Data security and governance
Protecting and governing data, including what AI can reach: Microsoft Purview for sensitivity labels, data loss prevention, insider risk, Data Security Posture Management (preview), Data Map and Unified Catalog, audit and eDiscovery, across Copilot, enterprise AI apps and third-party AI.
1 technology profile in this area
Overview
Microsoft Purview covers three areas. Data security includes Data Loss Prevention, Information Protection (sensitivity labels), Insider Risk Management and Data Security Investigations. Data governance includes Data Map and Unified Catalog. Compliance includes Audit, eDiscovery, Communication Compliance, Data Lifecycle Management, Records Management and Compliance Manager. Microsoft Data Security Posture Management (DSPM), which brings these together for apps and AI agents, is in preview.
AI raises the stakes because assistants surface whatever a user can open. Purview applies to Copilot experiences and agents, to enterprise AI apps such as Microsoft Foundry, ChatGPT Enterprise and Claude Enterprise, and to other AI sites detected through the browser. Prompts and responses go to the audit log, can be found with eDiscovery, and can be kept or deleted with retention policies.
For SharePoint oversharing, SharePoint Advanced Management and Restricted Content Discovery handle site-level fixes. Restricted SharePoint Search is retiring and has blocked new enablement since July 31, 2026.
An organization with little regulated data can start with sensitivity labels and basic data loss prevention before adopting the wider suite.123
Problems it addresses
Sensitive data pasted into AI sites2
Endpoint DLP can warn or block users who paste sensitive information into third-party generative AI sites in a browser.
Copilot returning protected files2
With encrypting sensitivity labels, Copilot returns content only when the user has both VIEW and EXTRACT rights.
Risky AI use by insiders2
Insider Risk Management includes a Risky AI usage policy template that detects prompt injection attempts and access to protected materials.
Legal holds on AI chats2
Prompts and responses are stored in mailboxes, so eDiscovery can search, hold and export them like other content.
Finding and trusting data1
Data Map and Unified Catalog support a federated approach to governing data across the estate.
A typical workflow
See how AI is used2
Open DSPM, or DSPM for AI (classic), to view AI activity and apply recommended one-click policies.
Label content2
Enable sensitivity labels for Office files in SharePoint and OneDrive so labels on stored files are honored.
Fix oversharing4
Use SharePoint Advanced Management reports, Restricted Access Control and Restricted Content Discovery on high-risk sites.
Block leaks2
Create DLP policies for Microsoft 365 services and onboarded Windows devices.
Retain and investigate2
Set retention policies for AI interactions and use Audit, Communication Compliance and eDiscovery for investigations.
Technologies
- Suite
Microsoft Purview
Microsoft Purview is Microsoft's set of data security, data governance and compliance solutions. It labels sensitive data, prevents leaks with DLP, detects insider risk, catalogs data estates, and handles audit, eDiscovery and retention, including for prompts and responses in Copilot, agents and other AI apps.
- Security
- Data
No technology in this area matches. Clear the filter to see all.
Use cases
Documented implementations
County of Kern
In productionCounty of Kern: protecting data with Purview before Copilot
over 13 million files (measured)
Government, United States (Kern County, California)
Chow Tai Fook Jewellery Group
In productionChow Tai Fook: 400+ AI agents for jewelry retail staff
over 400 agents (measured)
Retail and consumer goods, Hong Kong (story dateline); store network across several markets
Make-A-Wish America (national office and chapters)
In productionMake-A-Wish America: unified data, Copilot and an IT help bot
roughly two-thirds of staff (measured)
Nonprofit, United States
Copilot Readiness Check
Is your organization ready for Microsoft 365 Copilot, and what to fix first.
Next steps
Publish a short, clear set of sensitivity labels and train staff on when to use each.
Run the SharePoint Content Management Assessment before a broad Copilot rollout.
Decide which third-party AI sites to allow, warn on or block.
Use the Copilot Readiness Check to see which data controls to put in place first.
Connected areas
Sources
- Learn about Microsoft Purview (Microsoft Learn) (opens in a new tab)
- Microsoft Purview data security and compliance protections for Microsoft 365 Copilot and other generative AI apps (opens in a new tab)
- Restricted SharePoint Search (opens in a new tab)
- Get ready for Microsoft Copilot with SharePoint Advanced Management (opens in a new tab)
Thank you. Your correction was sent.
The editors check it against the sources. If you left an email address, they may reply about it.