AI search optimisation: how to get cited by ChatGPT, Google AI and Perplexity

AI search optimisation means making your pages easy for AI answer engines to find, trust and quote. In practice that is solid SEO plus four extras: let the AI crawlers in, write answers that stand on their own, make your business a clear entity, and earn genuine mentions on sites these tools already trust. No special file replaces that.

The labels keep multiplying. GEO (generative engine optimisation), AEO (answer engine optimisation) and “LLM SEO” all describe the same goal: being one of the sources an AI tool cites when it answers a question in your category. Google’s own documentation now says plainly that, for Google Search, optimising for generative AI features is still SEO. Other tools work a little differently, and the differences matter for your robots.txt file and your measurement. This guide covers how each major tool chooses sources, what actually moves citations, where llms.txt fits, and how to track results. It is the same approach we use in our SEO services, where AI citations are reported alongside clicks and leads.

Key takeaways

• Google says AI Overviews and AI Mode draw on its core Search ranking systems. A page has to be indexed and eligible for a snippet before it can be cited.
• ChatGPT search and Perplexity use their own crawlers. Blocking OAI-SearchBot or PerplexityBot removes you from those answers, while blocking the training crawlers does not.
• llms.txt is a proposal, not a standard. Google Search ignores it. It costs little to add, so treat it as optional housekeeping, not a ranking lever.
• Pages that get cited answer a question in the first lines, use specific facts, and come from a business that is clearly identified across the web.
• Measure with Search Console’s generative AI report and Bing Webmaster Tools’ AI Performance report, then judge success by enquiries, not citation counts alone.

What is AI search optimisation, and how is it different from SEO?

Classic SEO aims for a ranked blue link. AI search optimisation aims for a mention or a cited link inside a generated answer. The two overlap heavily, because most AI answer engines retrieve pages from a search index first and then write a summary grounded in those pages. If your page is not in the index, or is not among the results retrieved, it cannot be quoted.

Google is direct about this. Its guide to optimising for generative AI features on Google Search, first published in May 2026, says AI Overviews and AI Mode rely on Google’s core ranking systems, that “AEO” and “GEO” are SEO for the search experience, and that content which is not technically sound or helpful enough to rank is unlikely to appear in AI answers.

What changes is the shape of success. A user who reads an AI answer may never click. Pew Research Center’s 2025 browsing study of 900 US adults found people clicked a traditional result in 8% of visits when an AI summary appeared, against 15% when it did not, and clicked a link inside the summary in only 1% of visits. That is why being named in the answer, with your brand and your point of view intact, now matters as much as the click.

How do ChatGPT, Google AI Overviews, AI Mode and Perplexity choose sources?

Each tool has its own crawler, index and controls. The table summarises what site owners can act on.

Google AI Overviews and AI Mode

Both features use retrieval from the Search index. Google describes “query fan-out”: the model fires off several related searches to gather more information, which is why a page can be cited for a question it never targeted word for word. Google also warns that creating separate pages for every fan-out variation, mainly to manipulate AI answers, falls under its scaled content abuse spam policy. Preview controls such as nosnippet and max-snippet limit how much of your page can be shown, so check they are not applied by accident on pages you want cited.

ChatGPT search

OpenAI documents three relevant crawlers on its bots page. OAI-SearchBot surfaces websites in ChatGPT’s search features, and OpenAI states that sites opted out of it will not be shown in ChatGPT search answers. GPTBot collects content that may be used to train models, and it is controlled separately, so you can allow OAI-SearchBot and block GPTBot. ChatGPT-User makes visits that a person asks for during a conversation and is not used to decide what appears in search.

Perplexity

Perplexity’s crawler documentation recommends allowing PerplexityBot in robots.txt if you want to appear in its results. Perplexity-User fetches a page when a user asks a question and generally ignores robots.txt because the request is user-initiated.

Microsoft Copilot

Copilot and Bing’s AI summaries ground their answers in the Bing index. That makes Bing Webmaster Tools worth setting up even if Bing sends you little direct traffic, because it is currently the clearest window into AI citations (more on measurement below).

