes GenAI analytics, reports and assistants without naming a tool or defining live use. The report lists more than 2,200 training hours for SAP. Faster response and stronger controls are qualitative claims; no measured savings, cost or ROI is disclosed. No partner, user population, AI approval rules or autonomous control is specified. Tower and training counts are implementation scope, not economic outcomes. Company-wide emissions and social results are unrelated and are not attributed to this technology. [Relatório de Sustentabilidade 2024](https://www.petroreconcavo.com.br/Download.aspx?Arquivo=OXSl%2FB7SEmrx%2BHoNzBeFlQ%3D%3D#page=78) · 2025-07-14 · PDF / printed page 78: SAP, Projeto Waves and data lake [2024 Sustainability Report · Notice to the Market](https://ri.petroreconcavo.com.br/Download.aspx?Arquivo=6hzxLJkhGOrPzcEDtUKx3A%3D%3D&IdCanal=ckeQVNzVtM%2FZZ4T5tqRLLw%3D%3D) · 2025-07-14 · PDF pages 1–2; publication announcement and English-edition timing ## October 18 evening: supplier-to-operator maintenance recommendations NOV describes its agent interfacing with a Norwegian operator’s agent at a sulphate-removal unit. Real-time equipment diagnostics on the Max Platform support maintenance recommendations. The release does not establish automatic execution or identify an LLM. No measured accuracy, downtime reduction, savings, cost or ROI disclosed. Supplier-only account; operator, unit and user population unnamed. Approval, override and safety rules undisclosed. Separate service-rig monitoring in the same release is not evidence of AI. [NOV Reports Second Quarter 2026 Earnings](https://investors.nov.com/news-releases/news-release-details/nov-reports-second-quarter-2026-earnings) · 2026-07-28 · Significant Achievements: Norwegian sulphate-removal unit; separate remote service-rig monitoring paragraph ## October 19 morning: trading-data controls AWS describes a TotalEnergies Trading & Shipping system that matches structured and unstructured transaction data. The workflow supports transaction traceability and input-error detection. AWS lists Bedrock and ECS; the model, connectors and exception-handling process are unspecified. A named customer manager reports more reliable controls and fewer financial losses, without amounts, error rates, a baseline or an observation period. AWS page undated. Implementation date, trading-desk scope, user count and transaction volume unknown; an in-window launch is not established. Vendor-published customer testimony, not independent validation. No costs or net ROI. Matching records does not establish authority to place trades, release payments or change books. No specific model provider can be inferred from the Bedrock service description. [TotalEnergies puts Gen AI to work on its trading operations](https://aws.amazon.com/solutions/case-studies/totalenergies/) · Undated · Overview; Olivier Collon customer quotation; AWS Services Used ## October 19 additional: maintenance lessons and remote operations ### Murphy Oil 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. Article: 20 January 2025. Alert records begin January/March 2023, before this report’s rolling window; exact project start/end dates unknown. 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. How many alerts lead to useful work, and what does reviewing the rest cost? [Framework for AI- and ML-Based Predictive Maintenance for Offshore Rotating Equipment](https://jpt.spe.org/framework-for-ai-and-ml-based-predictive-maintenance-for-offshore-rotating-equipment) · 2025-01-20 · Methodology; Observations; Discussion: Project Planning and Management, Data Prerequisites, Domain Knowledge, Model Development, Performance, and User Experience; Figs. 4–6 ### NOV NOV reports controlling a full-size crane from about 20 km away and delivering one Aura-equipped remote-operated crane to an unnamed customer. Aura brings camera feeds, load information, equipment diagnostics and environmental data into the human operator’s screen. Customers are expected to start onboard before moving control ashore. No measured cost, safety or crew reduction. Two further cranes were being built for delivery by early 2027; these are planned deliveries. Article: 20 August 2026. Supplier test: Kristiansand to Søgne, Norway; exact test and first-delivery dates unknown. No AI/ML identified. Delivery does not establish shore-controlled customer operations or autonomous lifting. Customer, vessel, operational hours, interlocks and fallback arrangements are undisclosed. What must be proven onboard before a customer can safely move the operator ashore? [Aura moves remote crane operations from concept to reality](https://www.nov.com/news/aura-moves-remote-crane-operations-from-concept-to-reality) · 2026-08-20 · Remote-access paragraph; Kristiansand/Søgne test; first delivery; two planned deliveries; phased customer adoption ### Oceaneering Oceaneering says an Ocean Intervention II survey campaign began in April 2026 for an unnamed international operator offshore Trinidad and the U.S. Gulf. The scope combines geophysical and geotechnical work with surface-vessel-supported autonomous underwater surveys. U.S. Gulf pipeline inspections include non-contact cathodic-protection measurement. The supplier says combining work on one vessel and crew can reduce mobilizations, exposure and emissions. No measured savings, baseline or net ROI disclosed. Announcement: 30 June 2026. About six months of work expected, including roughly four in Trinidad and two in the U.S. Gulf; not verified completed durations. Autonomous survey is not proof of ML or LLM use. The vessel still has a crew. Operator identity, human supervision, intervention rules, achieved campaign length and comparative performance are undisclosed. Does combining survey tasks cut total vessel days once weather, rework and supervision are included? [Oceaneering Secures Multi-Discipline Cross-Regional Offshore Survey Contracts for Ocean Intervention II](https://www.oceaneering.com/oceaneering-secures-multi-discipline-cross-regional-offshore-survey-contracts-for-ocean-intervention-ii/) · 2026-06-30 · June 30 release: opening contract/program paragraphs; Trinidad/U.S. Gulf scope; simultaneous operations; supplier quotation; 2025 vessel upgrades ## October 20 morning: Oil India programme plan Oil India describes ten DRIVE 2.0 workstreams, including a control centre, a digital oilfield covering 380 producing wells, subsurface agents, maintenance and corporate workflows. The proposed systems would connect operating data with engineering and business records. Rig-planning approvals and safety notifications to designated people are described; AI decision authority and safeguards are not. The 380 wells are planned scope, not a verified connected population. No measured uplift, savings, implementation cost or net ROI. Publication and DRIVE 2.0 start dates unknown. Header: “11/09/2026 09:35 AM”; footer: “20/10/2026 11:28 AM”. These update strings do not date deployment. The 2019 start belongs to earlier DRIVE. One undated issuer plan, now accessible after an earlier blocked route. No completion dates, live user population or model providers established. Ten workstreams are not ten operating AI deployments. Which parts have been commissioned, and who can approve or override their recommendations? All ten components below are proposed scope; commissioning is unverified. - Operations data / IT, data & enterprise transformation: A shared control centre and OT data lake; corporate, field, pipeline and Rajasthan locations. Source: DRIVE 2.0 §1. - Connected wells / Production & field operations: Proposed sensors on 380 producing wells, remote monitoring/control, lift optimization, integrity models and virtual flow measurement. Source: DRIVE 2.0 §2. - Subsurface assistants / Exploration & geoscience: Proposed GenAI/agent tools linking subsurface and operating datasets for forecasts and drilling/workover recommendations. Source: DRIVE 2.0 §3. - Drilling and rig logistics / Drilling & completions: Rig scheduling, readiness/approval workflows and bit-selection support using operating and ERP data. Source: DRIVE 2.0 §4(a)–(b). - Surface-facility twins / Refining, LNG & process operations: Planned models at OCS Makum, FGGS Madhuban and GCS Hebeda, with engineering records and historian/SAP links. Source: DRIVE 2.0 §5. - Equipment maintenance / Maintenance & reliability: An asset registry and lifecycle/maintenance workflows using SAP S/4HANA; proposed sensor-based failure warnings. Source: DRIVE 2.0 §6. - Camera safety alerts / HSE & emissions: Existing camera feeds and edge processing would flag PPE, access and other hazards for designated people. Source: DRIVE 2.0 §7. - Project control / Finance, investment & strategy: Forecasts of project dates, cost overruns and bottlenecks, with scenario views and audit trails. Source: DRIVE 2.0 §8. - Purchasing / Supply chain & logistics: Requisition-to-order automation, bid support and procurement analysis linked to SAP S/4HANA and GeM. Source: DRIVE 2.0 §9. - Workforce tools / Workforce & learning: A proposed HR system joining