Process an Invitation to Bid

An Invitation to Bid is the document that starts the whole lifecycle: a project owner or GC sends it, and you need its details (project name, when the bid is due, who to contact) captured the moment it arrives. This first step does exactly that, and it works on the free Filler and Lite tiers since you are only building a template and reading values off a file. The automation half (mailbox routing) is a Pro feature, but the template you build here is the same one Pro reuses.

Build the template

Open Templates, click + New template, name it Invitation to Bid, and click + Add in the document bar to attach a sample ITB PDF as the source tab. Add the fields you care about (a Project Name text field, a Bid Due Date date field, a Contact text field), and for each one click where it should land on the page so it gets a write zone. The full canvas mechanics, resizing, nudging, moving a zone between pages, live in placing fields on the page.

The reason this is the easiest step is that an ITB is a read-only intake: you are not generating anything, just pulling structured data out of a document the AI has never seen. The field type does the heavy lifting. Declaring Bid Due Date as a date field forces the AI to return a parseable date instead of whatever phrasing the sender used, and the build-a-template primer covers why types are load-bearing for the dashboard later.

Steer the AI on the date

Set Bid Due Date's Value Source to Extracted so the AI reads it off the file, then add a field-level AI instruction in that input's box. ITBs bury the deadline in prose ("proposals are due no later than..."), so a hint like "the date proposals must be submitted by, usually phrased 'bids due' or 'proposals due', not the project start date" steers the model to the right line. This is the lowest, most reliable instruction layer: a per-value rule shows only when that value extracts, so it does not dilute the signal the way a template-wide rule would. The five layers, how they merge, and when to use a one-off run instruction instead are detailed in AI extraction and instructions.

Automate the intake (Pro)

Once the template extracts cleanly, stop uploading ITBs by hand. A Gmail automation watches your inbox and routes every match into a mailbox, tagged with a category and run through this template's extraction. This is a Pro ("The Operation") feature because it depends on mailboxes and the cases pipeline, but the setup is short. Connect Gmail in Settings → Account, open the Automations drawer from the bottom of the workspace sidebar, and bind a Gmail trigger to a destination mailbox plus this template.

Picking the template on the Extract Data node auto-resolves the file category, so the standalone category selector hides; you do not tag the file twice. Keep your filter mutually exclusive (one subject pattern or one label per process), because only one automation processes each message and you cannot pick the winner on overlap. Test with a single forwarded ITB before pointing a high-volume label at it. The full intake model lives in automating document intake, and the trigger filters (subject, sender, label, file types) in triggers and rules.

Next: with ITBs flowing in and parsed, process a bid confirmation to close the loop on the documents you send back.