How AI tools describe your business, and what is actually known.

ChatGPT, Gemini and Perplexity are separate answer systems, and Google has no documentation on how the other two describe a brand. What the vendors do publish is narrower: how their crawlers work and how their search features show sources. OpenAI's crawler documentation says OAI-SearchBot is used to surface websites in ChatGPT's search features, that sites opted out of it will not be shown in ChatGPT search answers, and that GPTBot, which relates to training, is a separate setting. OpenAI's help article adds that responses using web search may include citations, and that results and citations can be incomplete, outdated or incorrect. Perplexity's crawler documentation says PerplexityBot is designed to surface and link websites in its search results. For Gemini, Google's common crawlers page says the Google-Extended token covers use of crawled content for training Gemini models and for grounding in Gemini Apps, and that it does not affect inclusion in Google Search.

What is not documented is just as important. None of those pages explains how an assistant decides what to say about a particular business, or promises that allowing a crawler leads to a mention. We treat answer engines as probabilistic and changing often, so a single screenshot is an anecdote rather than a finding. Whether crawlers can reach your pages is something you control; how an assistant phrases a description is not something any vendor tells site owners how to control, so a review has to look at the output itself. OpenAI also says robots.txt rules may not apply to ChatGPT-User, which visits pages for certain user actions.

Cultured Digital covers this as the first part of a visibility review: how ChatGPT, Gemini and Perplexity describe you, recorded as what the tools actually said, with prompts and dates.

Google does document its own AI features. Its AI features and your website guide covers AI Overviews and AI Mode and says the best practices for SEO remain relevant: There are no additional requirements to appear in AI Overviews or AI Mode. To be eligible as a supporting link, a page must be indexed and eligible to be shown in Google Search with a snippet. The guide adds that meeting the requirements does not guarantee that Google will crawl, index or serve a page.

Google also says AI Overviews are shown only when its systems judge them additive to classic Search, and that they often do not trigger. Both features may use a "query fan-out" technique, issuing multiple related searches across subtopics, and they may use different models and techniques, so the links they show vary. Site owners can limit what is shown from their pages with nosnippet, data-nosnippet, max-snippet or noindex, and robots.txt directives for Googlebot remain the control over crawling for Search.

Search Console's help page describes AI Mode as grouping the user's question into subtopics and searching for each one at the same time, with links to web resources that support the information, and it treats a follow-up question in AI Mode as a new query. In the documentation we read, Google is the only one of these vendors that gives site owners reporting on how their pages appear in its AI answers.

Cultured Digital covers Google's AI search experiences in the visibility review, starting from what Google documents rather than from speculation about special optimisation. How these experiences appear in reporting is covered on the measurement page.

When a competitor is cited in an AI answer and you are not, the reason is rarely visible in the answer itself. Google's documentation gives some conditions: a cited page has to be indexed and eligible for a snippet, and because of query fan-out the supporting links can come from searches on related subtopics rather than only the question as typed. One reasonable inference, and it is an inference rather than something Google states, is that a competitor may be cited because it has a strong page on a subtopic you have not covered.

For ChatGPT and Perplexity, the documentation we read describes crawlers and citations but does not explain why one source is chosen over another. OpenAI says only that citations may appear and should be checked. So the "why" falls into three groups: what can be observed, such as who is cited for which prompts; what can reasonably be inferred; and what cannot be established.

Two Google pages bear on the comparison. The spam policies treat attempts to manipulate generative AI responses in Google Search as spam, so copying a competitor's apparent tactics is not automatically safe. And the people-first content guidance asks whether content provides substantial value compared with other pages in search results, which is a sensible test to apply when you put your page next to the one that is cited.

Cultured Digital covers where competitors are cited, and why, in the visibility review. We state which parts are observed, which are inferred and how confident we are, and which are unknown. Where the gap turns out to be authority rather than content, it connects to link gap analysis.