18 of 48 franchise stores never came up when buyers asked ChatGPT who to buy from. We fix that.
In September we asked ChatGPT, signed out, the questions buyers type about 48 franchised stores. 502 answers, all dated. When the buyer hadn't picked a store yet, 18 of the 48 came up in none of them, and a day later, on 1 open question in 6, the store flipped in or out. I find the outside records those answers lean on, correct them, and run the same questions again so you see the change with a date on both sides.
Dated readings. Check any of them.
stores never named in an open buyer answer
open questions where the store flipped in or out overnight
Temecula, October 7: ChatGPT named 3 to 6 businesses per answer. Google Maps lists 20 to 31 for the same searches.
checks in Google's top three map results, reviews up from 45 to 52
on Google maps, from outside the top three
wrong directory listings gone in a week
Three answers a buyer can get about your store.
The engine answered the question and your store wasn't in it.
You were in the list. The buyer was pointed somewhere else.
The engine picked you, and said why.
There is a fourth, and a mention counter scores it as a win: the avoid list. I read for it on every run.
What ChatGPT said about 48 franchised dealers.
ChatGPT answers with total confidence. Ask it the same question tomorrow and the answer can change. We measured how often.
ChatGPT, signed out, 24 and 25 September 2026. 48 franchised stores in California, New York and Virginia, every question asked twice, a day apart.
When the buyer hadn't picked a store yet, it was a coin flip. 18 of the 48 stores came up in none of their open answers.
Same question, one day later, the store moved in or out of the answer 1 time in 6. The first store on the map changed 6 times in 10.
It quoted DealerRater in 101. Those are outside records about the store, and that is where the answer gets decided.
That is the part we fix. We find the records ChatGPT reads about your store, correct them, and ask the same questions again, so the answer changes with a date on both sides.
The published dealer study is enrolling now: three stores, twelve weeks, every fix approved by the store, published with the store's name whatever it shows. Until it publishes, the closest automotive proof on record: a Huntsville performance shop named first by ChatGPT on 3 buyer questions, October 2, 2026, signed out and dated.
Run by someone who sat on your side of the desk.
Storefront Audit is a Temecula, California company that shows car dealerships and local businesses how they appear in AI answers like ChatGPT and Google's AI Overviews, then fixes what holds them back. Founder Adrian Marin brings 18 years of dealership management and 27 years of retail management. Its AI Visibility Index tracks how often a business is named, cited, and recommended across the major AI engines.
Adrian Marin spent eighteen years in car retail, from the floor to the general manager's office. He was general manager of Carlock Nissan of Jackson, Tennessee, from February 2015 to November 2017: a 52-person store selling more than 1,500 units a year. By his own store records, CSI rose 100 points and service and parts profit rose $40,000 a month.
He trained salespeople, F&I managers and sales managers. Now he measures what ChatGPT and Google say about stores like his, signed out and dated, and fixes the records behind the answer. The measurement is his, and so is the phone call.
What a market looks like when it is measured.
Hickory, North Carolina. Perplexity. The same 15 questions, four runs across 22 days. Seven stores unnamed in July, seven on 8 August, seven the morning of 14 August, nine that afternoon. Four were never named once, in any of the 60 answers recorded. Three of those four are franchise stores: a Hyundai store, a Mitsubishi store, a Volkswagen store.
Dallas, Texas. ChatGPT, signed out, 31 August 2026. Twenty buyer questions. Six of the twenty most-reviewed used-car dealers were never named. The six carry 453 to 2,228 Google reviews each.
In Hickory, one store went unnamed across three runs over 22 days. Four hours after the third run, we asked the same 15 questions again, and that store was named four times. We had changed nothing there in between. That's why one check is never enough. From the West Coast to the Northeast and Southeast, we've logged hundreds of dated buyer answers.
Yours runs the same way. Request your market's index.
The method, in four steps.
Every reading is dated and taken signed out, so you can ask the same question yourself and check our work.
Measure
Buyer questions, signed out, on ChatGPT, and on Google AI Mode for client readings. Every answer saved with its date.
Find the source
Which outside records each answer leans on: review sites, directories, data aggregators.
Fix and log
Correct the records. Every fix is written down the day it goes live.
Read it again
The same questions, the same conditions, on a schedule, with a dated reading on both sides of every fix. In the published study, a store we are not working is read alongside as a control.
The month.
The Index.
The same buyer-question battery runs again on your market every month, signed out, dated and verbatim. Same questions, same conditions, so month over month is a real comparison and not a fresh opinion.
The delta.
What moved and what didn't, against last month's capture. Every month carries its own intervention list off that delta, and the work that month comes off the list.
Vendor reads.
Two AI vendor pitches a month, read against a fixed written scorecard, back to you inside five business days.
The call and the one-pager.
Forty-five minutes a month to walk it, plus one page you can carry upstairs and defend without me in the room.
No passwords, no DMS access, no CRM access.
Nothing in this scope needs a login from your side.
There is no maintenance month.
Month one clears the wreckage. The months after are when the position gets built, because the engines keep re-reading their sources and your competitors keep moving. You can cancel any month, there is no contract. And the day the work stops, the engines keep re-reading and the gap starts reopening. The month the intervention list runs dry is the month you should cancel, and I will tell you when it does.
Pricing, in writing.
Groups are priced on the call, because the market count and the call change.
Every month the same buyer questions run the same way, signed out and dated, and you see what changed in writing.
Local businesses.
Not a dealer? Same discipline, built for owner-run businesses. The free audit shows where you sit on Google's map today and what to fix first. Dated results are on the results page.
Start the free auditYour store, measured the same way.
Or skip the form
Know you want the call? Grab a 15-minute slot on my calendar directly, no waiting on a reply.
Book a 15-minute call