Step 1: let AI crawlers reach your pages

Many sites lose AI visibility through a firewall rule, a security plugin or a CDN bot setting that nobody remembers switching on. Before writing anything new, run this technical checklist:

1. Open your robots.txt and confirm Googlebot, Bingbot, OAI-SearchBot and PerplexityBot are not disallowed from the pages you want cited.
2. Decide separately whether to allow training crawlers such as GPTBot. Blocking them is a business choice and does not remove you from ChatGPT search answers.
3. Check your CDN or hosting firewall. Bot-fight modes and rate limits can block AI crawlers even when robots.txt allows them.
4. Make sure key content is in the HTML as text, not locked in images, PDFs or content that only appears after a click.
5. Confirm important pages are indexed in Google Search Console and Bing Webmaster Tools, and that none carry a stray noindex or nosnippet.
6. Keep pages fast and stable. A slow, script-heavy page is harder for every crawler to render.

If your site runs on WordPress, our technical SEO checklist for WordPress walks through each setting. For sites with deeper crawl problems, our technical SEO service starts with a crawl of exactly these issues.

Step 2: write pages that are easy to quote

AI answers are built from passages. A passage gets used when it answers the question clearly and can be lifted without the surrounding page. Google says you do not need to break content into tiny chunks or rewrite it for machines, and that its systems understand synonyms. What helps is ordinary editorial discipline:

• Answer first. Put a direct, two to four sentence answer under each question-style heading, then add detail.
• Be specific. Prices, dates, measurements, named standards and real examples are more quotable than general claims.
• Add what others do not have. Google’s guide asks for a unique point of view and “non-commodity content”. Your own data, process photos, case results and local rules are hard for anyone else to copy.
• Use tables for comparisons. A table of options, costs or specifications is easy to extract and easy to verify.
• Show who wrote it. Named authors with real credentials, and an about page that explains your business, support trust for every engine.

This is the core of our content marketing work: fewer, better pages built around the questions buyers actually ask.

Step 3: entity and schema work

AI systems do not just match words; they try to understand which business, product or person a page is about. An “entity” is that thing, with consistent facts attached: name, address, founding date, what it sells, where it operates. The more consistent those facts are across the web, the easier it is for any engine to connect a question to your brand.

Practical entity work looks like this:

• One canonical description. Write a two-line description of your business and use it on your about page, Google Business Profile, LinkedIn page and directory listings.
• Organization structured data. Google’s Organization markup documentation recommends properties such as url, logo, sameAs (links to your official profiles), address, contactPoint, legalName and foundingDate, placed on the home page or a single about page.
• Consistent name, address and phone. Mismatched details across listings make it harder to tell your business apart from similarly named ones.
• Product, service and FAQ markup where it is true. Structured data must match the visible text on the page.
• Author pages. For health, finance and legal topics, author pages with qualifications matter a great deal. For a UK private medical practice we worked with, clinical author markup and visible credentials were central to winning searches in a Your Money or Your Life category.

Be clear about one point: Google says structured data is not required for its generative AI features and there is no special schema for them. Schema helps machines read your facts accurately and keeps you eligible for rich results. It is not a switch that turns on citations.

Does llms.txt help with AI search?

llms.txt is a proposal published by Jeremy Howard in September 2024. According to the llms.txt specification, it is a Markdown file at the root of a site with a required H1 (the site name), a short summary in a blockquote, and sections of links to the pages an AI agent should read, plus an optional section for secondary material. The intent is to give AI agents a short, curated map of a site at the moment they need it.

The evidence that it changes visibility is thin. Google’s generative AI guide says you do not need new machine-readable files, AI text files or Markdown versions of pages to appear in Google Search, and that Google Search ignores llms.txt, so it neither helps nor harms there. Other services may read it, and agent-style browsing tools are starting to look for it.

Our view: if you have a documentation-heavy site, a SaaS product or a large knowledge base, a well-maintained llms.txt is cheap and may help agents navigate. For a typical business website it is low priority. Do it after crawl access, content quality and entity consistency, never instead of them.

Step 4: earn genuine mentions on sites AI tools already trust

AI answers often cite a mix of sources: brand sites, publishers, review platforms, forums and comparison pages. If your brand never appears outside your own site, you are relying on one page to carry the whole case.

