Jamf's AI governance tool names a problem you already have

On July 1, Jamf shipped AI Governance, a native, OS-level control plane for AI tools running on Mac fleets. The feature list is aimed at enterprise IT, but the problem it names lands on every small team that's let AI onto its laptops: you probably can't say what's running where.

What actually happened

Jamf's new capability does three things inside the Mac management platform admins already use. Visibility: shadow-AI discovery surfaces the AI tools, agents, and LLM runtimes across a fleet — including CLI-based developer tools and background agents — using its existing telemetry agent. Control: IT can define a sanctioned list of tools, push access policy at scale, and scope different postures to different teams. Governance: an executive AI posture report gives leadership a snapshot of who's using what, with SIEM compatibility for compliance.

At launch it recognizes Claude Code, Claude Desktop, and OpenAI Codex — the exact tools a modern dev team is quietly installing. Jamf also cites its own survey finding that organizations with deeply integrated AI are 40% more likely to report an incident than those still in the exploration phase. Adoption and exposure move together.

Why it matters for your business

You don't need a thousand Macs to have this problem. A five-person shop where two engineers run Claude Code, someone's pasting customer data into a chatbot, and a background agent has a token to your repo — that's shadow AI, and nobody wrote it down. The risk isn't the tools; it's that no one can answer "what has access to what."

You probably won't buy Jamf for a team that size. But the discipline is free, and it's the same discipline: know what AI is running on your machines, sanction a short list instead of pretending nothing's installed, and keep an audit trail so a security question has an answer that isn't a shrug. The tooling going enterprise-grade is the signal that this graduated from "nice to have" to "table stakes" — the same way device management and password managers did.

The operators who get burned won't be the ones who banned AI. They'll be the ones who never knew what they'd already let in.

Sources: PR Newswire (Jamf), Yahoo Finance.