A practical look at where point AI runs out of road, and what it takes to make agentic AI defensible enough for the exam room 

September 30, 2026 | 10 AM EDT | 3 PM BST | 4PM CEST 

KYC has quietly become the hardest place in the enterprise to get AI right. The volume is relentless, the moving parts are interconnected and every outcome carries consequences, from fines and exam findings to lost trust. It is exactly the kind of environment where automation should deliver the greatest value, yet many programs have hit a ceiling.  

This session explores how orchestration determines whether AI delivers meaningful results. We'll discuss the shift from point AI that automates individual tasks to governed, agentic AI that works across onboarding, screening, due diligence and ongoing monitoring as one continuous process. We'll also look at why a governed Agentic Layer, supported by shared data, policy controls, explainability and a unified audit trail, is essential in a regulated environment. 

Most importantly, we’ll focus on trust. Agents that operate without sufficient context, orchestration, governance or human-in-the-loop oversight create risk. We'll discuss the principles that distinguish regulator-ready agentic AI from disconnected automation and provide a practical roadmap for adopting out-of-the-box, custom and third-party agents without adding unnecessary risk. 

What you’ll take away: 

• Why disconnected, bolt-on AI reaches a ceiling and governed orchestration is what allows intelligence to compound 
• How the analyst’s role changes when agents handle data-gathering, freeing humans to focus on judgment 
• What moving from calendar-based periodic review to event-driven, continuous monitoring actually requires 
• A practical test for exam-ready agentic AI: full context, governed orchestration, explainability and human override at any point 

Speakers:

Speaker 1:

Nelson Cardoso

Senior Product Manager, KYC

NICE Actimize 

Speaker 2:

Ben Marsh

Global Head of Client Lifecycle Management, KYC

NICE Actimize