Presented by:
Hey everyone,
Ever wish you had a group of dealer pals to exchange ideas and experiences with? CDG Circles offers that and more.
Learn more here.
— CDG
Last week, I dug into The Presidio Group's new two-part AI research released on Sept. 15 and 29.
What jumped out: 11.2% of dealers in Presidio's survey named tech-driven efficiency as a top business driver for the year ahead. At year-end 2024, it was 6.4%.
Not a huge number, I know. But it nearly doubled in 18 months.
So, because more dealerships are heading this way, here's some context on where dealers are starting with AI in their stores, how they're landing on AI's short list, and what they're cleaning up so AI gets their reputation right.
Dealers are winning small with AI, while waiting on connected systems.
In Presidio's survey, 68% of dealer respondents reported positive overall experiences with AI. (That should make any shop nervous about exploring the use cases feel a little better…)
But if you're unsure where to begin: Brodie Cobb, CEO of Presidio, said in the report that AI is first helping solve repetitive tasks.
The chart below from part 1 of Presidio's AI report blends polls about usage.
Courtesy The Presidio Group
What's working now, and where there's room to grow, per Presidio:
Vehicle appraisal, and reputation and review management, rank among dealers' most-used AI tools in the survey, alongside service scheduling and automated call answering.
But cutting the time salespeople spend entering data and doing other routine tasks could, for the first time in decades, move the 10-cars-a-month benchmark, according to Presidio.
In one example Presidio cited, Hudson Automotive Group uses AI to check its monthly DMS invoice (which is about 3,500 lines long) against its contract, catching charges that don't match negotiated terms.
The Charleston, S.C.-based group also uses AI to schedule service appointments. It hasn't seen a workforce reduction or lower expenses, but Presidio says it has seen a significant boost in productivity.
Looking ahead: Cobb suggests patience for dealers who haven't seen the results they expected from AI (only 22% of AI-using dealers have seen the sales growth they hoped for, per Cox Automotive data cited by Presidio).
"The most honest answer is the one thing they should do [is] wait. There's going to be a lot of glass broken before it all works out," Cobb told CDG News via email.
Brodie Cobb
Presidio
For dealers starting from scratch, he advises against leading with one-off software tools that have some AI built in, since they may not integrate with a group's core tech platform.
He also says to keep an eye on the largest auto retail tech companies building end-to-end systems. Cobb expects the industry to make substantial progress integrating AI throughout dealerships and groups in the next 12 to 18 months.
Upside of a connected store: “If a customer says, ‘Don’t ever call me at 3 o’clock. I’m picking my kids up from school,’ an integrated agentic AI system will remember and learn from that,” Cobb said. “And guess what? It will never happen again.”
A quick word from our partner
A shopper who’s just browsing can look a lot like one who’s ready to buy, until you know what to look for.
This guide breaks down four real signals of buying intent so your team can prioritize by readiness instead of when a lead came in.
Dealers are fighting for a spot in AI's answers.
When a shopper asks AI where to buy a car, the answer is usually a short list of named dealerships, not a long page of links.
In fact: CarGenius, an AI visibility firm, tests that with the same non-branded questions real shoppers ask, such as "best Toyota dealer near [city]," on ChatGPT, Google AI Mode, and Google AI Overview.
Results are weighted by search demand in each city.
In its first reports, covering Toyota dealers across the 25 largest U.S. metros in July, the company found:
In Miami, AI named actual dealerships 99.6% of the time, and only three to six per answer.
In 17 of 25 metros, the AI favorite wasn't the store with the most reviews.
In 23 of 25 metros, dealers farther from the city center got recommended less often. AI's top pick was usually about 8 miles from the center, while the typical dealer in those metros was about 18 miles out.
In 18 of 25 metros, the No. 2 dealer was close behind AI's top pick, and no dealer anywhere dominated the recommendations.
CarGenius' public regional reports also rank every dealer in a metro by AI visibility.
