As AI adoption accelerates across investigations, compliance, and disputes functions, many organisations are moving faster on deployment than on governance. The result: growing pressure from regulators, boards, and internal stakeholders to not just use AI - but to prove it works, is controlled, and is defensible.

This interactive roundtable brings together senior legal, compliance, and investigations leaders to explore how organisations are moving from experimentation to operationalisation - balancing speed, efficiency, and risk. Together, we will discuss what “good” looks like in practice: from evaluating AI outputs and managing model risk, to ensuring auditability, transparency, and trust in high-stakes environments.

Rather than theory, this session focuses on real-world approaches to adopting AI in regulated environments - what’s working, what’s not, and where leading organisations are placing their bets.

 LLMs on bad data hallucinate with confidence. We unpack what "AI-ready" really means for UBO, sanctions, PEP and adverse-media data. Data structured for machines, not humans. Ten years of building this layer before the hype — and why the data, not the model, is the real differentiator in compliance AI.