MC1280556 - Microsoft Purview for agents: AI observability and insider risk management now generally available

Introduction

As organizations deploy AI agents that can access enterprise data and take action on behalf of users, governing and securing agent activity becomes critical. Microsoft Purview will extend its data security and compliance capabilities to include AI agents, helping organizations gain visibility into agent behavior, assess risk, and apply consistent governance controls.

Microsoft Purview for agents will introduce Data Security Posture Management (DSPM) – AI Observability and Insider Risk Management for agents. DSPM helps organizations understand where sensitive data is, assess data-related risk, and continuously improve their security and compliance posture.

This message is associated with Microsoft 365 Roadmap ID 516032.

When this will happen

Public preview: Rollout started in December 2025 and is expected to complete in late April 2026.
General availability (Worldwide): Rollout will begin in early May 2026 and is expected to complete by late May 2026.

How this affects your organization

Who is affected

Microsoft 365 administrators responsible for security, compliance, and governance
Organizations using or planning to deploy AI agents
Tenants with Microsoft 365 E7 or Agent 365 subscriptions

What will happen

DSPM – AI Observability and Insider Risk Management for agents will move from preview to general availability. Learn more about activation.
With Microsoft 365 E7 or Agent 365 subscriptions, admins will be able to:
- Monitor and analyze AI agent activity across enterprise environments
- Identify potentially risky or non-compliant agent behavior
- Apply governance policies aligned with organizational requirements

Insider Risk Management for agents

Microsoft Purview Insider Risk Management correlates various signals to identify potential malicious or inadvertent insider risks, such as IP theft, data leakage, and security violations.
Admins will be able to create policies based on their own internal policies, governance, and organizational requirements.
The solution is built with privacy by design:
- Users are pseudonymized by default.
- Role-based access controls and audit logs help ensure user-level privacy.
These capabilities will not be available without Microsoft 365 E7 or Agent 365 subscriptions.
There will be no user impact unless AI agents are deployed and governed by admins.

What you can do to prepare

Review whether your organization currently uses or plans to deploy AI agents that access enterprise data.
Activate your Microsoft 365 E7 or Agent 365 subscription if you plan to use DSPM – AI Observability and Insider Risk Management for agents.
Review and update internal security, compliance, and AI governance policies.
Inform security and compliance stakeholders about this upcoming capability.

Learn more: Overview of Microsoft Agent 365 | Microsoft Agent 365 | Microsoft Learn

Compliance considerations

Question	Answer
Does the change introduce or significantly modify AI, ML, or agent capabilities that interact with or provide access to customer data?	Yes. This change will move DSPM – AI Observability and Insider Risk Management for agents to general availability, enabling monitoring and risk analysis of AI agent activity that can access and act on enterprise data.
Does the change alter how admins can monitor, report on, or demonstrate compliance activities?	Yes. Admins will be able to use Microsoft Purview to monitor AI agent behavior, correlate risk signals, and apply governance and Insider Risk Management policies for agents.
Does the change modify, interrupt, or disable audit logging capabilities?	Yes. Audit logs will be used as part of Insider Risk Management signal correlation for AI agent activity, supporting investigation and compliance workflows.
Does the change include an admin control?	Yes. Access to DSPM – AI Observability and Insider Risk Management for agents will be controlled through Microsoft 365 E7 or Agent 365 subscriptions and managed in Microsoft Purview.
Does the change alter how existing customer data is processed, stored, or accessed?	Yes. The change will introduce new observability and analysis of how AI agents access and act on existing customer data, without changing where the data is stored.