onboarding, workforce analytics/chatbot, access, time, benefits, training and contractors. Source: DRIVE 2.0 §10. [Digitalization in OIL — DRIVE 2.0 programme](https://www.oil-india.com/digitalfootprint) · Undated · Digitalization in OIL; DRIVE 2.0 §§1–10; §2 380-well scope; §4 approvals; §7 notifications; header and footer update strings ## October 20 evening: availability, content and a planned appraisal workflow ### Atlas Energy Solutions Kodiak reports a 64-day uptime streak for one driverless truck in Atlas’s Permian Basin fleet, plus five trucks with streaks above 45 days. The truck and autonomy software stayed powered on and available for service. This is physical driving automation, not a language-model assistant. No one was in the cab; remote support and intervention boundaries are not quantified. As of June 30: 35 Atlas-owned and operated driverless trucks and over 40,000 cumulative paid driverless hours. These are supplier-reported fleet figures, not the uptime of every truck. No fleet availability rate, savings or net ROI. Announced October 20, 2026. The 64-day streak ran June 30–September 1. The fleet and paid-hours figures are as of June 30; 100 trucks by mid-2027 and a year-end highway launch remain plans. Available does not mean continuously driving or hauling. No mileage, dispatch utilization, collision/exposure comparison or cost denominator disclosed. Kodiak AI is separate from Kodiak Gas Services (AI-193). How much paid hauling did those available hours produce, and what did remote support, maintenance and autonomy cost per load? [Kodiak Driver Sets New Uptime Record By Running 64 Consecutive Days on an Unsupervised Driverless Truck](https://kodiak.ai/news/kodiak-driver-uptime-record) · 2026-10-20 · Opening uptime definition and dates; Atlas fleet as-of paragraph; future fleet and highway plans · Full original supplier release retrieved (HTTP 200) and read. Supplier-reported availability and fleet figures; no independent reliability, safety or financial audit. ### SPE / OnePetro SPE and i2k announced extended automated classification of OnePetro content. Separately, SPE says customers often wanted its technical content inside tools they already used. i2k describes metadata tagging using more than 20 domain knowledge bases. SPE’s original JPT account says EnRG-LLM remains licensable and identifies Halliburton Landmark as its first content-license customer in an undated follow-up paragraph. No measured retrieval accuracy, time saving, token saving, license revenue, usage population or net ROI. General claims about i2k’s geotechnical agents do not establish that SPE uses every advertised workflow. Classification release: October 20, 2026. JPT article: September 10 publication header, with an undated following-conference licensing update. Neither dates completed classification rollout. The complete issuer release was verified on syndication pages; the Business Wire original timed out. This is not an abandoned EnRG-LLM product. The two announcements do not prove that classification replaced the model. 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. [SPE extends i2k Connect information-classification collaboration](https://wrex.marketminute.com/article/bizwire-2026-10-20-society-of-petroleum-engineers-spe-partners-with-i2k-connect-for-information-classification) · 2026-10-20 · Opening announcement; classification paragraph; SPE CEO statement · Complete issuer-authored Business Wire release read on an accessible syndication page. Distributor original timed out; identical release also retrieved from AOL. No independent outcome validation. Distributor URL: https://www.businesswire.com/news/home/20261020770911/en/ [Industry Executives Offer Practical Advice on How To Use Modern AI Models Wisely](https://jpt.spe.org/industry-executives-offer-practical-advice-on-how-to-use-modern-ai-models-wisely) · 2026-09-10 · SPE’s AI Journey; licensing-demand discussion; following-conference Landmark paragraph; cost and metadata discussion · 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. ### bp Halliburton’s August contract for bp’s Bumerangue appraisal campaign includes LOGIX automation and remote operations offshore Brazil. The integrated drilling and evaluation scope connects planning, execution and evaluation. The release says LOGIX will be deployed; it does not identify an LLM, individual control permissions or human override arrangements. No number of wells, completed automated wells, operating comparator, quantified benefit, contract value or net ROI disclosed. Original announcement August 25, 2026; checked October 20. This is older evidence newly added to this review, not a launch announced today. A contract is not proof of completed field use. Do not transfer LOGIX results from Eni’s Indonesia well or another customer to Bumerangue. Recorded as a named-operator extension of existing LOGIX coverage, not a second platform. Which drilling tasks will run automatically, who can intervene, and how will bp compare outcomes with a credible baseline? [bp awards Halliburton integrated contract for Bumerangue field appraisal in Brazil](https://www.halliburton.com/en/about-us/press-release/bp-awards-halliburton-integrated-contract-for-bumerangue-field-appraisal-in-brazil) · 2026-08-25 · Opening and contract-scope paragraphs; LOGIX sentence · Full original supplier release retrieved (HTTP 200) and read. Future-tense scope; no completion or measured result established. ## October 20 evening scheduled: bounded execution and dispatch coverage ### U.S. Energy Development Corporation 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. 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. 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. Ask for the all-in cost and benefit bridge, the baseline calculation and the log of exceptions, overrides and rejected adjustments. [The Bottleneck Isn’t the Model. It’s the Execution.](https://www.aogr.com/magazine/cover-story/the-bottleneck-isnt-the-model-its-the-execution) · 2026-07 · July 2026 cover story; A New Optimization Philosophy; Eric’s Inside View; author biographies ### Ecopetrol Ecopetrol says its ISA-built energy decision platform managed 173 MW at Rubiales from March 2026, across five generation centres and the national grid. CADE supports choices of electricity source, timing and volume. Sixty workers were trained; that is not a user-adoption count. The earlier report describes a Rubiales dispatch-model MVP being built in 2025; the newer platform coverage does not independently confirm that model’s commissioning or AI control authority. The company describes the covered supply as 20% of group demand. Another 931 MW and 80% coverage are plans, not achieved scope. MW is operating power coverage, not energy saved, emissions avoided or ROI. Issuer announcement May 4, 2026; Rubiales coverage reported from March 2026. Older evidence newly checked, not a new October launch. No measured saving, implementation cost, LLM/model detail or human approval/override rule. The name “autonomous” and platform capacity do not establish autonomous AI dispatch. Group-wide generating capacity is a different population. Which decisions are actually automated, and what changes in delivered electricity cost after all platform and operating costs? [Grupo Ecopetrol inaugura centro para optimizar su gestión energética](https://www.ecopetrol.com.co/wps/portal/Home/es/noticias/detalle/grupo-ecopetrol-inaugura-centro-para-optimizar-su-gestion-energetica) · 2026-05-04 · May 4 heading and Bogotá dateline; Rubiales operating coverage, ISA partnership, projected expansion and training paragraphs ## October 21 morning: engine scope, confirmed diagnoses and solver attribution ### PTTEP PTTEP says it launched X.brain, an internal AI engine grounded in its own operating knowledge and data. The November 2025 feature describes support for analysis and recommendations. Eight specialist agents across subsurface, drilling, production, supply chain and corporate work were planned; the article does not establish that all eight went live. No X.brain user count, adoption rate or measured return is disclosed. The same article separately reports LogSight interpretation falling from two days to one hour and DORA drilling savings described only as “millions of US dollars annually”. Those are product-specific claims, not demonstrated X.brain results. Published 3 November 2025. Launch date, operating population and subsequent rollout status unknown. Missed historical evidence, not a new October 2026 announcement. No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons. Ask which agents people actually use, what data they can reach and which decisions still need a named human approver. - DORA: anticipate stuck pipe · Drilling & completions: Drilling Operation Real-Time Analyses combines machine learning with physics-based simulations to predict stuck-pipe risk and support parameter adjustments. The issuer claims annual savings of millions of US dollars, without an exact figure, well population, comparator or full costs. Automatic control and human approval rules are not disclosed. (PTTEP feature: DORA example) - LogSight: interpret well logs · Reservoir engineering & field development: AI supports petrophysical log interpretation. PTTEP reports a reduction from two days to one hour, without sample size, quality comparison, calendar period, human sign-off rule or full costs. (PTTEP feature: LogSight example) [PTTEP’s AI Vision for a Smarter, Sustainable Future of Energy](https://www.pttep.com/th/newsroom/special-features/1107/pttep-s-ai-vision-for-a-smarter-sustainable-future-of-energy) · 2025-11-03 · Page date; X.brain introduction; eight-agent plan; DORA and LogSight examples ### Energy Solutions / MNVR AI 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. 