AI Agents, Governance, and the Security Gaps Most Organizations Don't See

Surprisingly, most organizations are trying to govern AI agents before they even know what they have. In a conversation between Chris Johnson and Ed Baker from Netrix, this critical blind spot takes center stage. As conference presenter Minty pointed out, visibility, ownership, and monitoring must come before governance, not after. What does it really mean to understand how your clients use AI? The answer might involve interviewing nurses, vets, and paralegals to uncover real-world usage patterns, because the gap between assumed and actual AI behavior is where risk lives. Ninety percent of MSPs now use AI, yet most focus narrowly on internal efficiency rather than comprehensive security. Ed warns about the "confused deputy" problem, where employees unknowingly spread incorrect AI-generated information and create organizational chaos. That chaos, that organizational confusion, can spiral quickly without proper guardrails. Ultimately, adding layers of tools without understanding their configurations can paradoxically worsen your security posture. If you work in managed services or security, this episode offers genuinely practical thinking you can act on.