Know why deals went cold. Fix them.

What it does

Diagnoses every stalled deal. One actionable recommendation.

1. 1
   Schedule Trigger — every day at 9am

   Cron: 0 9 * * *. Runs daily so reps walk into the morning with diagnoses on every deal that went quiet, before the trail gets cold.
2. 2
   Fetch Stalled Deals — HTTP Request

   Pulls every deal with no activity in 11+ days from your CRM (HubSpot, Salesforce, Pipedrive, Follow Up Boss). Each deal returns deal ID, owner email, contact email, contact name, deal value, last stage. Adjust the threshold in the URL.
3. 3
   Sample Deals Data (replace this)

   Two sample stalled deals let the workflow run out of the box. Delete this once Fetch Stalled Deals is wired up.
4. 4
   Split Deals

   Iterates through each stalled deal so the rest of the workflow runs once per deal. n8n handles the loop automatically.
5. 5
   Fetch Email Thread — Gmail

   Pulls the most recent thread between the rep and the contact. Searches by contact email. If you use Outlook or another mail provider, swap the node and adjust the filter.
6. 6
   Parse Thread + Recommend — OpenAI

   gpt-4o-mini reads the thread and outputs JSON: the cause (price_objection, timeline_issue, competitor_mention, response_delay_pattern, scope_creep, decision_maker_change, other), the specific quoted line that signals it, and a concrete next-action recommendation. ~$0.001 per deal.
7. 7
   Email Deal Owner — Gmail

   Sends one email per stalled deal to the deal's owner. Includes the contact name, deal value, days stalled, the AI's diagnosis, the quoted signal, and the recommended next action. Swap for a Slack DM if your team prefers Slack.

Use it in your business

Vertical-specific framing