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. 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. Can the team show every alert, inspection and resulting work order—including alerts that turned out to be wrong? - Control relay · Maintenance & reliability: A gas-pressure trip left a generator motoring and drawing grid power for over nine hours. The authors say ODIE identified a relay rather than the initially suspected $18,000 controller. One hour of labor and a $100 relay are described; $25,500 estimated avoidance already includes the hypothesized wrong-controller path. (Customer PDF p. 1, case 1) - Ignition transformer · Maintenance & reliability: An agent flagged degradation; operators confirmed it and replaced the transformer during scheduled maintenance. Estimated cost avoided: $4,641, without a measured counterfactual. (Customer PDF p. 1, case 2) - Valve recession / detonation · Maintenance & reliability: An alert prompted operator inspection. A head replacement could then be planned. Estimated avoidance: $4,000; the account does not confirm the planned replacement was completed. (Customer PDF p. 2, case 3) - Silent NOx sensor · HSE & emissions: A sensor fault reportedly absent from traditional monitoring was identified. Replacement was proposed for the next convenient opportunity. Estimated avoidance: $1,000; completion is unverified. (Customer PDF p. 2, case 4) - Fouling gas filter · Maintenance & reliability: The team checked and confirmed a restriction and changed the filter during scheduled maintenance. Estimated avoidance: $1,874, based on the authors’ hypothetical failure progression. (Customer PDF p. 2, case 5) [Five Failures Across Five Systems](https://www.mnvrai.ai/s/MNVR-Customer-Story-Energy-Solutions-5-Examples-09222026.pdf) · Unknown · PDF pp. 1–3: seven generators; five fault cases; estimated costs; six-week period and retained diagnoses [ODIE — Operations Diagnostic Intelligence Engine](https://www.mnvrai.ai/odie) · Unknown · Human in the loop / operators make decisions section ### Quantum Capital Group Microsoft says algorithm changes in an updated version of Quantum’s field-planning tool reduced solve time by 50–65%. The integer-linear-programming optimizer selects feasible pads and wells. A Copilot Studio agent explains outputs. The numerical improvement belongs to the optimization algorithm; the source does not attribute it to the language model. 50–65% shorter solve times in an updated version, said to compound over hundreds of runs. The source does not specify absolute times, comparable hardware or the evaluation period. Do not multiply this with the separate three-weeks-to-20-minutes comparison. Microsoft story dated 30 April 2026; later operator identification remains QB Energy from Quantum’s 2026 report. This is an enrichment of AI-054, not a new deployment. Vendor/customer reporting with no full-cost return or independent performance validation. The version comparison is a different baseline from the manual-workflow comparison. Engineers retain plan review; no capital approval is delegated. The reported solver improvement is not attributed to the language model. Separate the solver, data preparation and assistant when testing which change actually improves the planning decision. [Field development planning at Quantum Capital Group](https://www.microsoft.com/en/customers/story/26476-quantum-capital-group-azure-app-service) · 2026-04-30 · Field development planning solution; Improved speed, scale, and precision; updated algorithm paragraph ## October 21 additional: daily use, commercial automation and field modeling ### Devon Energy Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data. The call distinguishes access to company data, AI-assisted coding and interpretation of operating data, and a next phase of process redesign. It does not establish that every example runs inside ChatDVN. Separate autonomous lift controls remain in AI-006. No active-user count, task-quality comparison, ChatDVN-specific cash benefit or full cost is disclosed. The call does not corroborate the older interview’s 15–30% productivity claim or AI-101 training count. Devon’s $1bn business-optimization target is not ChatDVN savings. Call dated 6 May 2026; CEO describes three years of operation, an approximate history rather than an exact start day. This is newly linked original evidence, not a new launch. Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status. Which jobs do employees finish through ChatDVN, and does the full review process improve time, accuracy and follow-through? [Devon Q1 2026 earnings call — internal ChatDVN workflow](https://s2.q4cdn.com/462548525/files/doc_financials/quarterly/2026/q1/Q1-2026-DVN-Q-A-Webcast-Commentary-Transcript.pdf) · 2026-05-06 · PDF and printed p. 4; Clay Gaspar prepared remarks, three waves of AI ### SLB SLB says its cementing technology had passed 100 remote and automated jobs by July 2026, following first-quarter commercialization. The release also names an automated dual-slurry operation on Shell’s Velox-1 in Egypt. The 100-plus figure combines remote and automated jobs; it is not 100 fully autonomous wells or a Shell-only fleet count. Job count shows reported operating reach. The release gives no controlled cycle-time, cement-quality, safety or cash-return comparison for this cohort. Issuer release 24 July 2026; first-quarter 2026 commercialization described. Individual job dates and the remote/automated split are undisclosed. Automation and remote operation do not establish machine learning or LLM use. Human supervision, control envelopes, fallback and override rules are not specified. The same release’s AI Production Studio is a pre-field demonstration environment, not proof that those capabilities ran on Shell’s well. Which steps execute automatically, what still needs the crew, and how does job quality compare under similar well conditions? [SLB Announces Second-Quarter 2026 Results](https://investorcenter.slb.com/news-releases/news-release-details/slb-announces-second-quarter-2026-results) · 2026-07-24 · Core / Technology Highlights; automated cementing paragraph and Shell Velox-1 follow-up ### PETRONAS PETRONAS reports Integrated Asset Management across eight fields, combining physics-based models with near-real-time data for surveillance and production optimization. The report says the workflow helps teams identify production opportunities faster. Its own passage does not attribute AI to IAM. Nearby PROTEAN and ERMAI AI descriptions concern different systems. The issuer says IAM improved production efficiency and value realization in 2025, without giving an amount, baseline, comparison group, attribution method or full costs. Eight is a field count, not eight models or platforms. Operating period 2025. Source register dates the 2025 integrated report 27 February 2026; exact posting day was not re-audited. Printed p.55 / PDF p.29 newly inspected. No field names, user population, human approval or automatic control permissions are disclosed. The chapter’s broad AI heading cannot establish IAM’s method. Do not transfer PROTEAN’s equipment footprint or ERMAI’s time comparison to IAM. What production decisions changed, and what incremental result can be separated from other field improvements? [PETRONAS Integrated Report 2025 — IAM](https://www.petronas.com/sites/default/files/uploads/PIR%202025/PETRONAS%20Integrated%20Report%202025.pdf) · 2026-02-27 · PDF p.29 / printed p.55; Turning Data into Production Upside ## October 21 evening: domain knowledge and a shared data foundation ### Shell Shell describes an initially unsuccessful search for pump-disruption precursors at Perdido. A field operator helped the data team refine its code and find a chemical signature. The team first drew a blank, later plateaued and considered stopping. The operator supplied equipment knowledge. Shell then reports a signature before 70% of disruptions, with one example about two days beforehand. The 70% is a reported precursor association in the analyzed events, not demonstrated live prediction accuracy. Event count, false alarms, held-out validation, intervention results, uptime benefit and full costs are undisclosed. Undated webpage and video; operating and evaluation dates unknown. Already mentioned in the coverage log, now structured as a readable case. It is not a new 