Getting named
Why doesn't ChatGPT mention your business?
Published · Updated Lees dit in het Nederlands
Because the pages it reads don't mention you. ChatGPT builds a recommendation from reviews, registers, directories, trade press and your own site, and one of four things is wrong with that material: nobody independent names you, your site never states what you do and where, your details disagree from one listing to the next, or the crawler cannot read your pages. Each has a fix, and you can check all four in an afternoon.
How we know
We asked. On 15 September 2026 we put the question "why doesn't ChatGPT mention my company" to ChatGPT and to Perplexity, logged out, in a fresh chat, with no history. Both gave the same list, in different words. ChatGPT:
Your company may have weak web visibility. If authoritative websites rarely mention it, there's less reliable information for models/search systems to draw on. [...] Independent reviews, industry publications, directories, partnerships, and citations can matter more than simply having your own website.
Perplexity added one line worth keeping: "One conversation does not add your company to ChatGPT's general knowledge." Telling the assistant about yourself changes nothing. The four reasons below are their list, with what Van Found sees when it runs the same questions for clinics and law practices in the Netherlands.
1. Nobody independent mentions you
The assistant trusts what other people say about you more than what you say about yourself. A review on Google, a listing in a professional register, an entry in an association directory, a mention in a trade article: each is a second source confirming you exist and do what you claim. A business with a good website and none of those is, to the model, unverified.
This is the reason that matters most. Across 75,000 brands, mentions of the brand name on other sites predicted visibility in Google's AI Overviews better than backlinks or domain rating did, per Ahrefs research. If you fix one thing, fix this.
What we look at first: how many separate sites mention your name next to the word "review". Under five is thin.
2. Your site does not say what you do, for whom, where
"We deliver exceptional care" tells the model nothing it can match to a question. "Hair transplant clinic in Amsterdam Noord, FUE and DHI, consultations in English and Dutch" does. An assistant answering "which Amsterdam clinic does FUE with minimal scarring" needs those words on a page, in a sentence, in the first paragraph.
Most professional sites hide the specifics in a brochure PDF or a photo caption and lead with a slogan. The slogan is what the crawler reads first.
What we look at: the first sentence on your homepage that names your service, your city and your customer. On most sites it is below the fold, or missing.
3. Your details disagree across the web
Your Google profile says "Kliniek Jansen", your KvK entry says "Jansen Medisch B.V.", your Zorgkaart page has the old address, and your site has a phone number nobody answers. To a person, these are the same business. To a model matching entities across sources, they are three or four weak ones, and it picks none.
Consistency is dull work and it is the cheapest fix on this list. One name, one address, one description, copied word for word into every place you appear.
What we look at: every listing that shows your business, side by side with your homepage. The mismatches are the work.
4. The crawler cannot read your site
Assistants read the web through crawlers with names. OpenAI's is OAI-SearchBot. If your robots.txt blocks it, or your hosting firewall blocks unknown bots, or the facts that matter live inside JavaScript that never renders for a crawler, the model sees an empty page. Both assistants named this reason unprompted, and both named the robots.txt file.
You cannot see this one from the owner's chair, because the site looks fine in your browser and blank to the machine.
What we look at: your robots.txt, and whether OAI-SearchBot, ClaudeBot, PerplexityBot and Google-Extended can fetch the pages that matter. Your web developer can answer that in ten minutes if you ask in those words.
A fifth reason, and it is not about you
The question. Ask "best clinic in Amsterdam" and the assistant returns the same four names it gives everyone. Ask "I'm in Amsterdam Noord with a torn ACL, who is fastest for recovery" and the list changes, narrows, and becomes winnable. The generic question is a beauty contest. The situational one is a match, and matches are where a smaller practice gets named. We cover that in its own piece.
Find out which reason is yours
Ask ChatGPT the question your customer would ask, in a fresh chat with memory off, and read the names. Then do the same on Claude, Gemini and Perplexity, because they disagree with each other.
You can do all of the above yourself, and some owners do. Most decide their afternoon is worth more in the practice than in a robots.txt file, and hand it to someone who does this every week. Either way, start by finding out whether you are named. Send us your website and we run the questions for you, free, across all four assistants. You get the names back within a few working days, and the reason, if yours is missing.
Van Found is an Amsterdam consultancy that gets clinics, law firms and other professional practices named when ChatGPT, Claude, Gemini and Perplexity answer. One client per category, at a time.