Linking automations to document sources
When you turn on Extract Data and pick a template, the automation links to that template's workspace document source — and the source's file category becomes the tag applied to every ingested file.
Last updated
When you turn on Extract Data and pick a template, the automation links to that template's workspace document source — and the source's file category becomes the tag applied to every ingested file. The source link is annotation, not file storage routing: files still land at the case's storage path, the source just supplies the category and tells the workspace which automations feed which document sources.
What you need
- A template in the workspace that's linked to at least one workspace document source.
How linking works
In the Extract Data node, you pick a template. The template is wired (via the source tabs in the document bar above its preview) to one or more workspace document sources — the PDF blueprints stored once at the workspace level. Each source carries a file category.
- If the template has exactly one linked source, the automation uses it automatically.
- If it has more than one, you pick the Incoming document type explicitly — that selection determines which source's category gets applied.
If you turn extraction off, the source link goes away too — pick a category directly on the Destination node instead.
Why category matters here
The category is the join key that connects the file to the rest of the workspace:
- The autofill picker only shows templates that consume that category.
- Send templates can attach the file by category.
- Case-detail visibility filters use it.
If neither the template's source nor an explicit Destination category resolves at run time, ingestion is skipped on purpose — uncategorized files would break those downstream filters.
The trigger ↔ destination relationship
- One automation has one trigger and one destination.
- One Gmail label or Drive folder can be referenced by multiple automations (rare; usually one per label or folder).
- The first automation whose trigger matches a message is the only one that runs — order matters.
If you need different processing for sub-types of the same source (different sender groups, different filename patterns), use multiple narrower automations rather than one broad one.