2026 deployment. The transcript does not establish a live alert or control workflow, prevented disruptions or realized savings. The two-day example is not an average warning time. Initial stalled analysis is not proof of an abandoned or failed field deployment. Could the team reproduce the result on later events, with all false alerts recorded, and show what operators changed because of it? [The Data Detectives — Solving the 10-year-old problem at Perdido](https://www.shell.com/what-we-do/digitalisation/artificial-intelligence.html) · Unknown · Artificial intelligence examples / Perdido video transcript; initial blank, plateau, operator input and refined precursor signature ### Quantum Capital Group Databricks describes Quantum using Lakebase, Unity Catalog and Lakeflow to give deal teams a shared, governed view of geological, production and financial data. A Postgres interface and PostGIS geospatial support connect familiar tools, including Tableau and Spotfire. Schemas, stored procedures, triggers and access controls help teams use consistent datasets. This is data infrastructure; the story does not establish an AI decision-maker. Vendor/customer-reported scale: 1.5 billion records across six vendor feeds and proprietary data, over 100 redundant tables removed, and over 11,000 monthly queries across a 150-person firm. Queries are not active users, completed deals or investment returns. Undated source; adoption and evaluation periods unknown. Extends existing sponsor platform AI-052, separately from QB Energy’s field-planning workflow AI-054. No named asset, portfolio-operator rollout, task-time or accuracy baseline, investment-approval rule, full cost or measured return. The story’s investment-performance language is not evidence of investment alpha or AI attribution. Can the team trace a deal assumption to its source and show whether shared data reduces reconciliation and review work? [Transforming energy deal evaluation with Lakebase](https://www.databricks.com/customers/quantum-capital-group) · Unknown · Building a consistent foundation; Lakebase relational capabilities; adoption and FAQ paragraphs ## October 22 morning: equipment alerts, people and the FAST roadmap ### TC Energy TC Energy describes an equipment-surveillance process built around specialists and field teams. FAST is its consolidation plan with Accenture and AVEVA, separate from Pipeline IQ and compressor-dispatch optimization. Alerts pass through triage, specialist investigation, corrective action and documentation. The proposed platform joins existing tools, AVEVA and AWS. SageMaker model development is marked development-only; language assistants and deeper SAP coordination sit under future work. The presentation credits existing surveillance with over $100m/year in avoided failure costs. FAST’s additional $250m over ten years is a forecast. Neither is measured LLM value or all-in net ROI. Undated AVEVA World 2026 presentation, internally dated 29 May. Page 16 schedules Canada for Q3 2026 and US/Mexico for Q4; page 23 describes a first release in under seven months. The 16 October opinion does not resolve completion. The deck’s release wording is inconsistent; scheduled quarters do not prove go-live. Benefits lack a disclosed counterfactual, cost schedule, currency designation and measurement period. Do not add the forecast to the annual claim or assign either to FAST’s future AI features. What reached production, who acts on an alert, and what remains after false alarms, specialist time and deployment costs? [Practical AI in Action: TC Energy’s Path to a Unified, Scalable Asset Reliability Platform](https://cdn.mediavalet.com/eunl/content/W8TklEW5w0msg_eHg0KrTA/gz5XgM6QhE6XBVDXhH9-pA/Original/Practical%20AI%20in%20Action%3A%20TC%20Energy%27s%20Path%20to%20a%20Unified%2C%20Scalable%20Asset%20Reliability%20Platform.pdf) · Unknown · PDF pp.14–16: advisory workflow, upcoming FAST and release roadmap; p.17: architecture; p.20: forecast; p.22: future AI work; p.23: release wording [How the oil and gas industry is putting data first](https://www.aveva.com/en/our-industrial-life/type/opinion/how-the-oil-and-gas-industry-is-putting-data-first/) · 2026-10-16 · Posted date; Improved reliability through AI monitoring section ## October 22 additional: investment governance and an adjacent voice-agent beta ### Diamondback Energy An official AI Lead – Value and Economics posting calls for someone to test AI business cases and distinguish actual impact from correlation. The role would set baselines and comparison methods, challenge benefit assumptions, compare forecasts with realized value, and recommend whether projects continue, scale, change course or stop. It works with Finance, product owners and the AI portfolio lead. No project results, filled-role evidence, implemented measurement framework, cost or return is disclosed. The posting was available when checked on 22 October 2026. Its publication date and whether the role has been filled are unverified. A job description establishes intended responsibilities. It does not show that these controls exist today or that any project was stopped. Existing Diamondback operating initiatives retain their own evidence. Who can challenge an AI business case, and what evidence would make the company stop funding it? [AI Lead – Value and Economics · R100885](https://diamondbackenergy.wd12.myworkdayjobs.com/DBE/job/Midland-TX/AI-Lead---Value-and-Economics_R100885) · Unknown · Job Responsibilities: valuation framework, baselines/comparison groups, benefits attribution, continue/scale/pivot/stop recommendations. Official public jobPostingInfo payload retrieved. ### Italgas A vendor-hosted talk describes Italgas testing a voice agent on a subset of customer calls, with people handling complex requests. A real-time language model uses governed knowledge and authenticated CRM tools for appointment work. Databricks supports retrieval and tracing; Azure hosts the voice model. Broader channels are planned. Speakers claim 50% fewer dropped calls and 40% more automated resolutions in beta. Test size, period, metric definitions, quality, costs and net return are absent. Undated talk. Full call coverage was planned for September, with no verified year or completion. It is not confirmed as a fresh announcement. The transcript’s earlier baseline figures do not reconcile. The summary omits the beta boundary. Do not reconstruct absolute outcomes, infer full rollout or treat these claims as measured upstream results. Does the bot complete the right task, escalate safely, and reduce total effort after corrections? [Building Production Voice Agents with Databricks and AI Gateway](https://answers.databricks.com/building-production-voice-agents-databricks-ai-gateway-w8XJClcgjx4) · Unknown · Full transcript: 02:04 and09:38 baseline conflict;10:10–16:57 architecture and CRM;19:28–20:15 limited beta, planned September rollout and claimed changes;21:05 human operators. ## October 22 failure-focused: a pressure-model test limit ### SPE pressure-prediction study 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. The authors used 10 inputs and an 80/20 training/test split. They describe a small, noisy dataset; the abstract does not establish testing on entirely unseen wells. 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. 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. 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. Would performance hold on entirely new wells, with the full error distribution and an operating decision attached? [Evaluation of Influential Parameters for Prediction of Flowing Bottom Hole Pressure in Dry Gas Wells Using Machine Learning](https://onepetro.org/SPENAIC/proceedings-abstract/26NAIC/26NAIC/SPE-234889-MS/802231) · 2026-08-10 · SPE-234889-MS; publisher-indexed abstract: dataset/split and Random Forest/ANN comparison paragraphs. Published 10 August 2026; conference 3–5 August 2026. ## October 22 evening: completion decisions, installed pilots and internal tools ### Vista Energy Vista says its own AI monitoring tool helps engineers make completion-design decisions in real time in Vaca Muerta. The tool identifies frac hits and helps adjust stimulation to protect parent wells. The investor-day slide says it has been used since March 2025. Vista reports less than 1% of completion stages lost in 2025 and one runaway stage avoided per well on average. These are issuer-reported technical outcomes. The same slide prints a cost-saving unit as ~150 $M per well; that dollar amount is unresolved and excluded from comparisons. No net ROI is established. Investor day: 12 November 2025. Implementation since March 2025. Results labeled 2025; exact evaluation dates, well count and later continuation not disclosed. No matched baseline, well or pad population, effect-isolation method, tool costs or independent validation. Approval and execution rights are not disclosed. The schematic stage numbers are not a project population. The cost-saving unit is ambiguous even in the original image; no dollar benefit is normalized or counted. What changed in