5 AI initiative profiles for ADNOC across 4 departments, with evidence grades and source links. Strongest evidence: B · Primary publication — Enterprise AI workforce training and agentic AI capability building.
B · Primary publication · Scaled · 8 fields · Production & field operations
complex tasks speed-up: up to 10x (Company-reported)
Source (2025-11-03)
B · Primary publication · Pilot · IT, data & enterprise transformation
accuracy improvement in major seismic interpretation aspects: 70% (Company-reported)
Source (2025-01-16)
B · Primary publication · Scaled · 40,000+ employees trained · Workforce & learning
Employees completing AI training: >40,000 (Company-reported aggregate); Monthly productivity hours gained: >70,000 hours/month (Company-reported; not isolated to training); Agentic AI training participants / agents created: >18,000 participants / >10,000 agents (Company-reported)
Source (2026)
B · Primary publication · Scaled · 120+ rigs · Drilling & completions
ADNOC claims 30–40% less engineering effort and capacity to oversee 2–3 times more rigs. It says the platform helps shorten incident response by 4–12 hours and avoid 1–2 days of downtime; observation period and frequency are unspecified.
Source (2026-08-04)
B · Primary publication · Scaled · 200 wells · Production & field operations
Abstract reports up to 2% oil-production increase and 5–30% lift-gas optimization at well level; these are not fleet-average results.
Source (2025-11-03)
Company and vendor statements are attributed, not independently validated. Evidence grades: A operator filing, B operator release, C vendor-reported, D secondary or unverified.
Later research and source checks
Supplementary notes can update existing programs or describe adjacent work. They are not extra adoption counts.
ADNOC Offshore’s drilling-report agent overflagged and sometimes needed overrides
A 2026 pilot at ADNOC Offshore tested whether software could reconcile rig sensor readings with daily drilling reports. The authors say the agent flagged too much in some instances, reviewers had to override it in others, and it sometimes added no value.
Five wells, two rigs, one offshore field. The abstract gives no frequency for excess flags, overrides or instances where the agent added no value.
Limits: The abstract documents review problems within the pilot. It gives no failure rate, net benefit, evidence of abandonment or confirmation of rollout. The same wells may appear in a separate offshore drilling-advice paper. The two samples are not added together, and no drilling-time result is attributed to report reconciliation.
SPE-236525-MS: Agentic AI in the Air-Gap at ADNOC Offshore · 2026-11-02
A two-well trial tested AI drilling recommendations in Abu Dhabi
Authors affiliated with ADNOC Onshore and Corva describe using machine learning to recommend drilling settings on two extended-reach wells in Abu Dhabi. The abstract does not name the field or well IDs.
The authors report more than 59 hours of on-bottom drilling time saved across both wells, 20% less than their top-quartile (P75) plan. They value the time at over US$208,000 in reported operational savings. Seven of eight sections beat the benchmark; one section reached 5,280 ft in 24 hours.
Limits: Only the original abstract was available. The plan construction, valuation, full costs and sustained performance are unverified. Treat the dollar figure as reported operational value, not net ROI. The trial’s overlap with the earlier AI-ROP collaboration or anonymous Corva cases is unresolved; do not add their figures or count a separate rollout.
SPE-235580-MS: AI-driven drilling optimization in extended-reach wells · 2026-11-02
ADNOC plans connected agents for field planning and commercial decisions
OpenAI and G42 will help connect field-development planning with operational, capital and commercial decisions.
Hours rather than months is a planning ambition, not an observed result. ADNOC separately reports existing ChatGPT Enterprise rollout to over 40,000 employees and adoption above 97%; the adoption definition and period are unspecified.
Limits: No costs, accepted plans, quality comparison, realized benefit or net ROI. Shared knowledge is planned; comprehensive communications capture and decision-outcome memory are not established. Existing workforce adoption is not Frontier adoption.
ADNOC announces OpenAI and G42 partnership · 2026-11-02
SLB repeats ADNOC’s rig-center scope, without a new rollout count
SLB’s Q3 release repeats that ADNOC’s AI-enabled operations center covers more than 120 onshore and offshore rigs.
No additional rollout count or newly measured benefit. Earlier ADNOC claims retain their own source, period and limitations.
Limits: 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.
SLB third-quarter 2026 results — selected operating and digital highlights · 2026-10-23