Service Advisor Agent
Tech notes become customer updates a non-mechanic actually understands.
A tech finishes an inspection and jots down what they see: "front pads 3mm, rotors scored, rear pads 6mm ok, slow leak at the front left CV boot." That shorthand means nothing to a customer waiting on a text. This agent turns it into a clear update — what was found, why it matters, what's urgent versus what can wait, and a plain-language estimate range for anything beyond the original job — ready for the advisor to review and send. It never invents a diagnosis the tech didn't make, and it flags real safety concerns for a phone call instead of a text. What used to take an advisor 10 minutes of translating per car becomes a 90-second read-and-send.
What this agent takes off your plate
- Every customer update is written in plain English, not shop shorthand
- Additional work found during a job gets flagged clearly, without every message feeling like a sales pitch
- Urgent safety findings are marked for a phone call instead of buried in a text
- Service advisors save 10+ minutes per vehicle translating and typing updates
- A consistent update format customers start to recognize and trust
The stack
Real costs & ROI
- Claude Pro or ChatGPT Plus$20/mothe free tier's message limits get tight once you're running updates on every car, every day
- Shared notes doc (Google Docs or your shop management system)$0/momost shop systems already have a tech notes field; a shared Google Doc works if not
- Zapier or Make (optional automation)$0–20/mofree tier is fine for manual copy-paste; a paid tier only matters if you want the whole flow automated end to end
The math
A service advisor translating tech notes into customer-ready language spends roughly 10 minutes per vehicle on a moderately complex job — reading the notes, deciding what matters, typing it out clearly. At 8–12 cars a day, that's 80–120 minutes of advisor time daily, or $40–60/day at $30/hour, roughly $900–1,300/month. This agent cuts that translation time to under 2 minutes per vehicle (reading and approving a draft), recovering most of that time for actual customer conversations and upsell walkthroughs. Even at the top of the $20–40/mo cost range, the payback is measured in hours saved per day, not per month — this is one of the fastest-paying-for-itself agents in the shop.
Build steps
Build your plain-English translation guide
The agent needs a dictionary of your shop's shorthand and a rulebook for how to explain findings without either underselling risk or scaring the customer.
Data you need
Pull 5–10 real tech notes from recent ROs, along with whatever the customer-facing summary ended up being (if you have one) so the AI can see the gap between the two.
I run an auto repair shop and I want to build a translation guide so an AI can turn technician notes into plain-English customer updates.
Here are real technician notes, and if I have them, what the customer-facing message ended up being: [PASTE 5–10 EXAMPLES]
Common shorthand and jargon my techs use that customers won't understand: [LIST — e.g. "OE", "flagged 1.5", part numbers, abbreviations like "CV boot", "TB", "PS"]
Please build me:
1. **A jargon-to-plain-English dictionary** — 20–30 common terms my techs use and the plain version a customer would understand
2. **A severity framework** — three buckets: URGENT (address before driving further), SOON (address within weeks/a few thousand miles), MONITOR (worth watching, not urgent). Build it as a **threshold table I fill in myself**: one row per common measured item (front/rear pad thickness, rotor thickness vs. minimum spec, tread depth, battery CCA vs. rated, belt cracking, leak severity) with a column for the URGENT / SOON / MONITOR cutoffs. Leave the numbers blank — my techs and I will set them from our own standards and the manufacturer specs, and the AI must never fill them in.
3. **A tone rule** — matter-of-fact and specific ("your front brake pads measured [X]mm; our shop recommends replacing at [Y]mm") rather than vague or alarming ("your brakes are dangerously low!")
4. **A "what's included in this estimate" clarity rule** — customers should always know what they're being asked to approve, in dollars, before reading further detail
5. **3 example translations** — one URGENT, one SOON, one MONITOR — so I can see the format in action
Save this as my translation guide — every future prompt will reference it.Connect the tech notes to a shared input
Trigger → AI action → delivery. The honest version first: Tekmetric, Shop-Ware, and Mitchell 1 do not expose a "new inspection note" trigger in Zapier or Make, so most shops start with copy-paste and only automate the intake. Manual path (works day one): tech finishes the DVI → advisor copies the raw notes and the approved job line into a Claude or ChatGPT Project that has the Step 3 prompt saved as its instructions → draft comes back → advisor edits and sends from the shop's existing texting tool. Automated intake (Zapier): Trigger: Google Forms — New Form Response (a 3-field form on the shop tablet: RO number, original job, raw notes) → Action: ChatGPT (or Anthropic) by Zapier — Conversation, Step 3 prompt in Instructions, form fields in Message → Action: Google Sheets — Create Spreadsheet Row with the RO number and the draft → Action: Slack — Send Channel Message to #advisor-drafts (or Gmail — Send Email) → advisor reviews → advisor sends the text. Make users: Google Forms — Watch Responses → OpenAI/Anthropic — Create a Completion → Google Sheets — Add a Row → Slack — Create a Message. If your system supports email-on-inspection-complete, swap the trigger for Email by Zapier — New Inbound Email and parse the body; otherwise the form is the reliable input.
Data you need
You'll need access to your shop management system's notes/inspection field (for copy-paste), or a Google Form your techs are willing to actually fill in consistently. Ask your shop-system vendor whether they have a Zapier app or webhook before assuming a live trigger exists.
Agent Pack
Unlock the Auto Repair Shop Agent Pack.
One payment unlocks all 3 blueprints in this pack — build steps, the complete system prompt, guardrails, a test checklist, and the maintenance routine — for $29.
one-time, unlocks all 3 blueprints in this pack on this browser. Secure checkout via Stripe.
More in this pack
More for Auto Repair Shop
Estimate Follow-Up Agent
Every unapproved estimate gets chased on a schedule, not when someone remembers.
Review & Reputation Agent
A review request after every completed job, and a reply drafted for every review that comes back.