Fal.Con

Breakout times have collapsed to minutes — but the fastest insider in your environment may be one you deployed on purpose. Hundreds of enterprises traced production incidents to AI agents last quarter. None involved an external attacker. An agent with tool access is an insider that never sleeps, chains actions faster than humans can review, and looks nothing like malware when it goes wrong. In this fireside, Tal Peretz and Nancy Wang as they share real incident patterns from securing agentic tools at scale: what a drifting agent looks like in telemetry, why point-in-time approval fails, and agents caught bypassing their own restrictions.

The Invisible AI Stack on Your Endpoints, and How to Bring It Into Falcon

Falcon protects your fleet, but a new class of risk lives in a layer no endpoint sensor was designed to read: static configuration. AI agents, MCP servers, skills files, and instruction overrides are spreading through user-space directories across your fleet,  granting AI systems access to databases, source code, and credentials with no approval, no audit log, no security review. In this session, you'll see how to continuously detect this class of shadow AI with Runlayer, and easily bring it into the Falcon platform for complete telemetry and remediation.

Join GuidePoint Security and Runlayer for an exclusive evening of conversation, cocktails and connection.

​An invitation-only dinner for CISOs, CTOs, Heads of AI, and technology leaders — a high-trust space to step away from the noise, the hype, and the vendor pitches.