For years, AI in financial crime compliance meant faster rules, better models, and smarter alerts. Agentic AI is something categorically different. These are systems that do not just flag anomalies, they reason, plan, and act. They execute multi-step investigations, adapt to new information mid-process, and operate across systems and data sources without waiting for human instruction at every turn.

The question for financial institutions is no longer whether agentic AI will reshape financial crime defense. It is whether they will be ready when it does. This session cuts through the hype and examines what agentic AI actually means for financial crime teams, what it can do today, where it is still maturing, and how institutions are beginning to deploy it in production environments.

Discussion spans three dimensions practitioners are navigating right now: operational deployment, where the first real use cases are taking hold; implementation, where integration complexity and workforce gaps are slowing progress; and output quality, where institutions are asking hard questions about explainability and governance.