completion decisions, and do the avoided stages still pay after tool costs and engineering review? [Vista Investor Day 2025](https://www.vistaenergy.com/docs/vista-investor-day2025-presentation-1wbmozw.pdf) · 2025-11-12 · Slide 19 / PDF p. 19; title, implementation date and results column. Date corroborated by SEC exhibit heading. [Vista Investor Day 2025 — SEC exhibit](https://www.sec.gov/Archives/edgar/data/1762506/000119312525276082/d948063dex1.htm) · 2025-11-12 · Exhibit heading: November 12, 2025; same slide 19 presentation. ### Saipem Saipem says it installed an AI predictive-maintenance system on the Saipem 12000, the first drillship in its fleet to receive it. With ADC Energy, the pilot reads equipment data to flag anomalies and anticipate failures. A separate BIP project for diesel generators on Saipem 7000 was still being implemented and awaiting tests. No measured downtime reduction, failure forecast quality, intervention count, cost saving or ROI disclosed. One installed pilot is not a fleet deployment. Announced 23 February 2026. Pilot installation described; operating evaluation period and later fleet rollout unknown. Saipem 7000 testing was prospective. The issuer explicitly calls the 12000 project a pilot. Fleet extension and the separate 7000 project remain plans. Human maintenance approval, automatic execution, fallback rules and measured outcomes are undisclosed. Did the alerts lead to useful maintenance interventions, with fewer false alarms and enough avoided downtime to cover full costs? [Saipem introduces an AI-based predictive maintenance system onboard the Saipem 12000](https://www.saipem.com/en/media/news/Saipem-introduces-an-AI-based-predictive-maintenance-system-onboard-the-Saipem-12000) · 2026-02-23 · Page date; Saipem 12000 implementation, ADC Energy, pilot and separate Saipem 7000/BIP paragraphs. ### Worley Worley lists four deployed tools for vendor comparison, engineering-document checks, internal knowledge search and customer-response drafting. AI.VendorSelect supports technical-bid evaluation; AI.HoldsFinder extracts design holds; AI.Assist searches proprietary engineering knowledge; Smart Response Generator helps draft responses to queries and expressions of interest. Four tool names are a program scope, not four independently measured deployments. No active-user count, measured time saving, accuracy comparison, full cost or return is disclosed. FY2025 results presentation dated 27 August 2025. Individual go-live dates, usage duration and later adoption unknown. These are Worley’s internal tools across energy, chemicals and resources. No upstream-only user cohort or client field deployment is established. NVIDIA, Microsoft and Dell appear as general partner logos, not confirmed partners for each tool. No autonomous vendor selection, design approval or customer communication is established. Which work do employees complete with each tool, and how much review remains before a bid, design or response is approved? [FY2025 full-year results presentation](https://www.worley.com/-/media/files/worley/investors/results-and-presentations/2025/wor-presentation-full-year-results-fy2025.pdf) · 2025-08-27 · Slide 31 / PDF p. 31: recently deployed Advanced Development Lab solutions. Date on cover. ### Whitecap Resources A Shakudo customer story describes Whitecap’s governed data platform and business-led AI prototypes, with IT bringing selected work into production. The account describes relational data and operating telemetry, on-premise language models and selective frontier-model access. Human review remains part of the approach. The page’s less-than-one-hour analytics headline conflicts with narrative describing an endpoint still being built toward. It is not counted as achieved savings or ROI. Company production and team size are not an AI user population. Vendor index dated 23 April 2026; story itself undated. Platform use described, but project start and evaluation dates are unknown. Vendor account with a named customer executive, without independent operator corroboration. Predictive analytics and agentic work are prospective. The conflicting time claim, model quality, active use and full costs remain unresolved. Which prototypes have become routine work, and what does the production team measure before approving wider use? [Whitecap Resources customer story](https://www.shakudo.io/customers/whitecap-resources) · 2026-04-23 · Key Results; govern-first platform, AI operating model, human review and endpoint being built toward. Page is undated; date comes from vendor customer index. [Shakudo customer stories — Whitecap entry](https://www.shakudo.io/customers) · 2026-04-23 · Whitecap customer-story listing dated April 23, 2026. ### Technip Energies Technip Energies says its transformation plan had launched more than 55 digital and AI initiatives and delivered more than 20 minimum viable products by the end of 2025. The plan began in 2024. An AI-for-all campaign launched in 2025 aims to upskill staff; neither completed training nor routine use by all employees is established. The counts mix digital and AI work. Launched initiatives and delivered MVPs are not equivalent to production deployments, active users, measured savings or net ROI. 2025 annual report published in 2026; exact publication day unverified. Progress is explicitly at end-2025. The URL timestamp is not treated as a publication date. Only the indexed original p. 41 excerpt was reviewed; direct PDF access returned 403. Full-report definitions and later rollout remain unreviewed. No per-initiative deployment status, active-user denominator, measured benefit or full-cost return verified. The earlier DAP source dates that program to 2025; the annual report dates a broader transformation plan to 2024. Their exact relationship is unresolved. Which MVPs survived evaluation, reached routine use and produced a measured result after implementation and review costs? [Technip Energies 2025 Annual Report](https://investors.technipenergies.com/system/files-encrypted/nasdaq_kms/assets/2026/03/10/14-52-00/TEN_2025%20Annual%20Report.pdf) · 2026 (exact publication day unverified) · Printed p. 41, Value creation / We strive for excellence: digital transformation plan and end-2025 output counts. ## October 23 morning: planning, execution and supplier evidence ### Azule Energy SLB describes Azule using one digital workflow from well design through rig execution offshore Angola. DrillPlan passes the plan into DrillOps automation. The supplier case identifies AI-driven operations; it does not isolate AI from engineering automation. Supplier reports 95% of engineering tasks automated, planning time of 192 versus 31 hours and 75% fewer manual handoffs. A 2,000-meter section had zero reported nonproductive time and 16% higher net rate of penetration than the field benchmark. These are different measures, not a common financial return. Q3 release published 23 October 2026. The related case study is undated. One well and a 2,000-meter section are described; exact operating dates unknown. No well identifier, wider well population, benchmark construction, review-effort adjustment, quality comparison or full cost. Execution is described as autonomous, but approval, override and fallback rules are undisclosed. No LLM or AI-only contribution is established; zero NPT for this section is not a long-term reliability rate. Does passing an approved plan into rig execution reduce total work once checking, exceptions and overrides are counted? [SLB third-quarter 2026 results — selected operating and digital highlights](https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f) · 2026-10-23 · PDF p. 16: Digital / Azule Energy paragraph; same case and key measures. [Azule Energy pioneers digital well planning and rig automation in Africa](https://www.slb.com/resource-library/case-study-with-navigation/di/2026/azule-energy-pioneers-digital-well-planning-and-rig-automation-in-africa) · Unknown · Location; Results: Achieving operational excellence; Products Used: DrillOps AI-driven workflows. ### SLB / unnamed Permian operator SLB reports AI injection control and anomaly detection on nearly 1,400 producing Permian wells, with remote recalibration and weekly reviews. Analyst takeaway: measure dosage accuracy alongside cost, and ask who reviews exceptions. Reported on-target injection: 63% to 92%; overtreatment: 17% to 3%; $4.4 million in net costs avoided in the first nine months. The supplier does not define the net-cost calculation or disclose a full-cost ROI. Published 23 October 2026. First nine months of the program; exact start/end dates not stated. Unnamed operator: no company attribution, core adoption or result-count addition. No controls, cost reconciliation, model error rate or approval/override policy. Do not infer a specific Permian producer or treat the figure as audited net cash ROI. What costs were included, and did lower overtreatment persist without compromising corrosion or scale protection? [SLB third-quarter 2026 results — selected operating and digital