AI agents are quickly moving from experimentation to real work—but without strong identity controls, they can become powerful new blind spots. This session explores how to think about agent identity in a modern enterprise, using Agent 365 and Microsoft Entra ID as the lens for secure, governed, and auditable agent operations.
We’ll walk through the core identity patterns behind agentic systems: how agents can be represented with first-class identities, how access can be scoped through blueprints and least-privilege design, and how different execution models—interactive, autonomous, and digital-worker-style agents—change the identity conversation. We’ll also cover the practical security questions identity teams care about most: authentication flows, access control, governance, sponsor accountability, protection signals, and how to reduce risk as agents interact with apps, tools, and data.