You have a Microsoft 365 subscription.

Requirement

Correct selection

Documents containing content that matches a custom regular expression must be classified automatically.

A sensitive info type

Contract documents in a standard format must be classified automatically.

A trainable classifier

The first requirement must use a sensitive info type because Microsoft Purview sensitive information types are pattern-based classifiers. Microsoft states that sensitive information types detect sensitive content by using patterns, and custom sensitive information types can be created when the organization needs its own detection logic. A custom regular expression is exactly this kind of detection pattern, so it belongs in a custom sensitive information type rather than a trainable classifier or EDM schema.

The second requirement must use a trainable classifier. Contract documents in a standard business format are document-category classification problems, not exact-value matching problems. Microsoft Purview trainable classifiers are designed to recognize types of content by evaluating example content and can be used as conditions for auto-labeling Office files with sensitivity labels. Microsoft also provides built-in trainable classifiers for agreement and legal agreement content, including legal agreements, statements of work, leases, employment agreements, contracts, and nondisclosure agreements.

An exact data match schema is wrong here because EDM is used to match structured sensitive records from a data source, not to identify contract document format. A data connector is for ingesting external data into Microsoft Purview, not classifying Office documents.