As teams in financial services increasingly use AI tools like ChatGPT in their day-to-day work, a key challenge is how to ensure the outputs are reliable.

This workshop focuses on the design of human-in-the-loop compliance AI workflows that can be relied on in operational settings.

You’ll learn how to:

• Design evaluation models that assess AI outputs for accuracy, completeness, and regulatory alignment
  
• Define human review points and design interventions that meaningfully improve reliability
  
• Use confidence scoring and guardrails to detect uncertainty, prevent drift, and constrain model behaviour
  
• Design for repeatability so outputs remain consistent across users, prompts, and time

Daniel Chatfield, Chief Information Officer (EU) at Monzo

Daniel Chatfield, Chief Information Officer (EU) joins the session to discuss how Monzo used AI to support regulatory work during rapid growth, the initial efficiency gains achieved, and the key barriers encountered, including sourcing, version accuracy, and reliability.

He will focus on where human review and judgement were required, how outputs were checked and escalated, and the practical lessons learned for building repeatable and scalable AI-enabled workflows.

Who should attend

Compliance leaders, policy teams, MLROs, risk and assurance professionals, audit teams, and anyone responsible for designing or overseeing AI-enabled compliance processes.