Source control records who committed code. It does not record who made the decisions behind it, and that gap is widening as AI agents take on more of the delivery process. This session explores how organisations can adapt accountability models for an AI-assisted development environment, arguing that everything downstream of an agent should be treated as AI-influenced and subject to a consistent verification approach. It examines the distinction between issues that automated controls can reliably detect and the judgement calls that require human ownership, clear accountability, and an auditable decision trail.

 

The session concludes with a practical framework for determining where human oversight remains essential, including a live demonstration of how teams can maintain meaningful human accountability even when development is increasingly driven by AI agents. Aimed at engineering leaders, CTOs, architects, and anyone navigating the changing relationship between AI, software delivery, and responsibility.

​