The next decade of franchise retail won’t be won by the dealer who buys the most AI. It will be won by the dealer who owns the record.
Ask a simple question at your next manager meeting: what happened to the customer who called about a used SUV three Saturdays ago?
You will get four answers. The internet manager has the lead source and the first two texts. The desk has a worksheet. Service has a repair order on that customer’s trade from last spring, and nobody in sales knew it existed. Marketing has them in a campaign built off a spreadsheet list that was stale before it deployed.
That is not a people problem. What the vendor market gave you was a collection of systems that don’t agree with each other. Each was bought to solve a real problem, and each arrived with its own login, its own dashboard, and its own version of your customer. Integration was a treaty between vendors, not a design decision made on your behalf.
Then AI showed up. For the most part, it showed up as one more layer sold on top of the pile.
Why bolt-on AI runs out of road
The demos are impressive. A tool answers a lead in seconds, writes a friendly text, sounds like a person. Nine months later it still doesn’t know that the customer it’s texting bought from you in 2023, has a lease maturing in four months, and declined a brake job in June.
That is not a bug, it’s the ceiling of the architecture. Intelligence layered onto fragmentation only scales the inefficiency. A bolt-on tool cannot get smarter, because it has nothing to learn from. It sees the slice of your business it was given access to and nothing else. It has no memory of your store.
Here is the uncomfortable part. Every dealership already has the data. Almost none of them have the record. The record is the single continuous story of one customer: the click, the up, the appointment, the deal, the delivery, the first oil change, the declined recommendation, the equity position two years later. When that story lives in one place, software can learn something worth knowing. When it lives in five places, all you can buy is automation that guesses faster.
Coaching is where it gets real
Sales coaching in most stores is a memory exercise. A manager remembers who blew a deal Saturday and delivers the lesson Tuesday to whoever is standing there. The reps who need the most help are usually the quietest ones in the room.
Now put coaching on top of a complete record. The system already knows every rep’s response times, appointment sets, shows, and closes, because it holds the whole lifecycle instead of a fragment. That turns the daily checkout report into something a GM can manage from: team performance in one view, reps grouped into bands so the middle of your roster stops being invisible, coaching notes tied to real behavior rather than recollection.
That is why AI Sales Coach was built inside CRM by Solera rather than beside it, with permissions, configuration toggles, engagement rules, and store hours all set by the dealer. AI in a store should not be an oracle. It should be a second set of eyes that never stops taking notes.
When a store commits, the pattern shows. Pat Lobb Toyota grew monthly sales 50% and lifted its close rate to 55%, an increase of 13 percentage points, and was live in under 30 days. That was achieved by aligning CRM structure, manager coaching, campaign-driven execution, and channel optimization on one platform. Read the caveat closely, because it is the lesson. Technology did not produce the number. Alignment did. Software made alignment something you can run daily instead of quarterly.
What this has to do with direct-to-consumer
Talk to dealer principals about factory direct sales and the conversation goes straight to franchise law. Fair enough, but that scoreboard is mixed: direct sales are permitted in roughly half of U.S. states, and one EV maker won another state-level fight this spring. Overseas, one major automaker just discontinued direct agency sales to private customers across Europe and handed distribution back to its dealers.
The legal front is not the real threat. Direct-to-consumer isn’t fundamentally a licensing model, it’s a data model. It works when one party holds the whole customer relationship end to end and can anticipate the next need. Anyone who assembles that relationship can compete for it: a factory, a national retailer, or a software company that started out as your lead provider.
Your advantage in that fight is enormous and frequently wasted. You have the trade, the service history, the warranty work, the equity position, the local relationship, and the technicians. Nobody upstream has that. But an advantage scattered across six systems isn’t an advantage, it’s an archive. In one record, it’s the most defensible position in the business, because the store that knows what a customer needs next doesn’t have to wait for them to start shopping.
The store is the operating system
The CRM question is the small one. The real question is what your store runs on.
Sales, service, finance, marketing, websites, inventory, titling, accounting, and the connected vehicle after delivery are not eight software categories. They are eight views of the same business. When they write to one customer record and feed one engine, every piece you add makes the pieces you already had smarter. Service history sharpens equity mining. Equity mining sharpens marketing spend. Marketing sharpens what you stock. Connected vehicle data carries the relationship past the curb, where most retention gets lost. Titling stops being a department outside the deal and becomes a property of the platform.
That is the difference between a software stack and an operating system. A stack adds tools and cost. An operating system compounds. The platform your store runs next quarter should be worth more than the one it runs today, with no re-implementation and no new vendor.
Before your next vendor meeting, run the audit on yourself
Software gets blamed for a lot of things that are really decisions nobody made. So start inside the building.
Ask yourself: if I fired every vendor tomorrow, what would my store actually still know about its customers? When was the last time a system told one of my people something they didn’t already know? Will this tool be measurably smarter in twelve months because my store used it, or is it the same product with a new interface? Can I cancel something because of this? And when I add the next piece, does this one get better?
A stack is a set of tools you own. An operating system is something your store is teaching every day, in every lane, whether anyone is watching or not. One depreciates. The other compounds.
The dealers who win the next decade won’t be the ones with the most AI in the building. They’ll be the ones whose store learns something new every day and puts it to work before the customer has to ask.