Why AI Agents Fail in Financial Decision Workflows - And How to Introduce Them Safely

​AI agents are generating enormous interest across financial services. The real challenge is not experimentation – it is introducing them safely into complex decision environments where governance, accountability and operational control cannot be compromised.

​In this discussion, Gavin Bennett (CTO, CreditLogic) and Eddie Dillon (CEO, CreditLogic) join Diana Alvarado (WWSO Sr. GenAI Specialist, AWS) to explore what changes occur when probabilistic AI enters deterministic decision systems and what decision infrastructure should look like in an agentic world.

​The session will focus on practical approaches to governing uncertainty, unblocking document and data bottlenecks and introducing agentic capabilities incrementally rather than through large transformation programmes.

What You’ll Learn

• ​Why agent initiatives often stall when introduced into real financial decision environments

• ​How probabilistic AI outputs change governance requirements

• ​Where agents can safely operate inside decision workflows

• ​How to introduce autonomy progressively while maintaining traceability and accountability

• ​Practical architecture considerations for scaling decision operations

​Sign up to watch live or receive a copy of the recording.