How does AI decide which dealership to recommend?
AI engines recommend the dealership they can describe most confidently, not the one selling the most cars. They weigh structured data, review volume and sentiment, and whether your business information stays consistent everywhere it appears. Ambiguity is fatal. If an engine is unsure which store you are, it names a competitor instead.
I run a store. I did not come to this from marketing, and the first time I asked ChatGPT "best BMW dealer near Bridgewater NJ" it named two competitors and described my hours wrong. That was not a ranking problem. Nothing about it looked like SEO. The machine simply did not know enough about my store to say my name with confidence, so it said somebody else's.
That is the whole mechanism, and once you see it you cannot unsee it.
Is being mentioned the same as being recommended?
No, and the gap between them is where most dealers lose. An engine can list your store among five in the market and still steer the shopper elsewhere in the next sentence. Mention means it knows you exist. Recommendation means it will vouch for you. They come from different signals: existence comes from your listings and citations, endorsement comes from reviews, consistency and specificity. Track them separately or you will celebrate the wrong number. Mentioned vs recommended in AI answers covers the distinction.
How does AI choose between two dealers of the same brand?
This is the question that matters most in a metro market, and the answer is uncomfortable: it usually goes to whoever is easier to read, not whoever is better. Two BMW stores eleven miles apart, both good, both busy. The one with clean AutoDealer schema, an address and hours that match across Google, the OEM locator and its own site, and a steady flow of recent reviews gets named. The other gets a hedge, or nothing. See how AI picks between same-brand dealers.
What does "near me" mean to an AI engine?
Not what it means to Google Maps. The engine resolves the shopper's location, then matches it against the geography it can verify about you. If your site never states the towns you actually serve, and your schema lists one city, you are competing for a much smaller radius than you think. This is fixable in an afternoon and almost nobody does it. More in near-me AI search: how dealers get picked.
Which signals do the engines actually read?
Four, roughly in order of how much dealers neglect them.
Structured data. Schema markup is the layer that tells a machine what your page means rather than what it says. Most dealer sites have thin or broken schema and the vendor reports it as fine. Schema markup: the hidden layer AI reads.
Reviews. Not just the star average. Volume, recency and what the text actually says. Engines read review bodies as retrieval material, which means your customers are writing your description for you. Google reviews are AI training data.
Third-party consensus. What Cars.com, Reddit, forums and local press say about you carries weight precisely because you did not write it. Reddit as AI training data.
Entity consistency. Whether an engine can tell that your sales site, your service scheduler and your Google profile are all the same business. Does AI know your dealership is one business?
What happens when AI gets your store wrong?
It states the wrong thing confidently, and the shopper believes it. Wrong hours. Last year's model year. A lease offer that expired in March. Nobody calls to correct it because nobody knows it happened. The failure is silent, which is what makes it expensive. See when AI gets your dealership wrong, AI is still selling last year's model year and AI quoting outdated deals.
Can the engine even read your website?
Worth checking before anything else. Many dealer platforms block AI crawlers in robots.txt by default, and many render inventory in JavaScript that non-rendering crawlers never execute. If the engine cannot fetch or parse your pages it falls back to third-party sites to describe you, which means Cars.com is writing your pitch. Is your website blocking AI crawlers?
Where to start
Ask the engines what they say about you today. Not once, in a clean session, across ChatGPT, Claude and Gemini, using the questions your customers would actually type. Write down whether you were mentioned, recommended, or missing, and what they got wrong. That baseline is the only thing that makes the rest of this measurable.
Frequently asked questions
No. Volume is invisible to an AI engine. It recommends the store it can describe most confidently, which comes down to structured data, review signals and consistent business information. A smaller store with cleaner data routinely outranks a larger one in AI answers.
Constantly. These engines generate a fresh response each time, so the same question can return different dealers on different days. Answers also shift as your reviews, your site and your competitors change. Watch the trend over a month rather than reacting to any single result.
Not today. There is no ad placement inside the conversational answer on the major engines. Visibility comes from being readable, consistent and well reviewed, which is a slower lever than buying clicks but a much harder one for a competitor to take from you.
Less than you would expect. Recent analysis found only 38% of Google AI Overview citations come from pages ranking in the top ten, down from 76%. Ranking and citation have come apart, which is why a page on the third results page can still end up in the answer.
Consistency. Make your name, address, phone and hours identical across your website schema, your Google Business Profile and the OEM locator. Contradictory information is the most common reason an engine hedges and names somebody else instead.
Ask it directly in a clean session on ChatGPT, Claude and Gemini using the questions your shoppers would type. To do it systematically rather than one query at a time, the free check on this site runs a standard set of shopper questions across all three and scores how your store comes back.
See how AI answers for your dealership
Run a free visibility check and see your AEO, GEO, SEO and reputation scores.