In high-stakes industries like oil and gas, understanding the origin and transformation of data is crucial. Just as a criminal investigation relies on reconstructing events to determine accountability, modern subsurface operations depend on accurate data lineage to maintain quality, compliance, and trust.
CGI explores this challenge with Agentic AI, which considers AI systems as operational agents with clear responsibilities, observable behaviour, and controlled access to data. These agents assist in tasks ranging from data engineering and subsurface interpretation to model evaluation and reporting, while automatically generating traceable lineage for both human- and machine-produced outputs.
Managing a growing workforce of AI agents requires clear governance. CGI’s Agentic Operational Framework (AOF) treats AI agents like human employees – assigning roles, controlling access, tracking activity, and allowing deactivation when needed. This ensures AI-driven processes are secure, auditable, and controlled.
The framework emphasizes data lineage, provenance, and policy-based access, creating a “digital passport” that makes data traceable, transparent, and trustworthy. Standardized artefacts like OSDU Data Definitions provide a consistent foundation for governance and auditability. Agentic AI enables organizations to integrate advanced AI into critical workflows while maintaining compliance and operational control.
CGI’s presentation, delivered by Michael van der Haven on Thursday (04 December), will highlight practical approaches to implementing Agentic AI and governance, offering insights for energy companies aiming to combine AI innovation with rigorous data stewardship.
The program can be found on the conference website.