AI receptionist auto shop demo: full diagnostic + loaner in 90 seconds
Ninety seconds. That’s how long the scripted call below takes from ring to confirmation text — a full diagnostic appointment booked, a loaner car class reserved, a drop-off slot held, and a follow-up SMS on its way. Below is a turn-by-turn walkthrough of how an ai receptionist auto shop flow handles a real-feeling inbound call without putting the caller on hold, transferring them to a service writer, or promising a callback.
This is a scripted demo written to show how the script works — it is not a recording of a real shop’s call, and the “Caller” is not a real customer. Every line, timestamp, and data point below was written for illustration. The goal is to show the shape of the flow, the decisions the AI makes, and the shop-side integrations that have to be wired up for this to work live.
What an ai receptionist auto shop call actually does
Before the transcript, the quick mental model. The Bureau of Labor Statistics projects steady demand for auto repair techs through the decade, and ASE-certified shops keep reporting the same front-counter bottleneck: techs and service writers can’t stop what they’re doing to answer the phone cleanly every time. An inbound call handler that lives outside the front counter — and that can actually book, not just take a message — is the lever.
Four things have to be true for the 90-second call to happen:
- The AI can see the live bay schedule for the next 14 days, including diagnostic slot lengths.
- The AI can see the live loaner fleet and shuttle route status.
- The AI knows the shop’s policies — minimum ticket for a loaner, insurance rules, age limits — because they were encoded once at setup.
- The AI can send an SMS confirmation with the drop-off time, loaner class, and what to bring, through the shop’s messaging tool.
Without those four, you are not running an AI receptionist. You are running a slightly fancier voicemail. Our auto shop answering service breakdown goes deeper on that distinction.
The scripted demo call, turn by turn
The scenario: a caller’s check-engine light is on. They want to drop the car off Thursday and need a loaner for two days. Shop is a six-bay independent running Tekmetric and a four-vehicle loaner fleet.
Legend:
- [00:SS] = elapsed seconds from ring pickup
- AI = the InstaNexus AI receptionist
- Caller = the (scripted) customer
- Italicized lines in square brackets = what the system is doing under the hood
[00:00] — Ring + pickup
[System: call routed from shop’s published number to AI endpoint. Caller ID looked up against the DMS; no match, so this is treated as a new customer.]
AI: “Thanks for calling Riverside Auto. This is the booking assistant — I can set up a service appointment, check on a loaner, or take a message. How can I help?”
[00:06]
Caller: “Yeah, hi — my check-engine light came on yesterday and I’d like to bring it in. Thursday if you have anything.”
[System: intent classified as book_diagnostic. Required slots: vehicle, reason, preferred date, loaner need, contact.]
[00:12]
AI: “Got it — a check-engine diagnostic for Thursday. What year, make, and model is the vehicle, and whose name should I put it under?”
[00:19]
Caller: “It’s a 2019 Honda CR-V. Marcus Delgado.”
[System: vehicle parsed as 2019 / Honda / CR-V. Customer record created with name Marcus Delgado. Diagnostic slot length set to 90 minutes per shop’s default for CEL on this model.]
[00:25]
AI: “Thanks, Marcus. I have two Thursday openings — a 7:30 a.m. drop-off or a 9:15 a.m. drop-off. Which works better?”
[00:32]
Caller: “7:30 is fine. Actually — do you have a loaner? I’ll need something for Thursday and Friday.”
[System: loaner request added. Checks fleet module: two loaners returning Wednesday evening, both compact SUVs, available Thursday 7:30 a.m. onward. Matches shop’s minimum-ticket policy: diagnostic + likely repair clears threshold, so loaner eligible.]
[00:41]
AI: “Yes — I can hold a compact SUV for you starting Thursday at 7:30, for two days. Quick question: will anyone under 25 be driving it, and do you carry your own auto insurance?”
[00:50]
Caller: “Just me, I’m 38, and yeah, I have State Farm.”
[System: under-25 flag = false. Insurance on file = customer-provided, to be verified at drop-off. Loaner hold confirmed for 2019-class compact SUV, Thursday 7:30 a.m. through Friday 5:00 p.m.]
[00:56]
AI: “Perfect. I’ll hold it under your name. We’ll verify your insurance card at drop-off — please bring it along with your driver’s license. One last thing: is there a cell number where I can text the confirmation?”
[01:04]
Caller: “Sure — 555-212-7788.”
[System: SMS number saved. Confirmation template populated with appointment time, loaner class, drop-off instructions, and what-to-bring list.]
[01:10]
AI: “Got it. So: Thursday, 7:30 a.m. drop-off for a check-engine diagnostic on the 2019 CR-V, with a compact SUV loaner held for two days. I’ll text you a confirmation in the next thirty seconds with the address and drop-off instructions. Anything else I can help with?”
[01:20]
Caller: “No, that’s great, thanks.”
[01:22]
AI: “You’re welcome. See you Thursday, Marcus.”
