Agency

Send home service shops weekly scorecards of their calls

Turn a plumbing, HVAC, or other home service shop's recorded phone calls into text with AI, then send the owner a weekly scorecard showing booking rate, missed upsells, and coaching notes for each person answering the phone.

The hard part: getting each shop's call recordings out of their phone system every week and writing scores the owner trusts enough to coach with.

Setup 3 daysRuns $60/moYour time A weekly check-inNeeds ClaudeDecision modeln8nAirtableStripe

Who pays

Home service shop owners pay a one-time setup fee and a monthly fee for the weekly scorecard. coreyganim suggests $500 setup plus $500 a month. That pricing is the poster's suggestion, not reported revenue.

How to build it

  1. Collect the shop's call recordings

    Find out where the shop's calls are recorded, such as CallRail or RingCentral, and set up a weekly export of the recordings with the rep's name on each.

  2. Turn calls into text

    Run the recordings through WhisperX, a free, open-source tool that turns speech into text and marks who is speaking. A rented cloud computer with a graphics card runs it fastest.

  3. Score each call

    Use a decision model, a fast AI that only makes yes/no or sorting calls, to mark each call: was the job booked, was an upsell offered. Use Claude to write two or three coaching notes per rep.

  4. Store scores in Airtable

    Save each call's scores and notes to Airtable so you can total booking rate and missed upsells per rep each week.

  5. Automate the weekly scorecard

    Use n8n to pull new recordings, run the scoring, and email the owner a one-page scorecard every Monday.

  6. Bill setup and monthly fees

    Charge the setup fee and the monthly plan through Stripe Payment Links.

Sources on X

  • coreyganim · Oct 19, 2026

    The poster lists WhisperX as a service offer: weekly call scorecards (booking rate, missed upsells, rep coaching) for home service shops, at $500 setup + $500/mo. The poster says the pricing estimates are purposely conservative and that the buyer only cares about the outcome, not AI or GitHub.

Ideas are curated from public posts and untested by us. Read the full note.