Home / Technology / Microsoft Unveils Agent 365: Taming the 'Shadow AI' Wild West
Microsoft Unveils Agent 365: Taming the 'Shadow AI' Wild West
5 May
Summary
- Agent 365 offers unified control for AI agents across diverse platforms.
- It addresses the emerging 'shadow AI' risk from unmanaged local agents.
- The platform detects and governs local AI agents on employee devices.
Microsoft has released Agent 365, a platform designed to manage AI agents, marking its transition from preview to general availability. This move underscores Microsoft's view that governing autonomous AI is a pressing, operational concern. Agent 365 acts as a unified control plane, allowing enterprise IT and security teams to monitor, govern, and secure AI agents wherever they operate.
The platform addresses the critical challenge of 'shadow AI,' which refers to autonomous tools and workflows employees install on their devices without IT oversight. Microsoft's new offering aims to discover and manage these local agents, mitigating new enterprise security risks. Agent 365 provides a centralized registry and policy engine, offering a single view of all agents within an organization's environment.
Key features include the ability to discover and manage local AI agents on Windows devices, initially focusing on agents like OpenClaw. Microsoft plans to expand this discovery to 18 agent types by June 2026. The platform also offers asset context mapping and policy-based controls, enabling administrators to set guardrails and block malicious agent behavior.
Agent 365 extends its governance capabilities to rival cloud platforms such as AWS Bedrock and Google Cloud. Additionally, it integrates with Microsoft Entra network controls to inspect agent traffic and block malicious attacks. For organizations hesitant to run agents on endpoints, Microsoft offers Windows 365 for Agents, a public preview for sandboxed agentic workloads.
