Automating document intake
An automation watches Gmail or Google Drive and routes each matching file into a workspace mailbox: tagging it with a category, optionally extracting fields with a template, and creating or updating the matching case without manual upload.
Document Blueprint supports manual file upload. Automations let you skip manual uploads for incoming files. You configure an automation to watch a source. The app pulls the file in, reads it, and routes it. Unmatched, skipped, or failed items stay where they are.
Building an automation
To use file triggers, you turn on a connection in Settings > Account. Time triggers do not need a connection. Click Automations in the workspace sidebar. Click the plus button in the drawer header. Name the configuration in the Automation name field. You can also pick a color.
File automations start with two rows in the editor: the trigger row (named for its channel, for example Gmail trigger) and Mailbox routing. The plus beside Flow adds the optional steps the source supports: Extraction and Copy to Drive for Gmail, Extraction for Drive, and a Calendar reminder, Notify, Compose draft or Generate document step on any source. Case data and Date automations have no routing row; their reminder reads the watched key or the date field. When you add a reminder, its date field lists only the template's date-typed fields, and its title starts as {identifier}: {field:key}. Save appears in the editor header once something changed; click it when you finish. You can use Cancel on any row to discard edits.
Setting the trigger
The Trigger node sets what starts the run. Only one automation processes each overlapping Gmail message. You choose between four types.
- Gmail: Watches for incoming emails. You filter by Subject contains, Sender contains, Gmail label, and Read status. The status choices are All, Unread only, and Read only. All set filters must match. Gmail delivery is push-based.
- Drive: Watches a Google Drive folder. You supply a Folder ID *. The Folder name is display-only. The app ignores files inside subfolders. You can filter by File name contains and MIME type. Drive delivery uses a webhook.
- Time: Moves idle cases. You set From mailbox and After N days idle. The default is 30 days. You must pick different source and destination mailboxes. This runs on a daily schedule.
- Watch Case Data: Reacts when watched case fields change.
Gmail automations default to PDF, PNG, and JPEG file types. The File types allowlist exposes PDF, Images, and Custom options. An empty allowlist accepts all types. One email contributes at most 20 attachments; the run log notes how many were left out. Subject and file-name filters cannot contain quotes. The app filters attachments outside this list before ingestion. For Gmail, Capture From Email defaults to retaining Files. Turning Body off avoids storing full message text. The app always keeps sender and subject metadata.
Routing the file
The Destination node controls where the case lands. You pick a Mailbox to receive the case. A matching existing case moves to this mailbox.
When extraction is off, this node shows a Document selector. You associate the incoming file with a document type. The app skips ingestion if the category is missing.
Gmail automations can set a Drive copy folder name. The app copies ingested files to this folder. It creates the folder on first use. You can move the folder anywhere afterwards. This Drive mirror is best-effort. A copy failure logs an error but does not fail ingestion.
Extracting data with a template
The Extraction step is optional. Add it from the plus beside Flow, then select a Template for the automation. If the template has exactly one linked source, the app selects it. Multiple sources require you to choose the document type. This selection resolves the file routing and hides the standalone Document selector. You can click the Edit in template shortcut to open the editor. Click Done to commit the row draft.
A template with no ID field creates a fresh case for each file. The app marks the generated ID with an auto tag. This ID upgrades automatically when a run extracts a real identifier. When the template has an ID field, the app reads it. A match takes the file. No match starts a new case under that ID.
A capture panel lists delivery fields. You click + Capture from the email or + Capture from Drive. You add a field, name it, and select a part. The app adds these to the Data tab as template fields. Your template fields reading the email resolve whether or not you capture the body. A Message body field requires an instruction. It remains empty without one. If your template uses the other intake channel, the app adds a note. It tells you the automation leaves that field empty.
The Routing line states the grouping consequence of your ID choice. File name groups files yielding the same value. Folder name groups every file this automation ingests onto one case. Date parts arrive as date values. Keying on a date puts every file from that day onto one case. When no ID resolves, each Drive file keys its own case. Drive deduplicates a file ID within the resolved case.
How files move
Extraction does not happen automatically on every upload. A manual upload sits untouched until a template runs. The AI reads a file when an automation has an Extraction step. A teammate also triggers extraction by clicking Autofill, Extract, or Re-extract.
The extraction walks the template fields so results stay predictable. A teammate reviews the values in the case detail panel before generation. You trigger generation by clicking Generate or Generate from template.
When document content shifts between fills, turn on Document reflows in the Template details dropdown. This lets AI reposition fields at fill time. A template can also pull data From the delivery. This reads the email or Drive folder that carried the file. For a full worked example, the Tutorials build one end to end.
The app never sends email. You click Compose share to prepare an email. You click Open Gmail Draft. The draft lands in your Gmail, and you send it from there.
Past items and troubleshooting
Saved Gmail and Drive automations offer a Run on past emails or Run on past files section. This processes historical items from the source. Run never touches an item that is already in the app. When a finished Gmail run found such items, it asks whether to import them again: Overwrite, Save copies, or New cases.
The editor shows an Activity log. It identifies Processed, Skipped, and Error runs. It records the resolved case for each item. If a file lands on the wrong case, check the case Activity timeline. It identifies where the value was read and what it read as. It shows No ID read when nothing resolves.
You can correct the identifier from the timeline. A mismatch lets you move the file to a matching case. You can also start a new case from the file.
Automations can pause. The card shows an Auto-paused chip. A banner explains the reason. It might say Gmail is not connected. It might say Google Drive is not connected. It can also say Connected user removed from workspace. The app renews active connections automatically. You only reconnect when a banner requires it.
A file showed up in the wrong workspace. Check the automation trigger conditions like Subject contains or Folder ID. Check which Document the destination points at.