Why ChatGPT doesn't mention your business: 8 causes

The eight causes ClientRadar's fix engine actually finds, in the order a scan checks them.

1. No one who already ranks mentions you

This is the most common cause and the least mysterious: the domains the engines cited when a competitor was named — a directory, a review site, local press — don't cite the target business at all. ClientRadar's fix engine finds these domains directly from the run's own answers, ranks them by how often they show up, and turns each into a task named after the actual site: “Get listed on {domain}”, with the answer that cited it as evidence.

2. robots.txt is blocking the crawler that would cite you

A site check fetches robots.txt first and checks it against five crawlers by name: OAI-SearchBot, ChatGPT-User, PerplexityBot, Google-Extended and ClaudeBot. If any of them is disallowed from “/”, that engine's citations of the site are cut off entirely, however good the content is. The full guide is “Is your robots.txt blocking AI search engines?”.

3. No LocalBusiness or Organization schema, or it disagrees with reality

The site check looks for a LocalBusiness (or a subtype of it) or Organization block in the page's JSON-LD and compares its name, phone and address against the business's own profile. Missing schema means the engine has to infer these facts from prose, which it does less reliably; schema that disagrees with the truth — an old phone number, a moved address — can get repeated back as fact. “Schema markup for local businesses in AI search” has the exact fields checked.

4. No page actually titled or headed for the service asked about

A prompt like “roof replacement in Dallas” is matched against a target service and city; the check looks for a page whose title or h1 carries those words, checking up to eight internal pages chosen for the prompts asked, including a “/services” hub and the sitemap when the homepage doesn't link the page directly. A site that only says “Services” with no specifics gives the engine nothing concrete to cite for that particular question.

5. No FAQ content for the engine to lift

Engines answer question-shaped prompts almost verbatim from FAQ content when they find it, whether it's marked up as FAQPage schema or just a visible Q&A heading. A site check with no FAQ schema and no FAQ heading anywhere it looked leaves this open as a task, since it's one of the cheaper fixes available.

6 and 7. No HTTPS, or the page is empty until JavaScript runs

A homepage that can't be loaded over HTTPS is both a trust signal engines weigh and a broken-looking result to a customer who clicks through. Separately, if the raw HTML carries under 200 characters of visible text and either shows a “turn on JavaScript” notice or an empty app-mount element with scripts attached, the check flags the page as likely rendering client-side — invisible to a crawler that reads HTML before JavaScript executes, which is how these bots mostly work.

8. No llms.txt

llms.txt is a plain-text page at the site's root that states in a few lines what the business is and does, meant to be read by an AI crawler before anything else on the site. Its absence isn't usually decisive on its own, but it's the cheapest single fix on this list — this site's own is at /llms.txt.

What this list leaves out

Two things are deliberately not on it. First, paid placement: there is none to buy, on any of these engines, for this kind of answer. Second, review counts and star ratings in the abstract — they matter to the extent they show up in a page the engine actually reads, but there is no separate “AI reputation score” being computed here. A free scan checks a real site against this exact list and reports which of the eight it found true.

See where your own business stands

A free scan asks ChatGPT and Google AI Mode the questions your customers ask and checks the site against the list above.

Run a free scan