And in 12 of 25 metros, AI recommended a different dealer for service than for buying. Winning on sales doesn't carry over to service, CarGenius found.
What they’re saying: Frank Loeffler, CarGenius' head of AI services, told CDG News that the goal is a consistent message about your store, both on your website and off it.
“It’s a little bit different from SEO, where it’s very keyword driven," Loeffler said. “This is more about making sure in this new world that everybody’s saying the same thing about you, both on your site and off your site.”
Frank Loeffler
CarGenius
In other words, tidy your house: As we've reported, a strong, up-to-date website has always mattered. But it matters even more because now AI is checking it, too.
On Aug. 8, the share of ChatGPT's background searches that looked inside one specific website, such as a dealer's own site, jumped from 0.37% to 16.8%, according to Promptwatch data cited by CarGenius, as seen below.
Meaning, once ChatGPT has a short list of dealers, your website is one of the places it can go directly to research your store before recommending it, according to CarGenius.
How to make your site shine:
Be explicit about how you sell cars, according to CarGenius.
Generic lines like "committed to great customer service" won’t wow ChatGPT. Instead, spell out the cities you serve, the brands you specialize in, your service offerings, and how financing, trade-ins, and appointments work.
Owner interviews, staff introductions, dealership history, community involvement and YouTube videos help show what makes a store special, Loeffler said.
The testimonials page gets picked up fairly often and should ideally be updated at least weekly, Loeffler said. A couple of blog posts a month helps, too.
Put it in play: “The goal is to become a dealership AI consistently knows about, understands, trusts, and recommends,” CarGenius wrote.
Dealers are cleaning up what AI sees about them across the web.
In a 2026 Urban Science-Harris Poll study cited by Presidio, 40% of dealership executives said inaccurate or incomplete data has hurt performance.
The sources AI cites may surprise you.
Case in point: In CarGenius' July Toyota reports, Edmunds, Autotrader, and KBB didn't show up in the 10 most-cited sources in any of the 25 metros.
Not to mention, 16 of 25 metro leaders with a Yelp listing had less than 3 stars on Yelp, compared with a 4.5-star average on Google, even though Yelp feeds AI answers in most markets.
And yes, even details hidden in a website's code count.
CarGenius CEO and founder Corey Lydstone described one dealer whose website metadata still listed an old phone number. When asked for the dealer's number, ChatGPT gave one that didn't appear anywhere on the visible site.
"It's not just consistency in what you can see; it's consistency in what AI is seeing too," Lydstone told CDG News.
Corey Lydstone
CarGenius
How to tighten it up:
Keep your Google Business Profile and Yelp current, and respond to complaints. AI engines are picking up Yelp more than before, Loeffler said.
Get each rooftop's basics right. A group's reputation can help, but every local store still needs its own fundamentals in place, Loeffler said.
That means accurate name, address, phone, hours, departments, brands, service capabilities, and review profiles for each store. For groups, consistent processes across stores also help AI understand the group as a whole, according to CarGenius.
(P.S. Even "auto" vs. "automotive" matters, since AI sometimes gets confused between dealerships.)
Lastly: Don't build your strategy on one source. Around Aug. 14, CarGenius saw ChatGPT nearly stop pulling Reddit pages into its research for find, buy, and service questions.
Promptwatch similarly measured an 86.4% drop in Reddit's share of ChatGPT citations over the same stretch. Because of this, CarGenius advises dealers to keep watching Reddit for real customer conversations, but not to chase Reddit mentions.
To review: "The sources AI relies on can change quickly," CarGenius wrote. "Your AI visibility strategy needs to work even when the source mix changes."
The bottom line: Presidio expects AI to prove the most productive operating technology auto retail has seen in more than 100 years within the next five years.
"It's a much larger comprehensive thing, but it's coming," Cobb said.
In the meantime, AI assistants already have favorite dealers, and in most markets, today's leaderboard can still change.
