MC420058: Microsoft Purview | Information Protection: New out-of-the-box trainable classifiers with auto-labeling support

Updated October 31, 2022: We have updated the rollout timeline below. Thank you for your patience.

We’re rolling out several new out-of-the-box pre-trained classifiers for use within Microsoft Purview Information Protection, Data Lifecycle Management, and Data Loss Prevention solutions. These classifiers can be used to discover and auto-classify sensitive content across categories such as IP/trade secrets, healthcare, operations, financial information, and HR-related information.

This message is associated with Microsoft 365 Roadmap ID 98099.

Utilize several new pre-trained and ready-to-use trainable classifiers to discover and auto-classify sensitive content across multiple common business functions (such as finance, HR, healthcare, legal, procurement, and IT), as well as more granular sensitive document categories including IP/trade secrets such as product development files, strategic planning documents, A files, and personal financial information. These classifiers can be applied as conditions in Data Lifecycle Management retention policies and Data Loss Prevention policies.

When this will happen

Public preview (North American region only) rollout will begin in late August and is expected to be complete by mid-November (previously late September).

Standard rollout (all regions) will begin in mid-November (previously early October) and is expected to be complete by end of December (previously end of October).

How this will affect your organization

Out-of-the-box trainable classifiers are pre-trained using diverse and large numbers of real-world samples to provide broad coverage of multiple common business functions such as finance, legal, human resources, mergers and acquisitions, and marketing communications. In addition to broad categories, this release also includes many granular classifiers that can detect sensitive business and regulatory documents. Together, these classifiers can detect more than 100+ sensitive content categories.

• In Microsoft Purview Information Protection, these classifiers can be added to server-side auto-labeling policies for files and emails to help you discover and auto-classify sensitive content across your Microsoft 365 data estate (SharePoint, Exchange, and OneDrive workloads).
• Microsoft Purview Data Lifecycle Management can use these trainable classifiers as a condition in server-side retention policies to detect and automatically apply retention labels on matching sensitive content in the same workloads mentioned above.
• The new trainable classifiers can also be added to Microsoft Purview Data Loss Prevention policies, and are supported across all DLP workloads, including SharePoint, OneDrive, Teams, Exchange, and Endpoint DLP.

The full list of available classifiers will be updated within technical documentation as rollout commences.

What you need to do to prepare

Get started with data classification in the Microsoft Purview compliance portal.

Learn more

• Learn about trainable classifiers
• Automatically apply a sensitivity label in Microsoft 365
• Automatically apply a retention label

Message ID: MC420058
Published: 26 August 2022
Updated: 31 October 2022
Platform: Web, World tenant