NorwayEnergy and utilities2025 – ongoingAker BP
Industrial AI agents for root cause analysis
Aker BP said AI agents cut engineers' time on equipment root cause analysis by more than 70%
Problem
When equipment on offshore oil and gas installations fails, reliability engineers run root cause analyses that require pulling maintenance records, sensor data and documents from many systems. The work is slow and repeats with every event.
Approach
In September 2025 the Norwegian oil producer expanded its partnership with industrial software company Cognite to use Cognite Atlas AI, a set of pre-built agents and a low-code workbench on the Cognite Data Fusion platform. Agents find relevant information, build cause maps and propose analysis steps; others process supply chain documents, support barrier management and register equipment automatically. A reliability engineer said other AI tools the company tested had fallen short.
Results
Takeaway
The agents searched an industrial data platform the company already used. The reported reduction of more than 70% covered root cause analysis; other agents were expected to save thousands of engineering hours.
Categories
- Industry
- Energy and utilities
- Business function
- Production and qualityLogistics and supply chain
- Innovation type
- Process innovation
- AI technology
- AI agentsGenerative AI
- Adoption
- Vendor solution
- Evidence
- Company's own claim
Sources
Source summaries
Aker BP and Cognite announce the expanded use of Atlas AI agents for root cause analysis, document processing, barrier management and equipment registration. The release says the root cause agents cut engineers' analysis time by more than 70%, and that equipment registration could save a significant number of engineering hours. It quotes the chief digital officer and a reliability engineer who says other AI tools fell short.
World Oil reports the expanded partnership and the companies' claim that agent-based root cause analysis cut engineering time by more than 70%. It lists current and planned uses, including predictive maintenance and production chemistry optimisation.
Full details are in the original sources.
