Securing AI Agents: Identity, Permissions, and Safe Actions

Giving an AI agent access to business systems raises a central question: what should it be allowed to do? This panel explores agent identity, least-privilege access, tool permissions, approval checkpoints, and the handling of untrusted instructions. Discuss how to define action boundaries and prepare for unexpected behavior without making useful workflows impossible. Attendees will gain a practical perspective on managing the risks that arise when AI moves from generating answers to taking action.

Registration opening soon. Check back here for the registration link, speakers and start time.