AI for Detailing Shops: A Crew, Not a Chatbot

Every software company selling AI for detailing shops slapped “AI” on a chatbot and called it a day. You know the one — the little bubble in the corner of a website that answers “how much is a full front?” and not much else. It's fine. It's also not doing the job that actually keeps a shop full.

Because in my shop, answering a question on the website is maybe two percent of the work. The other ninety-eight is follow-up. It's the quote I sent Tuesday that went quiet. It's the call I missed while I was squeegeeing a hood. It's the happy customer who just picked up their car and would leave a five-star review this morning if somebody asked — and nobody asks, because I'm already elbow-deep in the next job. One bored chatbot doesn't sell a full front, chase a quiet quote at 8pm, or catch a review at the exact moment someone's thrilled. Those are not one job. They're a dozen, and they're all different skills.

That's the whole problem with treating AI as a single feature to bolt on the side. Answering a brand-new lead is a completely different skill than collecting an overdue invoice. One is fast and warm and about not losing momentum; the other is patient and firm and about not burning the relationship. Ask one generic bot to do both and it does neither well. You wouldn't hand your reception desk, your sales follow-up, and your accounts-receivable to the same summer hire and expect all three to go smoothly. Software shouldn't get a pass on that just because it says “AI” on the box.

So Service VIN doesn't ship a chatbot. It ships a crew. Eleven specialist agents, each scoped to one job in your funnel, each doing that job the same way every time. They share your shop's data and your shop's voice, but they act independently — which is exactly why the crew covers the whole pipeline instead of parking on a website widget waiting to be asked a question.

Here's the roster, by the job each one owns. Think of it less like a feature list and more like a shift schedule — everybody has a post.

• Lead Responder — answers every new lead the second it lands from your site, a form, or a DM, with a warm on-brand text, so you never lose a job to a slow reply.
• Missed-Call Responder— texts back any call you couldn't grab, so a missed call becomes a live thread instead of a dead voicemail nobody checks.
• Quote Follow-up — nudges sent quotes that went quiet at the right cadence, and hands you back the warm ones ready to close.
• Install Reminder — confirms the appointment ahead of time and cuts the no-shows that leave your bay sitting empty.
• Review Requester— reaches out right after a finished install and routes happy customers to your Google profile, at the moment they're happiest.
• Invoice Chaser — watches unpaid invoices and sends one friendly nudge with the payment link once one has sat too long — the money conversation nobody enjoys, handled.
• Campaign Runner — sends targeted text drives to the right segment (seasonal PPF, ceramic top-ups, referral pushes) and runs an always-on win-back loop that texts cold leads to rebook.
• Message Composer— right in your inbox, turns a conversation into one tidy, ready-to-send reply in your shop's voice, editable before it goes.
• Call Intelligence — transcribes and summarizes the calls you take yourself, then drafts the estimates, texts and bookings the promises on that call created.
• AI Business Advisor — runs a full snapshot of your shop on a schedule and writes plain-language insights with what to do about them, so your numbers stop being a mystery.
• Voice Agent (beta)— picks up when you can't, chats with the caller in your shop's voice, and drops a ready-to-work lead with the full transcript. It's in opt-in beta, rolling out deliberately — the one part of the crew I'd call still-in-progress, and I'm going to be honest about that.

Ten of those are live today; the Voice Agent is the beta. Notice how few of them overlap. That's the point. Each agent is a specialist, not a jack-of-all-trades, because a specialist is what a specific moment in your funnel actually needs.

The first thing any sane shop owner thinks when they hear “eleven agents texting my customers” is: absolutely not, not without me watching.Good. That's the right instinct, and the whole thing is built around it.

Every agent ships switched off. When you turn one on, it starts draft-first by default. Draft-first means the agent does all the thinking — reads the lead, writes the message in your voice — and then stops. The message queues up on your Approvals tab. You read it, tweak the tone if it's a hair off, and hit send by hand. Nothing leaves your shop with your name on it until you've looked at it. For a week or two you basically get to audition each agent, one message at a time, and correct anything that doesn't sound like you.

Then, once an agent has earned it — once it's consistently writing the message you would have written anyway — you flip it from draft-first to autopilot and let it send on its own. That's the trust model in one sentence: it writes, you approve, until you don't need to anymore. You graduate each agent individually, on your own timeline. Maybe the Review Requester earns autopilot in three days because the ask is simple and low-stakes, while you keep the Invoice Chaser on draft-first for a month because money is touchier. That's your call, agent by agent.

