IndiaEnergy and utilities2026 – ongoingOil and Natural Gas Corporation
Agent-assisted offshore well modeling with Pipesim
ONGC processed approximately 600 well models overnight with supervised AI automation
Problem
Engineers manually built and calibrated well models for offshore production studies. Preparing hundreds of models could take months.
Approach
ONGC connected a small-language-model agent to Pipesim through a custom Python library and a model context protocol server. The agent validated inputs and ran modeling workflows. Engineers supervised execution and reviewed exceptions.
Results
Takeaway
Engineers retained responsibility for validating inputs and interpreting outputs. Reported labor savings were estimates against manual workflows.
Categories
- Size
- Large enterprise
- Industry
- Energy and utilities
- Business function
- Production and qualityIT and R&D
- Innovation type
- Process innovation
- AI technology
- AI agentsGenerative AI
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
ONGC engineer Aman Sharma described two applications of supervised well-modeling automation. He reported overnight processing of approximately 600 wells and execution of 370 simulation scenarios in under 1 hour.
Full details are in the original sources.
