Create: → A sensitive info type

Element: → Keyword dictionary

You want to classify documents in SharePoint Online based on a list of customer names stored in Customer.csv (1,000 entries).

In Microsoft Purview, there are several options for custom classification:

Sensitive info type (SIT):

Designed to detect predefined or custom patterns such as credit cards, IDs, or custom lists.

You can build a custom SIT using elements like keyword dictionary, regular expressions, and functions.

In this case, since you have a large CSV with 1,000 customer names, you upload that list as a keyword dictionary element.

Trainable classifier:

Used when you want AI to classify documents by training on examples of positive/negative samples (like contracts, resumes).

Not relevant here because you already have a structured list of customer names rather than needing machine learning classification.

Adaptive scope:

Determines where (users, groups, sites) policies apply, not what content is classified.

Not relevant here.

Correct pairing:

Create: A sensitive info type (SIT).

Element: Keyword dictionary (upload the Customer.csv file).