Identity resilience has long meant protecting accounts, governing access, and recovering from breach—but the definition of “identity” has changed. AI agents are already inside your enterprise, booking meetings, processing invoices, querying databases, calling APIs, and making autonomous decisions, all without the usernames and passwords your traditional IAM stack was built to manage. They authenticate via OAuth tokens, API keys, service accounts, and MCP connections, are provisioned quickly, documented poorly, and are often governed not at all. When one is compromised, the blast radius can dwarf a human account breach.

The session draws on a year of running AI agents in production for podcast booking, content editing, and publishing workflows, combined with more than two decades of identity experience at SailPoint, AWS, and Saviynt, to map what’s actually breaking and how to fix it. Attendees will come away with a clear framework for extending IGA, PAM, and Zero Trust programs to cover non-human and agentic identities—what to inventory, how to scope entitlements, where lateral movement risk hides in your agent mesh, and what true resilience looks like when half your “workforce” never logs in.