highlights](https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f) · 2026-10-23 · PDF p. 14: Technology Highlights / United States production chemistry paragraph. ### Eni SLB says Eni expanded petrophysics and geomechanics AI workflows using an on-premise Lumi platform. Analyst takeaway: the location of the data platform matters when deciding what information a new assistant or model can access. No task-level benefit, asset population or full cost quantified in this update. Eni’s existing portfolio metrics and targets remain separately sourced; none is allocated to Lumi. Published 23 October 2026; exact deployment and expansion dates undisclosed. Supplier statement, without independent operator corroboration of this specific expansion. Model type, user count, data permissions and decision authority are undisclosed. No inference that this uses an LLM or captures all company communications. Which petrophysics or geomechanics decisions changed, and what data and review boundaries apply? [SLB third-quarter 2026 results — selected operating and digital highlights](https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f) · 2026-10-23 · PDF p. 16: Digital / SLB and Eni paragraph. ### Brunei Shell Petroleum BSP awarded SLB an on-premise contract for Petrel and Techlog, with digital and AI capabilities across subsurface and production work. Analyst takeaway: a broad contract establishes buying intent and implementation scope; it does not show which AI tasks employees complete. No measured operating or financial result disclosed. Contract scope is not a user population or proof of completed deployment. Published 23 October 2026. Award date, implementation timetable and active-use dates unknown. Supplier reports an award. Live AI workflows, algorithms, asset coverage, approvals and full cost remain unverified. BSP is separately identified; this award does not establish a Shell-wide rollout. Which workflows will go live first, and what acceptance test separates installed software from useful work? [SLB third-quarter 2026 results — selected operating and digital highlights](https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f) · 2026-10-23 · PDF p. 16: Digital / Brunei BSP software contract paragraph. ### ADNOC SLB’s Q3 release repeats that ADNOC’s AI-enabled operations center covers more than 120 onshore and offshore rigs. Analyst takeaway: this is another source for an existing footprint, not evidence that a new set of rigs was added. No additional rollout count or newly measured benefit. Earlier ADNOC claims retain their own source, period and limitations. Published 23 October 2026; earlier operator disclosure was 4 August 2026. No new commissioning date. Same supplier relationship and already covered footprint; not independent validation. Rig oversight and coordination do not establish autonomous drilling control on every rig. Do not transfer an individual well’s performance to the rig-center fleet. What is monitored centrally, what action can the platform take and who approves a change on a rig? [SLB third-quarter 2026 results — selected operating and digital highlights](https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f) · 2026-10-23 · PDF p. 16: Digital / ADNOC Real-Time Operations Center paragraph. ## October 23 evening: human review in supplier qualification ### Petrobras Petrobras describes AVAL IA Wave 2 reviewing supplier-registration documents before a technical evaluator checks them. The evaluation team records its decision and feeds it back to the AI. Petronect handles registration; the notice does not identify the AI model or developer. No accuracy, workload, completed-case count, cost, saving or ROI reported. Undated notice checked 23 October 2026. Forms are scheduled to open from September 2026, with documents due 30 days after each questionnaire opens; actual opening and evaluation dates are not confirmed. Technical evaluators keep the final check and decision. The notice says the pilot may pause for adjustments; it does not say it failed or was suspended. No LLM, autonomous approval or full rollout established. Does the first AI pass reduce total review work while preserving the quality of supplier decisions? [Cadastro de Fornecedores — AVAL IA Onda 2 pilot notice](https://canalfornecedor.petrobras.com.br/pt/cadastro-de-fornecedores/sobre-o-cadastro-de-fornecedores) · Unknown · Nova Fase do Projeto Piloto AVAL IA: O que muda / Prazos dessa fase. Portuguese original; September 2026 form schedule is not the publication date. ## October 24 morning: equipment alerts and enterprise foundations ### Halliburton Halliburton describes LOGIX Unit Vitality detecting early problems in cementing equipment so maintenance teams can investigate and plan repairs. Sensors feed more than 400 equipment parameters into cloud models. Staff check warnings, arrange parts and schedule corrective work. The product page separately describes transferring approved job designs from VIDA and iCem into controllers; this is not proof of autonomous maintenance. The supplier reports estimated downtime costs avoided of $125,000–$150,000 and 91% better maintenance productivity in four deployments. It does not define the productivity measure or say whether the dollar range is per deployment or combined. Article published 22 October 2026; product page undated. Deployment dates, customers, assets, basins and equipment population are undisclosed. No baseline, evaluation period, false-alarm rate, cash reconciliation, full cost or independent validation. Estimated avoided downtime is not booked savings or net ROI. No named customer adoption or new counted result. Separate from LOGIX drilling results in AI-016. Did the warning lead to the right intervention, and did avoided downtime exceed inspection, repair and system costs? [How predictive intelligence reduces downtime and improves equipment reliability](https://www.halliburton.com/en/energy-pulse/how-predictive-intelligence-reduces-downtime-improves-equipment-reliability) · 2026-10-22 · A simple question led to a new approach; How predictive reliability works; What early risk identification delivers. [LOGIX Unit Vitality — cement equipment health monitoring](https://www.halliburton.com/en/products/logix-unit-vitality) · Unknown · Automated cement equipment health assessment; AI-driven models; Real-time monitoring closes the loop. ### ONGC ONGC says Project ICE connects finance, materials, HR, production revenue and other functions on SAP S/4HANA. The operator reports more than 500 locations, 24,000 users and 23 modules. It describes a three-site disaster recovery setup and a SAP RISE journey over the next two to three years; that roadmap has no section-specific date. These are reported platform-scale figures, not active AI users, measured uptime, productivity, cost savings or ROI. Undated page checked 24 October 2026. The program began with SAP R/3 in 2004. S/4HANA transition date and the roadmap’s starting date are unknown. Historical non-AI context only. AI-enabled innovation is a roadmap aim; no specific live AI workflow or decision authority established. A shared platform does not by itself prove clean data, adoption quality or AI readiness. Can teams retrieve the same trusted records across functions, and which transactions still require manual reconciliation? [ONGC Technology — Project ICE](https://www.ongcindia.com/web/eng/about-ongc/technology) · Unknown · Project ICE, five paragraphs. Site-wide last-updated stamp does not establish when this section or its roadmap was published. ## October 24: SLB data and AI thesis SLB has a credible route to earning more from AI: it can put specialized models into software and operating workflows that customers already use. The public evidence supports a platform, engineering and governed-access advantage. It does not support treating every customer’s downhole data as an unrestricted SLB-owned training pool. 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. [Read the evidence and charts](https://energyai.sunya.ai/assets/slb-data-ai.html) ## October 24 evening: pump prediction and energy recovery ### Petroleum Development Oman In its 21 April 2026 results, Weatherford reported that Petroleum Development Oman had moved ESP Predictive Analytics from pilot to operational use. The system runs within ForeSite Well Management System and combines it with PetroVisor’s machine-learning framework. The on-premise installation provides failure forecasts, run-life estimates and real-time pump-performance insight while meeting the customer’s cybersecurity requirements. The release does not say whether forecasts trigger automatic changes or how people review alerts and approve maintenance. No pump or well count, predictive accuracy, warning lead time, run-life improvement, avoided downtime, cost or ROI is disclosed. Disclosed 21 April 2026, found in this 24 October review. The release says operating use had begun; its exact start date and duration are unknown. This is missed evidence from April, not a newly announced October deployment. 