Why Doesn't ChatGPT Mention My Business?
Most of the time it comes down to one of five things: ChatGPT is answering from memory rather than searching, your site cannot be read by the crawlers that feed it, nobody else on the web describes you in the words people are actually asking, your own pages describe you three different ways, or — the uncomfortable one — your competitors are simply better documented than you are.
They are not equally likely, and only some of them are yours to fix. Here is how to work out which one applies to you.
First: is it answering, or remembering?
ChatGPT works in two modes, and the difference decides everything that follows. Sometimes it searches the live web and builds an answer from pages it has just fetched. Sometimes it answers from what it learned during training — knowledge fixed months earlier, with no link and no source.
You can usually tell which happened. A searched answer shows sources or link previews. A remembered answer is plain prose with nothing to click.
This matters more than any other single factor. If it answered from memory, nothing you publish this week will change that response. If it searched, everything you publish is in play.
Reason 1: it never searched
This is the most common explanation and the most frustrating, because it is not a fault in anything you did. The model was asked a question it felt confident answering, so it did not look anything up — and that confidence came from a snapshot of the web taken before you existed, or before you mattered.
There is no direct fix. What there is instead is a slower one: every mention you accumulate now becomes part of what the next model learns. That is a genuinely long game, measured in model releases rather than weeks.
What you can act on immediately is the searched answers, which brings us to the four reasons you actually control.
What the research actually shows
Before going through the four fixable reasons, it is worth knowing what actually correlates with a brand showing up in AI answers — because the honest answer surprises most people who assume this works like classic SEO.
Ahrefs studied 75,000 brands against their appearance in ChatGPT, Google AI Mode and AI Overviews, using Brand Radar data from December 2025. The strongest single signal was YouTube mentions, correlating at roughly 0.737 — stronger than anything else tested. Branded web mentions (being talked about across blogs, articles, forums and directories) correlated at 0.66–0.71. Backlinks and Domain Rating, the two metrics most SEO strategy is built around, correlated far more weakly — Domain Rating at just 0.266 for ChatGPT specifically. Number of site pages barely mattered at all, at roughly 0.194. Publishing more of your own content is not the lever most people think it is; being talked about elsewhere is.
The same research found something specifically useful if you are a smaller or newer brand: ChatGPT showed the weakest correlation with traditional brand-authority signals of any of the three platforms tested. That does not mean ChatGPT is random. It means ChatGPT appears to be the most reachable of the major AI platforms for a brand that has not yet built the kind of search-volume and backlink profile that Google's AI Mode leans on.
Separately, Princeton's GEO-bench study found that adding citable statistics to a page improved its visibility in generative answers by roughly 41%, and that citing external sources produced a 115% visibility lift specifically for lower-ranked, less-authoritative pages — an equalizer effect that matters more to a small business than to a brand that is already dominant.
Put together: the fix is rarely "publish more." It is "get talked about, with evidence, somewhere other than your own site."
Reason 2: your site cannot be read
AI crawlers are separate from search engine crawlers. They follow their own rules, respect their own robots.txt entries, and are generally less tolerant of content that only appears after JavaScript runs. A site can be perfectly crawlable for Google and effectively blank for an answer engine.
How to check, in thirty seconds: open your page, view the raw source, and search for a sentence you can see on screen. If that sentence is not in the raw HTML, there is a good chance an AI crawler never sees it either.
Then open your robots.txt and look for entries naming AI crawlers specifically. Plenty of sites block them without anyone ever having made a deliberate decision to.
Keep this reason in proportion, though. A July 2026 crawl of nearly 10,900 domains by HasData's AI Crawler Block Index found that only 7.9% of ordinary websites block GPTBot in robots.txt — the figure jumps to 50.5% for news publishers specifically, who have a direct revenue reason to. For most small and mid-sized businesses, an outright robots.txt block is not the likely culprit. What is more common: 39.5% of the sites that do declare a GPTBot ban never actually enforce it at the server level, and llms.txt — the file some vendors push as an AI-visibility fix — sits at just 7.4% adoption and, per the same study, no major AI crawler has committed to reading it as authoritative. A missing llms.txt file is very unlikely to be your problem. JavaScript-only content and CDN challenge walls are far more common culprits than either of those.
Reason 3: nobody else describes you the way people ask
When a model assembles a shortlist, it leans heavily on third-party pages: review platforms, roundups, comparison articles, directories, YouTube videos. Your own site establishes what you are. Other people's pages establish that you belong in the set — and per the Ahrefs correlation data above, this is measurably the strongest lever available to you, well ahead of backlinks or additional pages.
How to check: ask Perplexity the question you care about and read the sources beneath the answer. It shows its working. If your brand is missing, you can see exactly which pages were used instead — and that list is your to-do list.
This is usually the highest-leverage gap, and the one companies are most reluctant to work on, because the fix happens on someone else's website: getting reviewed, getting included in a roundup, getting mentioned in a forum thread, getting talked about on video.
Reason 4: you describe yourself three different ways
If your homepage says one thing, your pricing page says another and your LinkedIn profile says a third, a model blending all three produces something vague. It is not confused about whether you exist. It is confused about what to call you.
How to check: ask an assistant what your company does. A clean one-sentence answer means your identity is clear. A hedged, padded answer that could describe four different companies means it is not.