This is the part that separates a crew you actually trust from a black box you cross your fingers over. You're never handing the keys to something you haven't watched drive. Every agent has its own instructions, its own daily cap, and its own send window, and you can flip any of it back to manual the instant something feels off. The autonomy is real, but it's autonomy you granted on purpose, not a default you have to go hunting through settings to undo.

Here's the distinction I care about most, and it's a boring technical one that turns out to matter enormously: the guardrails are enforced in code, not written in a policy doc I hope the model remembers.

Two things specifically. First, quiet hours. No agent ever texts a customer at 2am, full stop — every send checks your shop's send window and timezone before the message is allowed to leave. It doesn't matter if you told the Campaign Runner to blast a win-back drive and it's the middle of the night; the message waits until morning because the code won't let it fire outside your window. Second, opt-outs. The instant a customer replies STOP, they're out, honored immediately, and the required opt-out line rides on every message the crew sends. Add per-shop daily caps and per-lead cooldowns so nobody ever gets over-messaged, and the crew simply can't become the shop that annoys people.

Why does “in code” matter versus “the AI is instructed to be polite”? Because a model instruction is a suggestion. On a bad day, with a weird input, a model can wander off its instructions — that's just the nature of the thing. A hard check in the code can't. It's a wall, not a note taped to the wall. For your brand and for staying on the right side of texting rules, that's the difference between “usually fine” and “can't misbehave.” An agent that occasionally texts an opted-out customer at 2am isn't a helpful crew member — it's a liability, and it's built so it can't be one.

Every shop is also walled off from every other shop. An agent reads only your data — your leads, your quotes, your conversations — and can never see or act on anyone else's, and your customer information is never used to train an outside model. The crew works your funnel and nobody else's.

The reason eleven specialists beat one generalist is easiest to see when you lay them out across a single job's lifecycle. A lead doesn't just “come in” — it moves. New inquiry, sent quote, booked appointment, finished car, invoice, and eventually, if you don't stay on it, a cold name that drifts off. Each of those moments needs a different touch, and each has an agent whose entire job is that one moment.

Walk it with me. The timeline below is an illustrative pass through one job, showing which crew member owns each handoff — from the first text back to a lead all the way to the win-back loop that revives the ones who went quiet. Nobody drops the baton, because at every stage somebody's already holding it.

Let's talk plans, because I hate when software is coy about this. The crew splits along one clean line: texting agents on one tier, the calling suite on the next.

The texting agents — Lead Responder, Quote Follow-up, Review Requester, Install Reminder, Missed-Call Responder, the Invoice Chaser, and the Campaign Runner with its win-back loop — run on the Growth plan, CAD $199 a month. That's the tier built around the agents that visibly make money, and it pairs the crew with the Automations engine that carries the follow-up sequences forward on rails. The full calling suite — Call Intelligence and the beta Voice Agent — lives on Command, CAD $399 a month, the tier for shops where a single answered call a month covers the difference.

Every tier includes a monthly allowance of AI actions so the crew runs without a surprise bill — 100 a month on Essentials, 500 on Growth, 1,500 on Command. And here's the part that matters if you've got more than just yourself in the shop: pricing is flat per shop, not per seat. Growth and up come with unlimited team logins. Adding your front-desk person, your other installer, whoever — doesn't add a dime. The crew works for the whole shop at one price, which is how it should be. You're buying a shop's worth of help, not renting it by the head.

Let me bring this back to why any of it exists, because it's not about looking futuristic. I don't care if my shop looks like the deck of a spaceship. I care about getting out of the bay.

We're all installers before we're business owners. You got into this because you love the work — the film going down clean, the reveal when the tape comes off, the customer's face when they see it. You did not get into this to spend your Sunday night typing follow-up texts to quotes that went quiet, or to lie awake wondering which happy customer you forgot to ask for a review. That busywork is the tax on running a shop, and it's the exact stuff that keeps you chained to your phone instead of standing at the bench doing what you're actually good at.

That's the whole case for a crew instead of a chatbot. A chatbot answers a question and waits for the next one. A crew works the funnel while you work the car. It's not robots for the sake of robots — it's buying back the hours you burn every week on follow-up nobody enjoys, so you can close the laptop and pick up the squeegee. Turn on one agent, watch it catch a job that would've slipped, and add the next when you're ready. That's the point of the whole thing: less time being your own overworked receptionist, more time being the installer you set out to be.