Microsoft Discovery Session: Protecting the Agentic Enterprise: From Discovery to Defense

As agents multiply, the awkward questions follow: which ones exist, what can they reach, and what happens when someone tricks one. Alym Rayani (VP, Microsoft Security) and Angela Byers (Senior Director, Microsoft Security) work through discovering agents, enforcing least-privilege access, and stopping sensitive data being overshared through AI interactions.

There's a demo covering AI-specific threats including prompt injection and model abuse, then a section on Security Copilot agents that map exposure, investigate live attacks and tighten access policies automatically.

Microsoft rate it intermediate (200 level), so it assumes some security background. It's aimed at security, identity and compliance decision makers and practitioners, and it's naturally all Microsoft tooling.