The fix costs nothing but attention: pick one sentence describing what you do, and use it everywhere, word for word.
Reason 5: your competitors are simply better documented
Sometimes there is no technical problem at all. Your site is crawlable, your positioning is consistent, and the model still names three other companies — because those three have more coverage, more reviews and more independent pages describing what they do.
This is not a bug to fix. It is a gap to close, and per the correlation data above, it closes through accumulated third-party mentions and video coverage — not through a redesign, and not through publishing volume.
The four query types that reveal different problems
Not every prompt tests the same thing. Asking "what is [your brand]" tests something completely different from asking "what are the best options for [your category]" — and the two can produce wildly different results for the same company. Running all four types below, in order, tells you which of the five reasons is actually yours.
What this looks like for four different businesses
The five reasons play out differently depending on what you sell. Four examples, across four business types:
A local dental clinic
Direct brand queries work fine — the model knows the practice exists. Category queries ("best dentist in [city] for implants") come back empty. The clinic's Google Business Profile is solid, but almost no independent coverage exists beyond that: no local press mentions, no patient-written reviews with any detail, nothing on YouTube. This is Reason 3, and the fix is unglamorous: encourage detailed reviews (not just star ratings), get a mention in a local "best of" roundup, and consider a short video answering a common patient question — video mentions are the strongest single correlate in the Ahrefs data, and dental content performs well there.
A SaaS project-management startup
Comparison queries are the problem. Ask ChatGPT to compare project-management tools and it names Asana, Monday and ClickUp — never the startup, even though its product page is well-written and technically sound. The gap is Reason 5: those three competitors have years of accumulated third-party coverage — review-site pages, comparison articles, Reddit threads, YouTube walkthroughs — and the startup has almost none. The fix is not a better landing page. It is getting listed in review platforms (G2, Capterra), earning inclusion in "best PM tools" roundups, and seeding genuine comparison content that others can cite.
An e-commerce supplement brand
Problem-solution queries ("what should I take for X") never surface the brand, even though its products are relevant and its site converts well. Health and wellness answers draw heavily on third-party evidence because models are cautious with anything adjacent to medical claims — a brand's own product copy carries little weight here. This is Reason 3 compounded by category sensitivity: the fix is earning coverage on independent review sites and forums where real customers discuss results, since self-authored claims about health outcomes are exactly the kind of content models are built to discount.
A law firm
Direct brand queries return a vague, generic answer — "a law firm offering various legal services" — instead of the firm's actual specialty. The website uses different phrasing on the homepage, the attorney bios and the practice-area pages, none of which agree on what the firm is best known for. This is Reason 4, and it is the cheapest of the five to fix: pick one sentence describing the firm's specialty and use it verbatim across the site, the Google Business Profile, and any directory listings.
What to fix first
Not all five reasons deserve equal effort. In rough order of leverage relative to cost:
- Fix identity inconsistency (Reason 4) first. It costs nothing but attention and can be done in an afternoon.
- Check for technical blockers (Reason 2) second. Also fast to check, though for most non-publisher sites this turns out not to be the issue — verify rather than assume.
- Start earning third-party mentions (Reason 3) as an ongoing effort. This is the highest-leverage lever in the research, and the slowest to move — treat it as a standing priority, not a one-time task.
- Target video specifically. Given the strength of the YouTube-mentions signal, even a handful of third-party or owned videos addressing common questions in your category is disproportionately useful.
- Track the competitive gap (Reason 5) by category, not overall. You may already win some query types and lose others — fix the ones that matter most commercially first.
- Accept Reason 1 and move on. Do not spend a single hour trying to influence what a model already learned in training. Put that time into reasons two through five instead.
What not to assume
A few assumptions cause more wasted effort than the five reasons themselves.
Do not assume one bad answer means you are permanently invisible. Output is probabilistic. Ask the same question twice and you can get two different answers, naming two different sets of brands. That is not a malfunction — it is how these systems work. It is also why checking once tells you nothing useful. What tells you something is the same set of questions, asked repeatedly, over time.
Do not assume more content is the answer. Per the Ahrefs correlation data, the number of pages on your site barely correlates with AI visibility at all. Ten mediocre new pages will very likely do less for you than one genuine third-party mention.
Do not assume backlinks transfer the way they do in traditional SEO. They still matter for search rankings, but their correlation with AI-answer visibility is comparatively weak. Mentions — even unlinked ones — carry more weight in this specific context.
Do not assume every platform behaves the same way. The research above is ChatGPT-specific in places precisely because ChatGPT, AI Mode and AI Overviews weigh signals differently. A fix aimed at Google's AI Mode is not automatically a fix for ChatGPT.
Two things genuinely sit outside your control. Training data is already written — no amount of publishing changes what a model learned last year. And, as above, the output is probabilistic — a single check is a snapshot, not a verdict.
A twenty-minute check
Run these in order:
- Ask ChatGPT a direct-brand question about your company. Note whether it cites sources.
- Ask a category question — the "best options for X" version a real buyer would type.
- Ask a comparison question naming your closest competitor.
- Ask a problem-solution question describing what you actually solve, not your category label.
- View the raw source of your most important page and search for a sentence that is visible on screen.
- Open your robots.txt and look for entries naming AI crawlers.
By the end you will know which of the five reasons applies to you, and which query types are most affected. Four of the five reasons you can start on this week.