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. Ask which pumps and assets are live, how forecasts are validated, who acts on alerts, and whether failures or maintenance costs improved after full implementation costs. [Weatherford Announces First Quarter 2026 Results — PDO ESP analytics](https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18491) · 2026-04-21 · Technology Highlights → Production and Intervention; Oman / Petroleum Development Oman paragraph. ### Weatherford Weatherford reports the first commercial installation of its ForeSite Power Regenerative System retrofit on an existing U.S. rod-lift well. It captures and reuses energy generated during each pumping cycle. The supplier says the retrofit avoids replacement of the installed drive. This is energy recovery equipment; the release does not identify AI or a learning model in this installation. Weatherford reports over 50% energy savings on this installation. Metering method, comparison baseline, test duration and operating conditions are unspecified. No capex, total operating cost or payback is reported. Disclosed 20 October 2026 in Q3 results. Exact installation/measurement dates unknown. One unnamed well, not a fleet result. Anonymous single-installation supplier claim. Do not treat the percentage as company-wide energy savings, production uplift, net cash ROI or an AI result. The word ForeSite links a brand family, not proof this equipment used the PDO predictive model. Ask for metered before/after energy under comparable duty, uptime and load; retrofit cost and maintenance burden; and the well conditions under which the result transfers. [Weatherford Announces Third Quarter 2026 Results — rod-lift energy recovery](https://www.weatherford.com/investor-relations/investor-news-and-events/news/news-article/?ItemID=18871) · 2026-10-20 · Technology Highlights → Production and Intervention; first U.S. ForeSite Power Regenerative System retrofit paragraph. ## 24 October: false alarms and incomplete oversight These are operational lessons, not two new failed deployments. The MARS team changed its design and reached operational use; Havtil reports oversight gaps without naming a cancelled project. ### False alarms changed what the team could put into use 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. Preprint v2: 24 March 2026; first submitted 11 November 2025. Test uses 105 granules with known plumes and 95 without, containing 311 labeled events. Operational start dates differ by sensor. 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. Before scaling, how many false alarms and missed events will a team handle on the full live workload? [Operational machine learning for remote spectroscopic detection of CH4 point sources](https://arxiv.org/html/2511.07719v2) · 2026-03-24 · Version 2, 24 March 2026; sections 3.1, 5.1, 6.2 and Table 2; section 6.4 and Table 5. · 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. [Reducing False Positives in Operational Remote Sensing: Lessons from Methane Plume Detection](https://doi.org/10.1007/978-3-032-41086-3_46) · 2026-10-17 · Publisher abstract and First Online metadata; chapter pp. 721–736. Full subscription chapter not reviewed. · Editor read original publisher abstract and publication metadata. Subscription full chapter, experimental tables and code not obtained; no purchase or bypass. ### Some operators cannot yet account for all the AI they use 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. No respondent count, project denominator, incident rate, cost or return is disclosed on the reviewed page. Published 23 October 2026. Survey fieldwork dates and responding companies are not identified on the page; May 2025 guidance is earlier context. 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. Can the operating team name every AI system it relies on, including supplier features, and who owns its review? [Artificial intelligence (AI) in the petroleum industry is in need of better oversight and management](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/) · 2026-10-23 · Varying levels of maturity; An overview is essential for effective risk management; Prudent development and use. · Editor read original regulator publication and date. Underlying responses, named companies, response count and questionnaire not published on the reviewed page. ## 24 October: five SOL rounds on value, extraction and retrieval The West Brae account is an undated historical trial with a site-specific limited-value assessment. The other two notes are controlled tests, not evidence of cancelled operator AI systems. ### West Brae’s model worked, but the value case was weak 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. Undated historical account. Exact trial, publication and end dates unknown; not a new deployment in the rolling 24-month window. 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. Is there enough improvement left in the chosen asset to measure and pay for this tool? [Well Intelligence – Optimised Production](https://www.netzerotc.com/case-studies/well-intelligence-optimised-production/) · Unknown · Project overview; Industry value; Lessons learned. Undated historical account naming Marathon and West Brae. · 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. [Case Study: Enhancing Production with Well Intelligence AI Model](https://www.intelligentplant.com/case-study-well-intelligence) · Unknown · Challenge and Testing; Historical Data Review; Results and Impact. Anonymous client, no publication date. · 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. ### A well-record model picked the wrong depth field 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. Published 30 December 2024. Historical documents and 2024-era models; no prospective deployment date. 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. Does the test check the right depth, coordinates, units and source quality before values enter an operating record? [Information extraction from historical well records using a large language model](https://www.nature.com/articles/s41598-024-81846-5) · 2024-12-30 · 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. · 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. ### Text search missed information in engineering diagrams 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. Author manuscript posted 25 March 2026; proceedings marked 2026, exact issue day unknown. 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. Can the system retrieve the right information from drawings and tables, as well as prose, before anyone relies on its answers? [Evaluating Chunking Strategies For Retrieval-Augmented Generation in Oil and Gas Enterprise Documents](https://arxiv.org/pdf/2603.24556) · 2026-03-25 · 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. · 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. ## 25 October: procurement and the systems behind everyday work ### PTTEP 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. 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. 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. Which steps changed, who approves orders, and do the time gains persist after all operating costs? [PTTEP 56-1 One Report 2025 — Procurement X.Bot](https://www.pttep.com/en/document/viewer/stream/187064/56-1-one-report-2025) · Unknown · 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. [PTTEP organizes OpEx Day 2025, advancing operational excellence](https://www.pttep.com/en/newsroom/press-releases/1065/pttep-organizes-opex-day-2025-advancing-operational-excellence) · 2025-07-09 · Press release, procurement AI sentence in operational-excellence examples. ### BKV Corporation 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. Go-live 19 November 2025. Report letter is dated September 2026; exact report publication day unknown. 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. Does the common data model survive the next acquisition, and how much ongoing support does it need? [BKV 2025 Sustainability Report — Project Runway](https://www.bkv.com/2025-sustainability-report-governance) · Unknown · Governance: Project Runway; Informed by field operations; Collaboration beyond implementation. [BKV 2025 Sustainability Report — letter date](https://www.bkv.com/2025-sustainability-report) · Unknown · Letter from CEO footer: September 2026. ### Harbour Energy 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. Guide version 4, June 2025; exact day, system start and entity rollout unknown. 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. Where do invoices stall, which exceptions require people, and which supplier channel applies? [Harbour Energy Coupa Supplier Portal FAQ, version 4](https://www.harbourenergy.com/media/jbznhqcy/10-coupa-faq-english-harbour.pdf) · 2025-06 · 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. [Harbour vendor relations — distinct supplier channels](https://www.harbourenergy.com/vendor-relations/) · Unknown · Separate sections: vendors using SAP Business Network; ex-Wintershall Dea vendors. ### ConocoPhillips 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. Updated December 2024; January 2025 transition target. Undated live guidance still uses that future-tense wording; completion by business unit unknown. 