Useful mentions come from:

• Trade and industry publications that cover your category
• Original data or a short survey that journalists and bloggers can quote
• Expert commentary on regulation or market changes in your field
• Genuine reviews on Google and the platforms your buyers use
• Partner, supplier and association listings with accurate details

Google’s guide is explicit that chasing inauthentic mentions is not as helpful as it might seem, because AI features run through the same spam systems as core ranking. Planted forum posts and paid “listicle” placements tend to be short-lived and can damage trust with human readers too.

How do you measure AI search visibility?

Measurement improved a lot in 2026, though it is still partial.

Google announced the generative AI performance reports in June 2026 and rolled them out to all websites by the end of August. Microsoft launched AI Performance in Bing Webmaster Tools as a public preview in February 2026. Google’s guide also warns that no third-party tool has access to its internal ranking or AI systems, so treat “AI visibility scores” from vendors as estimates.

Report AI visibility next to the numbers that pay the bills: organic enquiries, calls and sales. A rise in citations that brings no new business is a vanity metric.

What does this look like in practice?

Three of our clients show how GEO builds on ordinary SEO rather than replacing it:

• Digiliance, a labour-law compliance SaaS, grew from 125 to 7,120 organic clicks a quarter after a rebuild, technical fixes and content structured with direct answers, definitions, tables and FAQ schema. Its AI citations in Bing rose from 5.8K to 14.8K. See the Digiliance case study.
• Standard Cold Press, a cold-pressed oil brand, built entity-based topic clusters around oils, uses and health questions. Bing’s AI Performance report shows 31.2K AI citations across an average of 29 cited pages. See the Standard Cold Press case study.
• ArtnovaAI, a new AI image tool, combined a technical clean-up with prompt guides structured for quoting. Organic clicks grew from 3K to 14.1K in six months and Bing AI citations grew five times. See the ArtnovaAI case study.

In each case the technical and content basics came first. The AI citations followed the same pages that were already earning search visibility, which is the strongest argument against shortcuts such as AI-only page copies or mass-produced question pages.

Frequently asked questions

What is generative engine optimisation (GEO)?

Generative engine optimisation is the work of making your content more likely to be cited or mentioned in AI-generated answers from tools such as ChatGPT, Google AI Overviews, AI Mode, Perplexity and Copilot. It builds on SEO: crawlable, indexed pages with clear, specific answers, plus a consistent brand identity and genuine mentions on trusted third-party sites.

How do I get my website cited by ChatGPT?

Allow OAI-SearchBot in robots.txt and in any firewall or CDN settings, because OpenAI says sites that block it are not shown in ChatGPT search answers. Then make sure your pages answer common buyer questions directly, are indexed by major search engines, and that your business is described consistently across reputable sites. Blocking GPTBot only affects model training.

Is llms.txt worth adding to my website?

It is optional. llms.txt is a proposed Markdown file listing the pages an AI agent should read. Google Search ignores it, and no major AI search tool has made it a requirement. For documentation-heavy or SaaS sites it can help agents navigate. For most business websites, crawl access and content quality matter far more.

Do I need schema markup to appear in Google AI Overviews?

No. Google says structured data is not required for its generative AI features and there is no special markup for them. Organization, product and FAQ markup still help machines read your facts accurately and keep you eligible for rich results, provided the markup matches what is visible on the page.

How can I track whether AI tools mention my business?

Use the generative AI performance report in Google Search Console for AI Overviews and AI Mode impressions, and Bing Webmaster Tools’ AI Performance report for Copilot citations. Add referral sources such as chatgpt.com and perplexity.ai in your analytics, and check your ten most important buyer questions in each tool every month.

How long does it take to appear in AI answers?

It depends on how strong your organic visibility already is. Pages that already rank well can start being cited within weeks of improvements. A new site needs the same months of crawling, indexing and authority building that classic SEO needs, because most AI answers draw from pages that search engines already trust.

If you want to know whether AI tools can read and cite your site today, we will check crawler access, indexing, entity consistency and your most important buyer questions as part of a free SEO audit. You can also see how AI visibility fits into our wider SEO services.

Sources

• developers.google.com/search/docs/fundamentals/ai-optimization-guide
• developers.google.com/search/blog/2026/06/gen-ai-performance-reports
• blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-
• developers.openai.com/api/docs/bots
• docs.perplexity.ai/guides/bots
• llmstxt.org
• developers.google.com/search/docs/appearance/structured-data/organization
• pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links