How do I make my company visible in AI?
Published on
To make your company visible in AI, three things matter in our measurements: that their crawlers can read your site, that sources they already consult talk about you, and that your pages precisely answer your customers' questions. The rest carries little weight: schema.org markup, the Q&A format, or the llms.txt file do not get a page cited. On a question that does not name you, ChatGPT, Claude, and Gemini mostly choose third-party sources: your site is necessary; it is not enough.
For what type of question? It all depends on the question your customer asks. If they name your company, the AI goes looking on your site, and your site does most of the work. If they ask a market question, such as choosing a service provider or a product, the AI chooses its sources itself: that is where most companies are absent, and it is mainly this case that this page covers.
What AI publishers and studies say
Google states that there is no additional requirement to appear in its AI answers, AI Overviews and AI Mode: a page must be indexed and able to be displayed with a snippet, and these features rely on the same ranking and quality systems as classic search. At OpenAI, a site must let the OAI-SearchBot crawler through to appear in ChatGPT's search answers. No publisher sells a place in the answers, and no one can guarantee being cited.
Studies converge on the weight of external sources. Across more than 25 million links cited by ChatGPT, Claude, and Gemini, Muck Rack counts 84% of so-called earned sources in May 2026: press, encyclopedias, research, institutions, and content from other companies. The link with Google ranking, on the other hand, is weak: according to Ahrefs in August 2025, only 8% of the pages cited by ChatGPT appear in Google's top 10 for the same query. Ranking well on Google remains useful for Google's AI answers; it is not enough for the others.
Three levers that can be measured
The first is access. A site closed to an AI's search crawler cannot be cited by it: in our measurements, none of the sites cited by ChatGPT blocked OAI-SearchBot. It is the simplest check, and the first one to do, because a setting inherited from an old security policy is enough to make a site invisible to an AI, whatever the quality of its pages.
The second is to exist in the sources that AIs already consult. In 72 to 91% of its searches according to our corpora, ChatGPT writes the name of a publisher it knows, and it cites only about one page in ten of what its search returns. A company that the AI has never seen named is neither searched for, nor found, nor cited. Being mentioned in the media, in specialist publications, or on third-party pages in your field is therefore what brings it into the AI's field of view. The third lever is the page itself: with ChatGPT, a page whose title repeats the question is cited 1.3 to 1.9 times more often, and with Claude, a paragraph of 50 to 100 words is cited 3.7 to 4.8 times more often than a short paragraph.
What carries little weight, despite what you read
Much widespread advice does not hold up when measured. Schema.org markup describes a page without getting it chosen: Ahrefs, by comparing pages that added it to control pages, measured no net increase in citations, and between 12 and 31% of the pages cited by Claude and Gemini carry none. The Q&A format does not help any more: between 74 and 93% of the pages cited by Claude and Gemini have a title that is not a question. The llms.txt file has no demonstrated effect, and Google writes that it is useless for its search.
Freshness counts a little, not always: with Claude, a page less than six months old is cited 1.2 to 1.3 times more often, but showing the year in the title lowered ChatGPT's citations on our French-language corpora. Reddit, YouTube, and Wikipedia are very present in what ChatGPT reads, but their pages returned by its search were cited between 0 and 1.5% of the time depending on the corpus. These levers are not useless in themselves; they replace neither access, nor third-party sources, nor a page that answers.
The solutions that exist, and what each one does
Four families of solutions are on offer. Tracking tools, from Peec AI to Semrush or Ahrefs, measure whether ChatGPT, Gemini, or Claude cite your company, for €20 excl. VAT to around a hundred dollars per month at the entry level. GEO agencies carry out an audit of your presence in AIs, most often between €1,500 and €5,000 excl. VAT, then ongoing support of €950 to €2,500 excl. VAT per month for an SME, according to industry guides. Press relations seek articles in the media: a press release counts mainly through the pickups it obtains, since press releases themselves account for only 1.1% of the cited links recorded by Muck Rack.
The fourth family is distribution: a service writes and publishes regularly for you, on your site and on third-party sources, where AIs go looking when the question does not name you. These solutions do not do the same work, and they combine: a tool tells you where you stand, an agency corrects and advises, press relations and distribution change what there is to read. If you do not know where you stand, start by measuring; if you already know that AIs do not cite you, one more measurement will change nothing.
What we do not know
Our measurements cover the API of GPT-5, Claude Sonnet 4.6, and Gemini 3.5 Flash with web search, not their applications nor Google's AI Overviews. We observe what distinguishes cited pages from discarded pages; we have not measured the effect of a publication on a company's citations.
Sources
- Google Search Central — “AI features and your website”, .
- Google Search Central — “Optimizing your website for generative AI features on Google Search”, .
- Muck Rack — “What is AI Reading?” (Generative Pulse), .
- Ahrefs, Louise Linehan — overlap between AI search and top 10, .
- Ahrefs, Louise Linehan — adding structured data and AI citations, .
- FaireDuBruit measurements, from September 4 to 16, 2026 — see “How we measured”.