Govern Sanctioned Enterprise AI

Bring the AI that runs outside Microsoft's Copilot family — custom apps you build and sanctioned SaaS such as ChatGPT Enterprise and Anthropic Claude — under Microsoft Purview, so their interactions are discovered, audited, and governed like the rest of your estate rather than remaining a blind spot.

Architecture

Core capabilities

Purview DSPM AI Interaction Discovery Microsoft Purview Data Security Posture Management

Use DSPM for AI as the front door to discover enterprise AI interactions and detect whether sensitive data is flowing into their prompts and responses
Entra-registered AI App Governance Microsoft Purview

Bring custom, in-house AI apps under the full Purview control plane by registering them in Entra and integrating the Purview SDK, so classification, labels, DLP, and compliance controls apply
Enterprise AI App Connectors Microsoft Purview

Connect sanctioned non-Microsoft SaaS AI — ChatGPT Enterprise and Anthropic Claude — so their prompts and responses become discoverable and auditable in Purview

Supporting capabilities 5

AI and Agent Activity Audit Microsoft Purview

Capture enterprise AI prompts and responses in the unified audit log for investigation and forensics
Communication Compliance for AI Interactions Microsoft Purview

Detect risky or non-compliant content in enterprise AI interactions where the connected app supports it
Insider Risk Signals for AI Activity Microsoft Purview Insider Risk Management

Flag risky AI usage from these apps and correlate it with broader insider-risk signals
AI Interaction Retention Microsoft Purview Data Lifecycle Management

Retain or delete enterprise AI interactions in line with retention policy
AI Regulation Assessments Microsoft Purview Compliance Manager

Assess governance of these apps against AI regulatory frameworks

Architecture

Core capabilities

Solutions