How Car Dealerships Show Up in ChatGPT Recommendations, and Who Fixes It

What ChatGPT said about 48 franchised dealers

Each question was asked twice, a day apart. That produced 502 complete answers. When the buyer hadn't picked a store yet, ChatGPT named the store in 84 of 168 answers. 18 of the 48 stores came up in none of their open answers. On an open question, the store flipped in or out of the answer overnight 1 time in 6, and the first store on the map changed 6 times in 10.

The gap by brand type was wide. The 19 domestic-brand stores came up in 2 of every 3 open answers. The 11 luxury stores came up in about 1 of 4.

What the engine reads before it names a store

In those same answers, ChatGPT quoted CarEdge in 146 of the 168 open answers and DealerRater in 101. Those are outside records about the store, and that is where the answer gets decided. Our working view is that a store with a clean website and a strong Google rating can still lose the open question when the pages the engine cites disagree about it.

Things worth checking: a sister store's rating printed beside your name, a service listing with a lower score than the sales listing, an old or merged dealership name on a directory, a wrong phone on an aggregator, and a manufacturer locator entry that doesn't match the Google listing.

The order to fix things in

1. 1

   Ask the buyer questions signed out, twice, and include the avoid question.

   Open ChatGPT in a private window and ask "best [brand] dealer near [city]", "most trusted car dealership in [city]", "where should I buy a used car in [city]", and "which car dealerships in [city] should I avoid". Repeat a day later. An answer that names your store inside a warning is a loss that a mention counter scores as a win.

2. 2

   Write down the sources it cites.

   The citations are the lever. For franchise stores they are usually CarEdge, DealerRater, Google and Yelp, not the store's own site.

3. 3

   Make every outside record say the same thing.

   Store name, street address, phone and brand on Google, Yelp, DealerRater, CarEdge, Bing Places, Apple Business Connect, the manufacturer locator and the aggregators. Keep sales, service and parts as separate, correctly named listings.

4. 4

   Claim the third-party profiles under your own login.

   A profile you haven't claimed can't be corrected, and it keeps feeding the engine whatever it last scraped.

5. 5

   Answer the questions buyers ask before they pick a brand.

   A page on your site for each open question you want to appear in: used cars in your city, trade-in, financing, service, pricing and fees. State the doc fee and the process in plain words. Engines quote plain statements.

6. 6

   Run the same questions again and compare dates.

   The engines keep re-reading their sources, and competitors keep moving. The comparison means something when the questions, the engine and the signed-out condition stay the same.

Who fixes it

Storefront Audit does steps 1 to 4 and 6 for dealers. AI Source Repair is $797 one time: your market's buyer questions are run clean-state, every source feeding each answer about your store is traced, the third-party profile feeding a wrong number is claimed under your login, corrections go to the aggregators mislabeling your store, and at day 30 the identical capture runs again with a before-and-after page. If anything in scope isn't handled by day 30, the work continues free until it is.

The monthly is $2,500 per franchise store, or $1,500 for a single independent rooftop, after a one-time $2,500 setup. The $797 AI Source Repair counts toward the setup, so the setup is $1,703. No contract.

Adrian Marin spent eighteen years in car sales, from the floor to the general manager's office. He reads the answers himself.