It would be nice to keep track of the percentage of the assets that are already classified, with descriptions, with Owners/Experts associated, etc. This can help Data Stewards to know which areas they should focus on when prioritizing classification review.

This feature, later on, could support workflow capabilities to automatically "certify" assets with higher degree of information fulfilled/reviewed.

Other than regular expression based classifications, modern data catalogs use machine learning to classify information. Purview should have that capability. Users should be able to plug built-in models and connect to custom models they develop and publish in Azure ML

Discovery & Classification integration with SQL Server/SSMS and Azure SQL/Synapse.

Suppose the customer has already put some effort on customize their labeling taxonomy columns (on SSMS our Security Centre) followed by the classification.

Will Purview integrate with this customization and recognize this classification or will need to configure Microsoftt 365 again?

Provide Healthcare specific classification from SNOMED. Systematized Nomenclature of Medicine (SNOMED) is a systematic, computer-processable collection of medical terms, in human and veterinary medicine, to provide codes, terms, synonyms and definitions which cover anatomy, diseases, findings, procedures, microorganisms, substances, etc.

Only the built in sensitive information types currently display in the Purview insights. Many customers use custom classification rules and would like to see these reflected in the insights reports

Often files have a naming convention that can be used to classify or label their content (such as this is a log file for application X, etc). It should be possible to classify and label file-based datasets by their name, similar to the ability to classify columnar data by the column name. The current file labeling based on office 365 is not a solution for this requirement.

Automatically find relationships between assets by correlating the schema and the data of the assets in the catalog.

When creating a custom classification rule, it should be possible to specify that the result of that specific classification is an entity in the business glossary

Customers would like the ability to make use of the trainable classifiers available in M365 to classify content with Purview

Many customers have already defined custom sensitive information types in the security and compliance center and would like to see these synced with Purview rather than needing to recreate custom classifications.

When defining a custom classification rules, the minimum count to trigger the custom classification (distinct count threshold) is currently 2. There are cases where even 1 occurrence of data matching a regular expression will identify the data as sensitive so this should be available.

The classifications in Purview overlap heavily with what is called Sensitive Information Types in almost all other M365 services. Although it is not a direct overlap, most of the classifications map to SITs. It would be good to have clarity on which of these are coming from M365 and which are unique to Purview.