Microsoft Launches Agent 365 to Manage Shadow AI Agents Across Enterprise Environments
What Happened — Microsoft announced the general availability of Agent 365, a cloud‑native service that discovers, inventories, and governs both SaaS‑based and on‑premise “shadow” AI agents such as OpenClaw and Claude Code. The platform adds integration points with Microsoft Sentinel, Defender for Cloud, and Azure Policy to enforce security and compliance controls.
Why It Matters for TPRM —
- Introduces a new third‑party risk vector: unmanaged AI agents that can exfiltrate data or execute malicious code.
- Provides a centralized control plane for continuous monitoring of AI workloads, helping organizations meet vendor‑risk policies.
- Aligns with emerging regulatory guidance on AI governance, making it a critical control for supply‑chain risk assessments.
Who Is Affected — Enterprises that use AI‑enabled SaaS tools, custom code assistants, or on‑premise AI models across any industry; particularly TECH_SAAS and CLOUD_INFRA customers of Microsoft.
Recommended Actions —
- Review current AI agent inventory and map any shadow agents to Agent 365 coverage.
- Validate that Agent 365 integration with existing SIEM/SOAR pipelines is enabled.
- Update third‑party risk questionnaires to include AI‑agent governance requirements.
Technical Notes — Agent 365 leverages Azure Resource Graph for discovery, integrates with Microsoft Defender for Cloud for policy enforcement, and supports API‑based onboarding of custom agents. No CVEs or vulnerabilities are disclosed; the release is a proactive risk‑management capability. Source: Microsoft Security Blog