Three dealerships. Twelve weeks. Published either way.
A lot of dealers will tell you that you cannot change what ChatGPT says about a store. So before choosing anyone, I measured the problem. Now I am choosing three stores to prove the fix, in public.
The problem, measured first.
On September 24 and 25, 2026 we asked ChatGPT, signed out, the questions buyers type about 48 franchised stores in California, New York and Virginia. 502 answers, each coded for which stores it named, on the map card and in the text, and which outside sources it named. Every question went out again the next day.
When the buyer had not picked a store yet, ChatGPT named the store half the time: 84 of 168 answers. 18 of the 48 never came up in a single open answer. A day later, whether the store got named flipped 1 time in 6. In the 168 open answers, ChatGPT named CarEdge in 146 and DealerRater in 101.
The study.
30 buyer questions, sales and service, on ChatGPT and Google AI Mode, signed out and dated. Read against a control store we are not working.
We fix the outside records those answers lean on. The store approves every fix, and each one is logged the day it goes live.
The same 30 questions again, on the same engines, with the control read alongside.
Published with the store's name, agreed in writing before week 1. The control stays unnamed.
Whatever it shows, we publish it. Read the method behind every reading.
GMs and owners who want their store considered.
Tell me your store and market. I will come back within one business day with times for a 15-minute call. Terms, and whether your store is named, are agreed in writing before week 1.
Put my store forward