Three issues in, and here we are again.

Reactivation isn't marketing. It's a relationship.

When we hear the word "marketing," most of us think acquisition — finding new customers, building new relationships, growing your base. That's what the word carries with it culturally and professionally. Marketing implies you're going out to get someone new.

But every product in this category is working a list of people who already bought a car from you. Already serviced with you. Already chose your store. There are no new customers here. There's a database of existing relationships — and an AI reaching back out to them.

That's not marketing in the acquisition sense. That's reactivation. And the distinction matters because it changes how you should be evaluating these tools entirely. The question isn't "how many people did it reach?" It's "how many of the right people came back?"

Why aren't they coming back in the first place?

Here's a question nobody in a vendor demo is going to ask you — but somebody should.

If a customer isn't returning to your service lane, an AI isn't the first question. The first question is why.

Was the last experience slow? Was the price a surprise at pickup? Did nobody follow up after the visit? The best version of any reactivation tool is the one you rarely have to use — because your service experience is so strong and your customer relationships so solid that people come back without being asked.

We're not naive. The real world is messier than that. Customers get busy. Life happens. A well-timed, well-targeted message from a store they trust genuinely does bring people back. Reactivation works. But it works best as a system that catches the customers who slipped through the cracks — not as a substitute for the experience that keeps them from slipping in the first place.

What good reactivation actually looks like.

Strip away the theater and here's what it looks like when it's done right.

• It's triggered by a real signal. Not a calendar. Not a blast schedule. A declined service. An overdue oil change. A lost customer who hasn't been back in 14 months. A recall notice. Something that gives the outreach a reason to exist beyond "we haven't talked to you lately."
• It works existing customers with a documented relationship. Not a purchased list. Not an OEM conquest manifest. People who chose your store — and who, with the right message at the right moment, will choose it again.
• It uses the right channel for the right customer. SMS and email for most people — because that's what the data says most people prefer and what converts best. Voice for the customer who hasn't responded to anything else. Channel should follow the customer, not the vendor's pitch deck.
• It measures one thing: booked appointments attributed back to the repair order. Not "conversations initiated." Not "engagement rate." Not "AI interactions." Revenue closed. That's the metric. If a vendor can't show you that number clearly and quickly, ask why.

You don't need a brain box. You need booked appointments.

Some vendors are selling dealers on the idea that the more their AI can do — the wider the conversation it can hold, the more topics it can handle, the more human it seems — the better the product. Dealers are being asked to invest time building out an AI's knowledge base, teaching it everything about their store so it can handle any conversation a customer might throw at it.

Think about what you're actually asking a reactivation tool to do: reach the right customer, at the right moment, with the right message, and book the appointment. That's it. An AI that can hold a freewheeling conversation about anything isn't a better reactivation tool — it's a different product entirely, with different risks.

A service advisor working your drive doesn't go off-script. They don't get pulled into conversations that have nothing to do with the service visit. They stay focused, handle the interaction professionally, and move to the next customer. Your AI should do the same.

We've heard from dealers whose customers came in genuinely upset because an AI kept a conversation going long past the point it should have handed off to a human — or went so far outside the original purpose that the outcomes were genuinely absurd. The capability to have an open-ended conversation sounds impressive in a demo. In practice it creates unpredictable outcomes that you, not the vendor, have to clean up. Focus and guardrails aren't limitations. They're exactly what you want.