Classification
The Classification policy uses AI to work out what each document is, such as an invoice, contract or resume, and tags its metadata with the matching types. Tagged documents are easier to file and search, and Routing can send each type to its own destination.
On a self-hosted server, Classification needs an administrator to set up AI with Document classification turned on (AI on your server).
Set up Classification#
- Open Processor → Pipelines and select Set up on Classification.
- Review the Classification labels. They are a built-in set of document types, shared across your whole team.
- Choose the documents to process. For editor files, Classification runs on Every upload by default.
- Select Create pipeline.
A document can receive several labels when more than one type fits.
Classify inside other pipelines#
Add a Classify step to any pipeline in the pipeline builder. Put it before delivery when you want to route documents by type. To route without AI, match on file properties instead; see Routing.
