AI agents are rapidly becoming active participants in business workflows, making decisions, taking actions, and accessing sensitive systems alongside humans and workloads. This shift is forcing organizations to rethink a foundational question: how do you securely control access when the user is an AI agent? As AI changes how access decisions are made, identities now span users, services, workloads, and AI agents operating at machine speed, exposing gaps between identity and network controls.  Join this interactive roundtable to explore how identity and access models must evolve to extend the controls used for the workforce to AI and machine identities. Together with your peers, we will discuss how organizations are defining agent identities, enforcing least-privilege access, establishing governance and accountability, and protecting against new threat vectors introduced by autonomous systems. Microsoft will share lessons learned from applying these principles internally, and participants are encouraged to bring their own experiences, challenges, and best practices to shape practical, real-world approaches to securing access in an increasingly agent-driven world.