As of May 1, Microsoft Agent 365 is generally available — and it answers a question that’s becoming more urgent by the day: who’s actually in control of all the AI agents moving into the enterprise?
What Agent 365 does
At its core, Agent 365 is a management platform for AI agents. Whether you’re running Claude-based agents, OpenAI Codex workflows, or local agents like OpenClaw on your machine — Agent 365 aims to bring them all under one roof. The platform extends Microsoft’s existing security and identity infrastructure (Entra, Intune, Defender) to cover AI agents.
In practical terms: IT teams can detect unsanctioned AI usage, restrict connections to approved destinations, filter risky file movements, and block prompt-based attacks before they lead to harmful actions. Sounds like enterprise jargon, but the problem is real — the more agents act autonomously, the more critical it becomes to have someone watching.
The multi-cloud angle
Here’s where it gets interesting: Agent 365 also syncs with AWS Bedrock and Google Cloud connections. IT teams can inventory agents across platform boundaries and, soon, manage their lifecycle — start, stop, delete. That’s a clear move toward cross-platform agent governance.
$15 per user per month
The price tag: $15 per user per month. Sounds like yet another line item on the growing Microsoft 365 bill, but for companies seriously working with AI agents, the alternative — having zero control — is considerably more expensive.
What this means for Claude users
If you’re using Claude in an enterprise setting, Agent 365 will likely be one of the first places where corporate IT can “see” and manage your Claude agents. That’s not necessarily a bad thing — it paves the way for broader adoption because IT departments finally have a tool to integrate AI agents into their existing governance frameworks.
The AI agent era just got an operating system. Whether it’s a good one or a constraining one remains to be seen.
Sources: Microsoft Security Blog, WinBuzzer, Microsoft Learn