Why doesn't ChatGPT recommend my business?

ChatGPT recommends businesses it has seen named, described consistently and reviewed on sources it trusts. If it never mentions yours, there are five usual reasons: your site is invisible to AI crawlers, third-party sources barely mention you, your brand data is inconsistent, your pages never answer a question directly, or rivals dominate the sources it cites. All five are measurable and fixable, and fixes usually show in AI answers within 4 to 12 weeks. This is the job of AEO (answer engine optimisation).

How ChatGPT picks businesses

When someone asks ChatGPT “who’s the best accountant in Bristol?”, the answer is assembled from three places: what the model learned in training, what ChatGPT’s search feature pulls from the web (which leans on Bing’s index and OpenAI’s own OAI-SearchBot crawler), and pages fetched live while answering. You are competing to be present, and well spoken of, in all three.

None of these work like Google rankings. A July 2025 Semrush study of more than 500 marketing topics found that when ChatGPT search cites a page, that page ranks in position 21 or lower in classic search for related queries almost 90% of the time (Semrush). Ranking well and being recommended by AI are related, but they are not the same job.

Reason 1: AI crawlers cannot read your site

Vercel and MERJ’s December 2024 analysis of AI crawler traffic found that none of the major AI crawlers render JavaScript, with the exception of Gemini, which runs on Googlebot (Vercel). If your site only shows its content after scripts run, GPTBot and the other AI crawlers see an empty shell. The same analysis found ChatGPT and Claude spending about 35% of their fetches on dead links, so broken URLs and stale sitemaps hurt more with AI than they ever did with Google.

The fix is unglamorous: content in the initial HTML, working URLs, an accurate sitemap, and a robots.txt that does not block the AI search crawlers: OAI-SearchBot, Claude-SearchBot and PerplexityBot. The training bots, GPTBot and ClaudeBot, are a separate choice.

Reason 2: nobody else talks about you

ChatGPT trusts third parties more than it trusts you. Profound’s analysis of 680 million AI citations (August 2024 to June 2025) found Wikipedia was ChatGPT’s most-cited source, at 7.8% of all its citations; among its ten most-cited sources, Wikipedia took 47.9% of those citations, Reddit 11.3% and Forbes 6.8% (Profound). A 2026 Quoleady study of SaaS “alternatives” searches found that 100% of the tools ChatGPT mentioned had reviews on Capterra and 99% had reviews on G2 (Quoleady).

Ahrefs’ May 2025 study of 75,000 brands points the same way: branded web mentions correlated with visibility in Google AI Overviews at 0.664, against 0.218 for backlinks (Ahrefs).

Reason 3: your brand data is inconsistent

Models hedge on brands whose name, description, category or pricing differ between their website, LinkedIn, directories and reviews. If you are “Smith & Co Chartered Accountants” in one place, “SmithCo” in another and “Smith Accounting” in a third, you have split your own vote. Pick one name and one description, and repeat them everywhere.

Reason 4: your content never answers anything

AI answers are assembled from extracted passages, and extraction favours pages that answer questions directly: a question-shaped heading, a plain answer under it, tables, and named statistics. The Princeton GEO study, peer-reviewed at KDD 2024, found that methods such as adding statistics, quotations and citations lifted visibility in generative engine answers by up to 40% (GEO paper). Fluent prose helped; keyword stuffing did not.

Reason 5: your rivals got there first

On commercial questions, AI engines lean heavily on comparison articles and “best X” listicles. Wix Studio’s AI Search Lab analysed over a million citations across ChatGPT, Google AI Mode and Perplexity in March 2026 and found listicles took 40.9% of commercial-intent citations, against 21.9% overall (Wix Studio). If every listicle in your category names three rivals and not you, the model repeats what it reads. Getting into those articles, or publishing an honest one yourself, changes what it repeats.

Does schema markup fix this?

Not on its own, and be wary of anyone who says otherwise. There is no solid evidence that schema markup directly increases LLM citations (one large study found no correlation). It still helps Google rich results and Bing’s understanding of your pages, and Bing feeds ChatGPT, so ship it, but as plumbing, not as the strategy.

What to do this month

1. Measure first. Find out what all six engines say about you today, against your rivals. The free AEO score does this for you.
2. Open the door: allow the AI crawlers in robots.txt, fix dead URLs, get your content into the initial HTML, register with Bing Webmaster Tools.
3. Answer questions on your pages: question-shaped headings, direct answers, tables, real numbers.
4. Get mentioned off-site: review platforms for your category, consistent listings, useful community participation, disclosed.
5. Watch it weekly. AI answers move over weeks and months, though ChatGPT search and Perplexity can shift within days on narrow questions. Whatever you change, you need the before and after.

How long do fixes take to show in AI answers?

Plan on 4 to 12 weeks from a fix going live to seeing it in AI answers. On our own site ChatGPT and Perplexity named us first, sometimes within days on narrow questions, while Google AI Overviews took 27 days. Of the fixes we shipped in August 2026, 63% got us named within five weeks, and those that moved took a median of 5 days to the first mention. The phases, the engine-by-engine timings and the full log are in how long AEO takes, and the short version is quick tip 22: publish, wait, re-ask, and judge at week eight.

Most published numbers in this niche come from vendors, including some above; treat magnitudes as directional. The directions, though, all point the same way. Figures checked 23 September 2026; re-checked quarterly.