Data Classification is well documented in the Imperva manuals.
One way it could help you is to identify sensitive data items that would may require additional monitoring.
You can use the output from a scan to setup table groups to use as additional criteria for policies.
For data classification you will need an account and privileges to the database you need to run a scan against. The list of privileges is in the documentation for various database types.
Then configure a database connection under the service in the site tree.
In Discovery & Classification - Scans Management
Configure a
scan profile and a
data classification scan to scan for the data types you are interested in.
The table group name is defined in the data type definition under
data types configurationOnce you run the scan, it may generate several thousand columns that need to be reviewed to find sensitive items and eliminate false positives.
Typically I have to export that list in a report and then send to the application owner to review as I do not know the structure of their databases.
Once an item is accepted, it will be added to a table group.