Initial setup for AI classification

After importing my docs and emails into multiple databases I discovered the AI doesn’t classify across these databases and that groups with subgroups should be empty with all docs kept in the subgroups.

I will reorganize accordingly which will take many days. :open_mouth:

Are there any other high level overarching things I should know before investing a lot of time structuring these databases for the AI to give me what I want? More specifically:

Does the AI analyze what I’ve already imported and will offer classification suggestions based on what is already there or does it just work on what it see me moving into place? For example I sort receipts by year then by month within each year. For it to sort imported email or scanned receipts can I simply import/scan into my Financial database’s inbox and the AI will see the dates of the recipes and suggest or auto classify into the correct year and monthly subgroup?

The AI assistants look at document Content, not Name, Creation Date or other metadata.

In my databases with topical organization by scientific discipline, policy issue and legislation the Classify and See Also assistants are very helpful. There’s typically strong coherency in the vocabulary of documents in each group, which means they have distinctive patterns of the contextual relationships of group contents.

In my financial database, in which organization of receipts, invoices and banking and investment data is organized by type and date, I wouldn’t rely on the AI assistants, and for that matter don’t need them, as I know where to file each new document.