Automotive software for the shop, not the vehicle
ZhoopZhoop is our automotive build. AI voice agents answer inbound calls and dial suppliers, across four branches in Toronto.
Your shop bleeds money through two phones. The one ringing while everyone is under a car, and the one nobody has time to dial for parts pricing. ZhoopZhoop automated both directions, which is what made it worth building.
What this work involved
4 things that decide this
- 01ZhoopZhoop is a genuine automotive build for a multi-branch collision repair business in Toronto.
- 02Inbound AI voice agents book appointments into the real shop diary. Outbound agents call suppliers to collect and compare parts quotes.
- 03Every call, transcript, recording, quote and appointment lands on one dashboard across all branches.
- 04Deepgram handles speech and turn-taking so a caller can interrupt, which is what separates a voice agent from a phone menu.
Businesses where a missed call is a lost job
ZhoopZhoop
AI receptionist and parts procurement for a multi-branch auto repair business.
Read the case study →
Maidily
Integrated operations platform for residential cleaning businesses.
Read the case study →
WorkMateAI
Dispute resolution, payments, and compliance for Australia's on-demand trades.
Read the case study →
Three service businesses, three revenue leaks
A repair shop, a cleaning company and a trades marketplace lose jobs the same way. The software followed the leak in each case.
| Build | The leak | What we built | Where it runs |
|---|---|---|---|
| ZhoopZhoop | Both phones went unanswered during the rush. | Inbound and outbound AI voice agents on one dashboard. | Collision repair, Toronto |
| Maidily | Scheduling and payment chasing ate the day. | Booking, routing, invoicing and messaging in one place. | Residential cleaning, United States |
| WorkmateAI | Money moved between strangers with no rules. | Held payments, verification and staged disputes. | Trades marketplace, Australia |
- RingCustomer calls, nobody free.
- AnswerVoice agent takes the call.
- BookInto the real shop diary.
- Dial outAgent calls parts suppliers.
- CompareQuotes side by side, one screen.
Inbound and outbound are the same technology pointed in opposite directions.
Customers hate phone robots. Here is the honest version
Most people have been trapped in a menu tree and will assume yours is the same. The difference a caller notices in the first ten seconds is whether the thing can be interrupted and whether it hands over cleanly.
So the escalation path gets designed before the script does. A caller who has just been in a crash should reach a human fast, and the agent should not try to sound like one.
- Let the caller interrupt. Turn-taking is what separates an agent from an IVR.
- Write the handover trigger first, then the rest of the conversation.
- Keep the recording and transcript. You cannot improve a call you cannot replay.
An IVR menu against a voice agent
Both answer the phone. Only one of them books anything.
| Criterion | Phone menu | What we built |
|---|---|---|
| Caller interrupts | Ignored. The menu finishes its sentence. | Handled. The agent stops and listens. |
| Booking | Takes a message for someone to action later. | Writes into the live shop diary during the call. |
| Outbound | None. Staff still dial suppliers by hand. | Agents call suppliers and compare parts quotes. |
| When it fails | Caller hangs up and rings a competitor. | Escalates to a human with the transcript attached. |
| Cost of a mistake | A lost job nobody records. | A logged call your manager can replay and fix. |
The business around the vehicle
A repair business makes its margin on the phone, the diary and the parts bill. ZhoopZhoop automated all three, across four branches, on one dashboard.
The same shape shows up in every service business we build for. Revenue leaking through an unanswered call, and branches that each work slightly differently. Maidily and WorkmateAI attack the identical problem from a different trade.
- 01AI voice agents in both directions: inbound booking, outbound supplier quote collection.
- 02WhatsApp booking assistant handling the same work over chat.
- 03Multi-branch dashboards with every call, transcript, recording, quote and appointment in one place.
- 04Role-based admin so each branch keeps its own schedule and suppliers.
The stack behind ZhoopZhoop
Voice and messaging
Platform
AI
“They will treat your vision like their own and build it that way.”
Ron Klabunde · Founder, SmartREI ↗
What shop owners ask us
01Will an AI agent book into our real diary or just take messages?
Into the real diary. ZhoopZhoop's inbound agent writes the appointment during the call, against live shop availability. An agent that takes a message has moved the work rather than removed it.
02What happens when a caller is angry or has just crashed?
The agent escalates rather than improvises. That trigger gets designed before the conversation script. The failure you must avoid is a machine cheerfully offering a Tuesday slot to someone standing beside a wreck.
03Can software collect parts quotes without staff on the phone?
Yes, and that is the half most shops do not expect. Outbound agents call your suppliers, collect availability and pricing, then put the quotes side by side. Nobody in the shop spends the afternoon dialling.
04What part of an automotive business do you build for?
The business around the vehicle: the phone, the diary, the parts and the branches. That is where a repair operation makes and loses its margin, and it is what ZhoopZhoop automates end to end across four Toronto locations.
05We run four branches that each work differently. Can one system hold that?
It can, and ZhoopZhoop does. Each branch keeps its own schedule and suppliers while calls, transcripts and quotes roll up to one dashboard. Forcing four branches onto one identical process is usually where these projects stall.
06What happens when the AI gets something wrong on a live call?
You find out, because the recording and transcript are kept. That is the operational difference between a demo and a production system. The fix loop is reviewing real calls weekly and tightening the prompts and handover rules against what actually went wrong.
