Someone two streets away just asked an assistant where to eat tonight. The answer held three names. We find out whether one of them is yours, and we send you the answers we recorded.
One free scan per email address, three questions, ChatGPT only, answer in your inbox.
The questions below are built from what you typed upstairs. Change the street, the category or the engines and run it again.
Every question is asked 3 times, so the same answer can come back different. Models change their wording between runs, and that is exactly what we are counting.
Illustrative figures. The real scan is free, takes two business days and arrives with the raw answers attached, so you can check every line yourself.
Request the real scanThe panel above is a demonstration. The free scan is real: we send five of these questions to ChatGPT, read what comes back and email you the result with the answers attached.
One free scan per address, five questions. The full audit behind the paid work runs twenty questions, three times each, on every engine you pick. We do not send anything else, and nothing runs until you click the link in the email.
The free scan is one per email address, and it is meant as a taste. The paid plans measure twenty questions, three times each, on ChatGPT, Claude, Perplexity and Gemini, every thirty days, and we do the repairs in between.
Already used your free scan? Reply to the report email and we will run the full audit as part of the paid work.
Your name as customers say it, your street, your category. That is the whole form. No card, no login, no website audit tool to learn.
Your twenty questions go to the engines you picked through their official APIs, three runs each. We read the text of every answer and check whether your name, or a spelling of it, appears.
The share is that count divided by the number of answers. Nothing is modelled, nothing is inferred from your website.
Report and log inside two business days, plus a call. We run a small number of free scans each week, because every question we send costs us money and we would rather run few and run them properly.
You do not get a login. You get a document you can read at the counter, the raw answers behind it, and a person on the phone.
Not a score our software guessed from your website. A question sent to a real engine, and the answer it gave back, kept with a timestamp.
Same category, same street, one question. Keep scrolling and the scan runs door by door.
Other tools crawl your website and estimate what an assistant would probably say about it. We ask the assistant itself, three times, and keep every answer.
Twenty questions a real customer would type about your category and your corner. Every engine you picked, three runs each, official APIs only. Every answer names somebody, and we count who.
Structured data on your site, then Google, Apple, Bing and Yelp, plus the platforms your trade actually uses, corrected until every profile agrees on your name, hours and address. Pages that answer the exact questions. Replies on every review.
Two pages every thirty days: your share of answers, the movement since last month, what we changed, what is next. The raw answers sit behind the report and you can read them line by line.
Twenty questions, the engines you pick, one written report inside two business days. You see the score and the things holding it down. No card, no call, nothing to cancel.
No. Classic SEO moves you up a list of links. This measures whether an assistant says your name out loud when somebody asks it for a recommendation. There are no positions and no links in an assistant's answer, only names, so either yours is in there or somebody else's is.
You can, and you should. Two things get in the way. Your own assistant knows you, remembers your chats and your location, so it names your business to you more often than to a stranger. And the answer changes every time it is asked, so one scan tells you nothing. We ask from a clean session, repeatedly, and count.
Yes, and that is expected. Our report is a record of specific answers at a specific time, not a forecast of what you will see in a minute. Every answer arrives with a timestamp so you can read exactly what was said and when.
Structured data on your site so a model reads your name, address and hours without guessing. Your listings on Google, Apple, Bing and Yelp, plus the platforms your trade actually uses, corrected until all of them agree. Pages that answer the exact questions we measured, in the words your customers use. A reply on every review from the last year.
Listings and structured data land in the first two weeks. The answers usually start changing between week four and week eight, because a model has to crawl the corrected sources before it can repeat them. You get the same measurement every thirty days, so you watch it happen instead of taking our word.
Then we simply tell you so, with nothing to sell. It happens, mostly with businesses that have a complete Google profile and a website that answers questions in plain sentences. The scan costs you nothing either way.
Pick your category and your corner. We run the questions, record the answers and send the report. And if the assistants already name you, we will tell you plainly and leave it there, with nothing to sell.
Request the free scan