10 booked jobs in the first week, on the calls they would have missed.

The problem we were asked to solve

Car Hub does what most independent shops do: takes inbound calls for diagnostics, brake work, suspension, tires, and the long tail of repair work that comes through Google searches and word of mouth. The owners are technicians first. When the phone rang while they were under a car, it usually went to voicemail. Some of those callers left a message. Most didn’t. They moved on to the next shop on the Google list.

The leak wasn’t marketing. They were ranking fine locally and getting plenty of inbound. The leak was the gap between the call coming in and somebody being free to answer it — and the gap between an after-hours form fill and the next morning, by which point the lead had already booked somewhere else.

What we installed

We kept their existing phone number and existing calendar. The AI receptionist sits between the two:

• Every unanswered call triggers an SMS within a few seconds: “Hey, this is Car Hub — saw we missed you. What can we help with?” The conversation continues over text.
• The agent is trained on Car Hub’s services, pricing ranges, service area, hours, and the questions that actually qualify a job (year/make/model, what’s wrong, how soon they need it).
• Once the lead is qualified, the agent offers two or three open time slots from the live calendar and books the appointment.
• A confirmation goes out by SMS, then a reminder the day before. After the job, the system asks for a Google review.

No rip-and-replace. The owners didn’t have to learn a new CRM, port a number, or change how they answer calls when they’re free. The system only fires when a human doesn’t.

What happened in week one

Ten jobs on the calendar that the shop wouldn’t have captured otherwise. Some were straightforward missed-call text-backs — caller didn’t leave a voicemail, would have been lost, instead got a text within seconds and was on the books ten minutes later. Others were late-night form fills from the website that were qualified and scheduled before the bay opened the next morning.

The bigger thing, harder to put a number on: the owners stopped spending the first hour of every day returning voicemails. The previous-day inbound was already triaged, and the calendar reflected reality.

What we’d do differently

Two things we tuned in week two and would recommend any shop bake in from day one:

• Tighter qualifying questions for diagnostics. Our first script asked too many questions before offering a time. Customers who knew exactly what they wanted (oil change, tire rotation) bounced. We pared the qualifying flow down to two questions for routine work and saved the deeper diagnostic triage for jobs that needed it.
• Owner-only escalation path. A handful of callers had complaints or were following up on existing jobs. We added an instant escalation: any message that looks like a complaint or references an existing repair gets forwarded to the owner’s personal phone with full context, instead of being routed through the booking flow.

Why this generalizes

Nothing about the install was unique to auto repair. The same workflow runs equally well for an HVAC company answering no-heat calls on a January Sunday, a roofer fielding storm-damage leads, a plumber handling burst-pipe emergencies, or a dental practice triaging new-patient inquiries. The pattern is the same: service business, single physical location or small team, phone that rings when nobody can answer, leads that move on if they don’t get a response inside ten minutes.

The AI doesn’t replace the owner. It replaces the gap between “phone rings” and “owner is free to answer it.”