Found on AI Search: What to Do and Ignore

Advice on getting a business found on AI search repeats the same list: optimise for AI, add schema, add an FAQ, write question headings, keep your details consistent, complete your Google Business Profile. Some of it is sound. Some has no support in anything the platforms publish.

This article tests that list against the documentation from Google, OpenAI and Perplexity. Where a source is silent we say so, and where we are giving our own reading rather than a platform’s wording, we say that too.

What the documentation says gets a page in

Google’s guide to optimizing for generative AI features says its AI Overviews and AI Mode rely on its core Search ranking and quality systems, and use techniques such as grounding in pages retrieved from the Search index and “query fan-out”, which issues several related searches to build an answer. To be eligible as a supporting link, a page must be indexed and eligible to be shown in Search with a snippet. The AI features page says there are no additional technical requirements.

The guide adds that a site must also be included in Google’s generative AI features through a Search Console setting. The Search generative AI control documentation says inclusion is the default for every property, and that excluding a site removes its links and content from those features. Meeting every requirement does not guarantee crawling, indexing or serving.

For ChatGPT, OpenAI’s crawler documentation says OAI-SearchBot surfaces sites in ChatGPT search, that sites opted out of it will not be shown in search answers, and that it recommends allowing it in robots.txt. Perplexity’s crawler documentation makes the same recommendation for PerplexityBot, and notes that a web application firewall may need a rule to let it through. Our reading: before any tactic, check that crawlers are actually let in. Our article on what each AI crawler setting does covers the detail.

Does AI search need its own optimisation?

Google’s answer is that optimising for generative AI search is optimising for the search experience, and so still SEO. Its guide has a section on what you do not need to do, and it names five things: llms.txt files and other special markup, “chunking” content into tiny pieces, rewriting content just for AI systems, seeking inauthentic mentions, and over-focusing on structured data. It says there is no ideal page length.

It also says that creating and keeping an llms.txt file is fine if other services use it, and that doing so neither harms nor helps visibility in Google Search, which ignores such files. The same guide tells readers to evaluate third-party “AEO” and “GEO” advice against official guidance, and Google’s page on third-party SEO advice adds that third-party tools do not have access to its internal ranking data and cannot guarantee performance.

Schema, FAQ sections and question headings

On schema, Google’s position is that structured data is not required for its generative AI features and that no special schema.org markup is needed, although it still recommends structured data as part of normal SEO for rich results. Markup should match the visible text on the page.

FAQ sections need care. Google’s changelog records that the FAQ rich result stopped appearing in Google Search from 7 May 2026, and that its documentation was removed in June. Our reading: an FAQ is still a good way to answer real customer questions for people, but we found no Google documentation that FAQ markup, or an FAQ section as such, helps a page appear in AI answers.

The same goes for question-style headings and answering in the first sentence. Google asks that content be organised in a way that helps readers, with paragraphs, sections and headings giving a clear structure, and that is good writing. It does not present those choices as a lever for AI visibility.

Content that gives an answer something worth citing

This is where Google puts the weight. Its guide says unique, valuable, non-commodity content will likely influence presence in generative AI search more in the long run than any other suggestion in it. It contrasts a first-hand, experienced view with a summary that restates what is already available, and says not to recycle what others have said or what a generative AI tool could easily produce.

It also warns against creating separate content for every variation of how people might search, including fan-out queries, when done primarily to manipulate rankings or AI responses, because that violates its scaled content abuse policy. It recommends supporting text with high-quality images and video where they help.

Local businesses: the Business Profile and the rest of the web

For local questions the documentation is more concrete. Google’s guide says Merchant Center and Business Profiles can help products and services be visible in AI responses and other Search results, and the AI features page lists keeping that information up to date as a best practice. The Business Profile help on local ranking says businesses with complete and accurate information are more likely to appear in local results, that local ranking rests mainly on relevance, distance and prominence, and that prominence draws on links and reviews. Google’s page on establishing business details adds verifying the site in Search Console.

OpenAI’s ChatGPT search help says ChatGPT may use an approximate location, and may share location with third-party providers to supply nearby business listings. It does not name those providers in that section. Our reading: keep your details accurate wherever you are listed. Popular advice that identical name, address and phone details across the web decide AI answers is not something we found any platform document, and Google says that seeking inauthentic mentions is less helpful than it seems.

Can you pay, and how long does it take?

Google’s Business Profile help says there is no way to request or pay for a better local ranking. The documentation we read says nothing about paid placement inside AI answers, so we make no claim either way. On timing, Google says crawling can take anywhere from several days to several months depending on how often its systems decide a page needs refreshing, and OpenAI says a robots.txt change can take about 24 hours to be picked up. Neither gives a timeline for being cited, and OpenAI notes that ChatGPT search results and citations can be incomplete, outdated or incorrect.

Checking where you stand

Google’s generative AI performance report in Search Console shows impressions in AI Overviews and AI Mode, grouped by page, country, date and device. It may not appear if the site has not received enough impressions. We cover how this sits alongside analytics in where AI-driven visits show up. For ChatGPT and Perplexity we found no equivalent report in their documentation. Our reading: server logs are the place to see whether their documented crawlers are visiting.

Where this fits at Cultured Digital: AI search tested against the documentation

Our AI & Search Visibility page says there is no “AI SEO” trick, only fundamentals pointed at new surfaces, and that no one can guarantee citations. The visibility review looks at how ChatGPT, Gemini and Perplexity describe a brand, and measurement covers AI referral traffic and the limits of reporting. We improve the conditions that make citation more likely and measure what changes. Related reading: the five fundamentals and what Google says about AI Overviews.

Getting your business found on AI search: what to do and what to ignore