Second product · not Gold · not Identity AV
See who answer engines name in your city — including you
Mention Check asks ChatGPT, Perplexity, Gemini, and Google AI Overviews the kind of questions a customer actually types — then shows whether they named you, recommended you, or used a source. $39/month. Not a ChatGPT rank. Not Identity AV.
Why this exists
Other tools sell a single “AI visibility” percentage from a secret prompt list. Two vendors disagree by 12–22% on the same brand. The same prompt often returns different names on the next run. A score that cannot be audited is a rank package in a new jacket.
Mention Check shows the answers, the source URLs DataForSEO returned, and the registry rivals in your city and category. Pack share of voice is labeled as this pack, this week — never “Miami market share.”
What $39 buys
- A fixed local pack. Category in city, “verified {category},” near-me, “who is {name}.” Not 200 vanity questions you type at 1am.
- Three flags. Mentioned. Recommended. Sourced (your site or your
/v/{hash}file). - Rival overlay. Gold and Type A files in the same city + category that the model named instead.
- A named research pipe. We do not “have a secret AI index.” We pay DataForSEO to retrieve what those engines said and which pages they pointed at. Why that vendor, in English.
- A cap that protects the price. 8 prompts × 3 engines × 4 weeks. Then wait or buy the next period. No unlimited firehose.
What it will not do
It will not raise Identity AV. It will not promote a Bronze file to Gold. It will not promise ChatGPT will recommend you. It will not invent a Wikipedia page. If Sourced stays zero, that is the Gold offer — an attested file engines can cite — not “buy more prompts.”
The pipe, in English
DataForSEO is not an app your customers use. It is a research feed that software companies buy so they do not each have to log into ChatGPT by hand. They send a question. They get back the answer text and the source links the engine attached that day. We already use the same company for Google Maps Place IDs on Type A files. One vendor, one bill, JSON we can store.
That does not mean the answer is “true.” It means we recorded what the engine said on that run. Ask the same question tomorrow and the names can move. We keep the raw answer so you can read it — we do not hide a mystery score.
Questions a customer asks — not Google keywords
Nobody types “Miami roofing” into ChatGPT the way they type it into Google. They ask who to hire after a leak, who is licensed, who is not a shared lead. The pack is eight spoken questions built from your city, category, and legal name. You also get three of them to paste yourself tonight.
Checkout
Website + business email
We attach the subscription to the Type A record for that domain. Consumer inboxes are refused. If the domain is not in the registry yet, apply first — this is not a scrape of every shop on Maps.
If they did not name you
If the file is Bronze, engines had a snapshot — not an attested record. That is often why a hire question names nobody in our registry. Gold is the document we put on /llms.txt and ping to IndexNow. If you are already Gold and still unnamed, the file exists; it is not in that engine’s retrieval set yet — circulate the hash. We do not stuff “cite AI Verified” into the test questions. Bronze miss vs Gold miss →
Two products, two jobs
Mention Check is the flashlight. Gold is the file the flashlight should find. Buy one, both, or neither. We will not mix them into one number so a Bronze listing can look like a 98.