Uploading and finding your files

Your documents do not live next to your data, they are part of it. A contract is attached to the case it belongs to, and you find it from that case just as easily as from search.

Two ways to upload

From a record. The documents tab of a record holds everything that concerns it. This is the everyday move: you are in the case, you add the document to it.

From the drive. You upload first, you attach later. Useful when a batch of documents lands before you know which case each one belongs to.

One document, several files

A document is not always a single file. An ID has a front and a back; a contract has its signed original and its appendices.

A document gathers its files. You upload the front, then the back, then the amendment: they stay together under the same document instead of drifting apart with no visible connection.

That is also what lets your agents make sense of them: they reason on a complete document, not on a handful of files whose link they would have to guess.

What the platform reads

An uploaded document is not an opaque block: its content can be extracted by optical character recognition, and an agent can then use it in a conversation.

That text, however, does not go into the search index. Searching for a sentence contained in a PDF will not bring it up.

Limits to know about

Files are checked on upload, on their real type and not on their extension. A .pdf that is not one gets rejected.

Uploads are capped in number and in size, and the upload rate is limited. A large import goes through bulk import rather than through one upload at a time.