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. Which units and invoice routes are live, and who owns the exceptions? [ConocoPhillips GEP SMART Introduction](https://static.conocophillips.com/files/resources/gep-smart-introduction_rusa-bu-website.pdf) · 2024-12 · PDF p.1 updated December 2024; p.3 transition target; p.5 intended benefits; p.6 supplier invoice flow. [ConocoPhillips technical upgrades to supplier systems](https://vendors.conocophillips.com/supplier-guide/technical-upgrades-to-supplier-systems/) · Unknown · GEP portal/EDI and Enverus routes; supplier training and implemented SSIS support. ## 25 October evening: three different kinds of evidence ### Ecopetrol reports IT gains while its wider AI program still faces data limits 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. The report covers 2025; exact publication day and AIOps start/evaluation dates are unknown. The 2027 target is prospective. 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. Which services improved, what changed against the prior baseline, and how much of the saving remains after running costs? [Ecopetrol 2025 Integrated Management Report — digital transformation and AIOps](https://files.ecopetrol.com.co/web/esp/aga2026/informe-integrado-gestion-informe-anual-gc.pdf) · 2026 (exact publication day unknown) · Printed p.314 / physical PDF p.315 of 482, Transformación Digital. AIOps, self-service and AI maturity are separate passages. ### YPF’s fuel-blending tool supplies estimates to refinery operators 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. 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. 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. Can YPF show achieved blend-time and quality results over enough batches to separate the tool from other operating changes? [YPF refinery blending — interview with BCG project lead Roberto Rodrigo](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) · Unknown · Sections Del laboratorio a la operación en tiempo real and Argentina, el primer caso de implementación. ### Dixstone selects shared project and fleet software; go-live remains unverified 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. Selection announcement dated 9 December 2025. Go-live and implementation milestones are not disclosed. 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. Which assets and workflows are live, what still needs implementation, and where does AI actually change the work? [Dixstone selects IFS Cloud for offshore services](https://www.ifs.com/pl/insights/news/dixstone-selects-ifs-cloud-to-power-global-offshore-operations) · 2025-12-09 · Dated release: selection, project controls and fleet maintenance paragraphs; About Dixstone separates affiliate status from Perenco E&P. ## 25 October: measured operating changes ### Perenco reports faster drilling across three wells 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. 2025 three-well campaign. Exact publication day and campaign dates unknown. Raw drilling-speed baseline, matched offset population, complete costs and AI-only contribution unknown. Did the gain hold on more wells after accounting for geology, equipment and full deployment costs? [Perenco enhances drilling efficiency with advanced automation solutions](https://www.slb.com/resource-library/case-study-with-navigation/di/perenco-enhances-drilling-efficiency-with-advanced-automation-solutions) · Unknown · Summary and Results; three-well 2025 campaign. ### Invoice handling fell from about five minutes to under one 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. Publication and operating calendar dates unknown; results observed within weeks of go-live. Twelve weeks describes implementation, not the evaluation period. Invoice volume, automated-scenario share, precise measurement window and full costs unknown. This is the Pakistan Shell-brand licensee, not Shell-wide adoption. What share of invoices qualifies, and how much staff time is actually released after exception handling? [From manual accounts payable to AI-powered finance in 12 weeks](https://spendconsole.ai/success-stories/from-manual-accounts-payable-to-ai-powered-finance-in-12-weeks/) · Unknown · Client profile; solution; results; automated-scenario qualification. ### A chemical-treatment redesign cut consumption without identified AI 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. Thirty-day observation; publication and operating calendar dates unknown. Population, throughput, cost calculation, installation costs and longer-term durability unknown. New chemistry and equipment prevent attributing the result to software alone. How much could treatment redesign achieve before adding AI, and what does it cost to sustain? [HSV-01 chemical treatment case study](https://apolloresourcesllc.net/apollo-resources-hsv-01-oil-treating-chemistry-effectively-eliminates-30000-ppm-of-h2s-in-senm/) · Unknown · Solution and Results; 30-day observation. ### Targeted chemical placement is another route to less waste 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. Paper dated 16 September 2025. Trial dates and evaluation duration unknown. Operator, geography, well population, raw baseline, cost perimeter and installation costs unknown. This does not establish all-in ROI or scaled deployment. Does targeted placement outperform surface optimization after installation and maintenance costs? [Maximizing chemical efficiency with rigless precision-based capillary injection](https://www.weatherford.com/documents/papers/2025/maximizing-chemical-efficiency-and-reducing-environmental-footprint-with-rigless,-precision-based-ca/) · 2025-09-16 · Public abstract, SPE-227643-MS; field-trial results. ### Better chemical dosing reduced waste across an existing well fleet 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. 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. 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. What costs were included, and did lower overtreatment persist without compromising corrosion or scale protection? [Permian field chemical automation case study](https://www.slb.com/resource-library/case-study-with-navigation/production-chemicals/permian-basin-field-automations-cs) · Unknown · Results and solution sections; 20-month dosing comparison and first-nine-month savings are different periods. [SLB third-quarter 2026 results — selected operating and digital highlights](https://investorcenter.slb.com/static-files/7045f96e-ee28-4096-934c-49eb4b37e66f) · 2026-10-23 · PDF p. 14: Technology Highlights / United States production chemistry paragraph. ## 26 October: results and setbacks ### A title model needed attorney-grounded regression tests 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. Publication, operating and evaluation dates unknown. 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. How does the error rate hold on unseen deeds, and who signs off on ownership conclusions? [Mineral-title evaluation and regression testing](https://www.toptal.com/case-study/oil-gas-tech-firm-improves-ai-mineral-title-accuracy) · Unknown · Challenge; Ground-truth AI Evaluation; Outcome. No publication/evaluation date displayed. ### High overall accuracy hid weaker rare-event detection 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. Application published 5 March 2026; exact underlying evaluation dates unknown. 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. Does the evaluation give rare, costly events enough weight? [AI approach for drilling operation insights — US 20260063025 A 1](https://patents.google.com/patent/US20260063025A1/en) · 2026-03-05 · Table 15 and section 5, Model Performance Comparative Analysis; published application, not a granted-patent or commercial-deployment claim. ### Complex plant drawings were harder than simple ones 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. Preprint 5 November 2025; evaluation and operating calendar dates unknown. 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. Which drawings still need manual reconstruction, and how long does correction take? [Case Study: Transformer-based Solution for the Automatic Digitization of Gas Plants](https://arxiv.org/html/2511.08609v1) · 2025-11-05 · Sections 3.1,4(Tables 1–3) and 5; top 100 relationship recall and complex-system limitation. ### One useful warning model did not make every coker use case useful 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. 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. 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. Did the prediction answer the actual operating decision? [Honeywell AI pilot aids coker unit operations at TotalEnergies refinery](https://www.controlglobal.com/show-coverage/honeywell-users-group/article/55383668/honeywell-ai-pilot-aids-coker-unit-operations-at-totalenergies-refinery) · 2026-06-11 · Steam perturbance, drum cycling, heater spalling and alarm-decision-support sections; Thomas Klein remarks.