Your enterprise may have hundreds of AI agents operating right now — and most security teams have no idea. As non-human identities multiply and agents gain the ability to connect to APIs, MCP servers, and sensitive data, the zero trust discipline organizations spent years building is quietly being bypassed.

In this episode, Tim Currie sits down with Michael Caplan, CTO of In Balance IT, to unpack what agentic security actually looks like in practice — from gaining initial visibility to implementing the right guardrails at runtime.

What we cover:

• Why organizations consistently undercount the agents in their environment
• The difference between binary and contextual guardrails — and why both matter
• Non-human identities: how they're over-provisioned and how they get compromised
• Prompt inj ...