Microsoft Agent 365, the company’s centralized platform for observing, governing, and securing AI agents across the enterprise, reached general availability on May 1, 2026. The platform now includes native discovery and policy enforcement for OpenClaw agents running on Windows devices, multi-cloud agent registry sync, and a new Cloud PC environment built specifically for agentic workloads.
From Preview to Production
Agent 365 was first announced at Microsoft Ignite in November 2025 as a control plane for managing the growing population of AI agents inside enterprise environments. The GA release expands the platform’s scope beyond Microsoft-built agents to cover third-party and locally installed agents, starting with OpenClaw.
The platform is included in Microsoft 365 E7 licenses or available standalone at $15 per user per month, according to Microsoft’s announcement.
OpenClaw Agent Discovery and Blocking
The headline capability: Microsoft Defender and Intune can now detect OpenClaw agents running on managed Windows devices. IT administrators can see which devices are running OpenClaw, enumerate the agents deployed, and apply Intune policies to block common execution methods.
“To help organizations address accelerating agent sprawl and the rise of unmanaged agents, we’re introducing new capabilities as part of Agent 365 so you can discover local AI agents, and apply appropriate controls, such as blocking unmanaged agents,” Microsoft stated.
Support for GitHub Copilot CLI and Claude Code is planned next, according to Thurrott.
Starting in June 2026, Defender will add asset context mapping for each discovered agent: the devices it runs on, MCP servers configured, associated identities, and the cloud resources those identities can reach. Security teams will be able to use this context to assess exposure, investigate agent activity including file access and network behavior, and define custom detections. Runtime blocking of coding agents exhibiting malicious behavior patterns, such as data exfiltration attempts, will also enter public preview in June.
Multi-Cloud Registry and Partner Ecosystem
Agent 365 is not limited to Microsoft-built agents. The new agent registry sync, launching in public preview, lets Microsoft 365 admins discover agents built on AWS Bedrock and Google Gemini Enterprise Agent Platform, then import them into a unified registry for monitoring and governance.
Microsoft has also partnered with Adobe, SAP, Zendesk, and Manus to make their agents fully manageable through the Agent 365 platform, according to Thurrott.
Windows 365 for Agents
Alongside the GA release, Microsoft launched Windows 365 for Agents in public preview. The offering provides Cloud PCs purpose-built for running AI agents at enterprise scale, managed through Intune with the same identity, security, and compliance controls applied to employee devices.
The preview requires an Agent 365 license, an Intune license, and an active Azure subscription.
The Control Plane War
Microsoft is placing a bet that the enterprise agent management layer is as valuable as the agents themselves. By positioning Agent 365 as a vendor-neutral control plane that covers Microsoft, AWS, Google, and locally installed agents like OpenClaw, Microsoft is trying to own the governance layer regardless of which agent ecosystem wins.
The $15/user/month price point puts agent governance on par with existing security and compliance tooling. For organizations already running Microsoft 365 E7, agent management is bundled in. That pricing strategy mirrors how Microsoft captured endpoint management: make it cheap enough that building your own doesn’t make sense, then make it mandatory through compliance requirements.
The June preview of runtime blocking and MCP server mapping will be the real test. Discovery is table stakes. The question is whether Defender can enforce meaningful policy on agents that, by design, are built to act autonomously.