Microsoft this month delivered security features that bring Anthropic Claude – an increasingly prominent AI model in the company’s strategy – under the visibility of Microsoft management tools, as well as new software to secure the execution of tasks by AI agents.
By extending visibility and control to Claude, the company is helping customers to more comprehensively secure customers’ agents, data, and identities. The integration provides visibility to Claude Enterprise and thereby helps tech and business leaders protect sensitive data across their AI estates – not just a single vendor’s tools and platforms.
Claude protections are being added to the Microsoft Purview security and governance platform; integration is powered by the Claude Compliance API, which Microsoft is supporting for centralized visibility and oversight into Claude Enterprise activity, including insights on Claude interactions and audit log signals.
The Claude Compliance API gives enterprise security and compliance teams using Purview programmatic access to two types of data:
- Conversation content – including chats, uploaded files, and projects — from Claude Enterprise. This allows admin teams to apply consistent security, monitoring, and Data Loss Prevention (DLP) policies to Claude, in keeping with their current tools and practices
- Events across both Claude Enterprise and the Claude Platform. This includes user logins, admin actions, and configuration changes so that security teams gain a unified view of how Claude is being used across the organization
Microsoft is one of 28 software developers (currently) supporting the Claude Compliance API; those companies span categories including DLP, data security, security operations, identity, AI security posture management, and more.
Secure AI Agent Execution
The company is moving to secure execution of actions by AI agents through the new Windows 365 for Agents. Now in public preview, Windows 365 for Agents works with Microsoft Agent 365 – the “control plane” for AI agents — to determine the work any agent is authorized to perform, while governing the agent’s access and behavior, using shared organizational policies and identity details.
Windows 365 for Agents enables organizations to define agent behavior, set up organizational policies, manage permissions, and maintain visibility into what agents are doing across the enterprise, whether those agents are built by Microsoft or third-parties. The company said Windows 365 for Agents is designed to help organizations move from early agent experiments to IT-managed, production-ready deployments.
Windows 365 for Agents defines where an agent executes the work; specifically, it works with Cloud PCs that enable agents to operate their own fully managed, auditable desktops and applications. The Cloud PCs serve as a purpose-built, IT-managed environment for running agents at enterprise scale. Managed by Microsoft Intune, the Cloud PCs bring the same identity, security, and compliance model IT already uses to provide agent execution oversight.
With Microsoft 365 for Agents, Microsoft is extending its same trust model and controls to AI; that’s an important development because IT and business leaders alike need to ensure AI operates securely and can be applied confidently to drive greater business productivity – within trusted security frameworks.
More Agent Governance Insights:
- Microsoft Governance Tools Ensure AI Agent Play Within The Rules at Runtime
- Microsoft Outlines Security, Governance, and Interoperability Features Coming to Agent 365
- Security Dashboard Strengthens Control Over Expanding AI Ecosystems
- Agent 365: The Platform That Keeps CIOs in Charge as Agents Proliferate
Ask Cloud Wars AI Agent about this analysis