[System: call wrap. Appointment written to Tekmetric. Loaner hold committed. SMS dispatched. Call summary + transcript posted to shop’s shared inbox for the service manager’s morning review.]
[01:30] — SMS delivered
Hi Marcus — confirmed for Thursday at 7:30 a.m. at Riverside Auto, 412 W. Main. We’ll run a check-engine diagnostic on your 2019 CR-V and have a compact SUV loaner ready at drop-off. Please bring your driver’s license and insurance card. Reply STOP to opt out.
That is the whole call. Ninety seconds of conversation plus an eight-second SMS handoff. No hold. No transfer. No promise the shop can’t keep — the AI only offered 7:30 after confirming the bay was open, and only held the loaner after confirming fleet availability.
Where the 90 seconds actually come from
Reading the transcript, the thing to notice is how little of it is social padding and how much of it is data routing. Map each turn back to a slot and you get a clean picture:
| Turn | Seconds | Slot captured | System action |
|---|---|---|---|
| Greeting | 0–6 | — | DMS lookup |
| Reason | 6–19 | reason, vehicle, name | Customer record created |
| Appointment | 19–32 | date, drop_off_time | Bay schedule queried + slot held |
| Loaner | 32–50 | loaner_needed, under_25, insurance | Fleet availability checked + policy validated |
| Contact | 50–64 | sms_number | Customer contact saved |
| Recap + SMS | 64–90 | — | Confirmation dispatched |
Human service writers can absolutely run this flow well. The challenge isn’t capability — it’s consistency across five service writers, three shifts, and a front counter that has a customer standing at it. The AI’s advantage is not intelligence. It’s that it runs the same clean script on call #1 and call #141, at 7:02 a.m. and at 6:58 p.m. For the cost math on what slipping that consistency actually adds up to, see what missed calls cost an independent auto shop.
What would break this demo in a real shop
A fair question: what are the failure modes for an ai receptionist auto shop flow once it meets real operating conditions? Three show up most often.
Stale loaner data. If porters aren’t logging returns in real time, the AI will book loaners that are still out. Fix the habit before you fix the software — no tool can outrun a loaner log that’s updated twice a day.
Missing policy encoding. “Do under-25 drivers get loaners?” needs a yes or no answer in the system prompt. If the AI has to guess, it will either over-promise or be annoyingly restrictive. Write the policies down; feed them in once.
Diagnostic scoping. Not every check-engine light is a 90-minute job. Some are a 4-hour tear-down. The AI should default to the safe slot length and flag the ticket for the service writer’s morning review — which the demo transcript does with the summary-to-shared-inbox step.
We compare this flow to alternatives — human answering services, call forwarding, voicemail — in our AI receptionist vs. answering service pillar. None clear the 90-second bar on a multi-slot booking the way a tightly scripted voice AI can.
The objection: “90 seconds is a marketing number, not a real call”
Fair. The 90-second framing is a scripted demo target, not an average. A realistic inbound mix — first-time callers, returning customers, pricing questions, “is my car ready yet?” status checks — averages longer for diagnostic bookings and shorter for status checks the AI answers directly from the DMS.
The point isn’t that every call finishes in 90 seconds. It’s to show the shape a clean booking call can take when the AI has the data it needs and the policies are already decided. An oil-change reschedule runs in 30–45 seconds. A towing-in-a-panic call ends with a warm handoff to the service manager. The AI’s job is to hit the right script for the shape of the call.
For more on how these scripts are structured, see our auto repair phone intake playbook and service writer AI deep dive. The auto repair vertical page covers the full inbound stack.
Frequently asked
Q: Is this actually a recording of a real shop’s call? A: No. This post is a scripted walkthrough, written to show the shape of the flow. Every timestamp, line of dialog, and customer detail is illustrative. A live demo on your own phone number uses your shop’s real schedule, loaner fleet, and policies.
Q: Does the AI sound like a robot? A: The voice layer uses modern text-to-speech that most callers can’t distinguish from a human on a short booking call. We disclose the AI on the first interaction per the shop’s preference — some owners want “this is our booking assistant” up front, others want it mentioned only when asked.
Q: What happens when the caller asks something outside the script? A: The AI either answers from its knowledge of the shop (hours, address, service menu, pricing ranges) or warm-hands to a human voicemail or mobile number with a transcript of what’s been captured so far. It never makes up an answer.
Q: How quickly can a shop get an ai receptionist auto shop flow like this live? A: With a Tekmetric-style DMS and a documented loaner policy, the typical setup is under a week. The bottleneck is not software — it’s writing down the policies and schedule defaults that aren’t already explicit.
Q: What does it cost compared to hiring a dispatcher? A: A fraction of a full-time front-counter hire, covering nights, weekends, and overflow. See our ai receptionist vs. hiring a dispatcher breakdown for the cost math.
See the 90-second demo on your own phone
The scripted demo above is the shape of the flow. The honest test is hearing it on a call with your shop’s schedule, your loaner fleet, and your policies plugged in. We’ll run a 15-minute live demo on a dedicated test number — you call it, you listen, you decide.