AI is the new dashboard. Eighty-two percent of dealers use it. Nine percent say it made them money.
The arithmetic carries it.
Six weeks ago I wrote about the mirror.
The dashboard. The screen full of clicks, sessions and conversions that shows us motion and lets us call it progress.
I said the mirror wasn't lying. It just couldn't tell us whether the reflection was the outcome our business exists to create.
A lot of you wrote back. Some agreed. Some pushed. All of it made the argument better.
Since then, the industry has bought a new mirror.
It's called AI.
And it is the most convincing mirror ever built.
Let me be clear before I go further. This is not an anti-AI piece. We use it. We build with it. I believe it will change how vehicles are marketed and sold more than anything since the internet.
That is exactly why it deserves a hard look.
Here is the arithmetic.
Cox Automotive tracks AI use in dealerships every quarter. In its most recent tracker, 82% of dealers said they use AI.
Nine percent said it had improved profit.
Read that again. Eighty-two. Nine.
Dealers expected revenue growth from AI at a rate of 69%. Twenty-two percent say they've seen it.
And roughly one in three dealers using AI don't measure its impact at all.
That is not a technology problem.
That is the mirror.
Walk any trade show floor this year and you'll see it. At NADA, the number of AI-focused exhibitors went from about 10 to more than 40 in a single year.
AI that answers the website chat.
AI that texts the lead at 2 a.m.
AI that books the service appointment.
AI that writes the ad, prices the car and summarizes the CRM.
And every one of them arrives with a number.
Thirty percent more leads.
Twenty-five percent more appointments.
Forty percent more showroom visits.
I don't doubt that the activity is real.
I doubt that anyone has checked it against the sold file.
Because here is what almost every one of those numbers has in common: the vendor measured itself.
Appointments set by the AI, counted by the AI.
Leads generated by the platform, reported by the platform.
Conversions defined by the same system that gets paid for them.
We have seen this movie.
25 years ago the platforms taught us to count clicks, and we let their definition of success become ours.
Now the definition is getting smarter. It talks. It writes. It sounds like a person.
It is still measuring what it can see.
And what it can see is still not a sale.
There's a second shift happening, and it matters just as much.
The shopper is changing how they look.
Sixty-three percent of in-market shoppers say they plan to use AI on their next vehicle purchase. When Google shows an AI summary, people click through to websites about half as often.
So the research is moving somewhere the dashboard can't follow.
A shopper asks an AI which store to buy from. The AI answers. The shopper searches your name, clicks your ad, and books.
Your report credits paid search.
The conversation that actually chose you never shows up.
The mirror just got a blind spot, right at the moment it got more confident.
The big players see where this is going, and to their credit, they're moving.
This week Cox Automotive announced it is embedding Fullpath's customer data platform directly inside the VinSolutions CRM. That's a smart move. It puts the customer record at the center of the AI, which is exactly where it belongs.
But connecting the data is not the same as proving the result.
A unified record tells you everything a shopper did.
It doesn't tell you whether your marketing caused them to buy.
Only one scoreboard can answer that. It's the same one it has always been.
The sold file.
So here is my suggestion. Not a product. A discipline.
Before you renew any AI tool, ask five questions.
Who measured the result? If the vendor graded its own homework, you don't have a result yet. You have a claim.
Measured against what? Appointments and leads are activity. Match them to the vehicles you actually delivered.
Compared to what? What happened in the stores, months or households that didn't get it? Without a comparison, lift is a guess.
Which people? Can the tool tell you which real households it reached, or only which sessions and devices? A bot can book an appointment. It can't sign a buyer's order.
What did it cost per vehicle sold? Not per lead. Not per conversation. Per vehicle.
If your provider can answer all five, keep them. They're worth it.
If they can't, you're looking at a mirror. A very impressive one.
None of this means slowing down on AI.
It means putting the person back in front of the platform, one more time.
AI should help us find the right people sooner. Answer them faster. Follow up when a human would have forgotten. Those are real gifts, and dealers who use them well will win.
But the AI is a tool.
The person is the point.
And the sale is the scoreboard.
We built
MuddVision to keep score that way. Start with a known audience. Reach it across every channel. Match the sales back to the people we set out to reach, whatever tools touched them along the way, ours included.
You don't need us to do it. You need the discipline.
The last twenty-five years taught us what happens when we let the dashboard define success.
We don't have twenty-five years to learn it again.
Reach the person. Earn the sale. Tell the truth.
That's still the work.
Rob Mudd is co-owner of Mudd Advertising in Cedar Falls, Iowa, an agency his father founded in the family basement in 1981 and built around a simple principle: advertising should ultimately be accountable to the sale.
