Skip to main content

AI Search

Generative Engine Optimisation (GEO): What It Is and How It Works

What generative engine optimisation (GEO) is, how AI search engines choose sources, and how to get cited. See how AI search optimisation works.

  • 10 November 2026
  • 8 min read
  • By Kaan TURK
Kaan TURK, author

Generative Engine Optimisation (GEO) is the work of making a business easy for AI search engines to find, understand and cite. GEO builds on Search Engine Optimisation with a clear entity, consistent facts, passages that answer a single question and structured data, so ChatGPT, Perplexity and Google AI Overviews can recommend the business.

What is generative engine optimisation?

Generative Engine Optimisation is a digital marketing strategy for being named and cited inside AI-generated answers. Wikidata classes it as a marketing strategy and a branch of Search Engine Optimisation. Instead of aiming only for a ranked link, GEO aims for a mention or a citation inside the answer that ChatGPT, Gemini or Google AI Overviews writes for the searcher.

The name comes from a 2023 research paper by Pranjal Aggarwal and colleagues, published in the proceedings of ACM SIGKDD 2024 (read the paper on arXiv). The paper found that GEO methods can boost a source's visibility in generative engine responses by up to 40%. Read that as a best case from its experiments, not a promise for every site.

For a business, GEO turns a familiar question into a new one. The old question was where you rank. The new question is whether the AI answer names you at all, and which sources it trusted to get there. That second question is where my AI search optimisation work starts.

Is generative engine optimisation a real thing?

Yes. Generative Engine Optimisation is a real discipline with measurable results. A dental clinic I worked with had more than 500 five-star reviews but was missing from ChatGPT. A competitor was listed on a review directory that ChatGPT cited, and the clinic was not. Two weeks after I claimed and optimised the clinic's profile on that directory, ChatGPT listed it.

I found the gap by checking the sources ChatGPT cited when it recommended dentists for the clinic's main local search. The competitor appeared on BirdEye, a review platform, and the clinic had no profile there. So I claimed and optimised the clinic's BirdEye profile, and ChatGPT then started showing the clinic at the top for that search.

It is one case, and it ran alongside a 12-month SEO engagement in which 49 of the clinic's 93 tracked keywords reached Google's top three. So I treat it as evidence of a mechanism, not as a promise for every business: AI tools can only recommend a business they can find in the sources they trust. The full numbers are in the dental clinic AI search case study.

Case Study
2 weeks to ChatGPT recommendations
Dental Clinic AI Search
A clinic with 500+ five-star reviews, missing from ChatGPT until its profile on a cited review directory was claimed and optimised.
Read full case study →

How does generative engine optimisation work?

Generative Engine Optimisation works on the sources AI answers are built from. Google's AI features use pages from Google's own index, while OpenAI and Perplexity send their own crawlers to surface websites. GEO makes sure those systems can reach your site, recognise your business as one clear entity and find a passage that answers the question.

For Google, the rule is explicit. To appear as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet. OpenAI uses OAI-SearchBot to surface websites in ChatGPT search, and PerplexityBot does the same job for Perplexity. These are the four steps I use; the order changes with the gap the baseline finds.

Baseline: what AI answers say about you today

I start by asking ChatGPT, Perplexity and Google AI Overviews the questions a real buyer would ask. For each answer I record whether the business is named, which competitors are named and which sources the answer cites. The sources tell you where to work. The full method is in my guide on how to measure AI visibility.

Entity: one clear set of facts everywhere

Your business should look like one entity wherever it is described: the same name, role, address and description on your website, your Google Business Profile and the directories that cover your industry. The dental clinic shows why the directories matter. Its reputation was strong, but it was missing from a directory ChatGPT cited, and that gap kept it out of the answer.

Citability: passages that answer one question each

In my experience, traditional rankings depend heavily on a website's overall authority and SEO fundamentals, while AI search looks for an information block about one specific attribute. So each question on a page gets its own answer, in the first sentence under its heading, naming the subject instead of "it" or "this". Semantic SEO is how I structure those blocks.

Access: AI crawlers, structured data and llms.txt

The most common mistake I see is blocking AI crawlers through robots.txt or Cloudflare, which leaves a website with low visibility in AI answers. Cloudflare announced in 2025 that it became the first internet infrastructure provider to block AI crawlers by default, so if your site sits behind Cloudflare, check its AI crawler setting. In robots.txt, the search crawlers are the ones to allow:

User-agent: OAI-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

You can still disallow GPTBot, which OpenAI uses for training, without leaving ChatGPT search. Google-Extended is a separate token for whether Google may use your content to train Gemini models. Add structured data that matches the visible page, and treat llms.txt as optional.

Which AI search engines does generative engine optimisation cover?

Generative Engine Optimisation applies to every AI system that writes answers from web sources: ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, Gemini and Microsoft Copilot. These tools do not share one index. Google's features draw on Google's index, OpenAI uses OAI-SearchBot and Perplexity uses PerplexityBot, so I check each tool separately.

I do not believe in a universal GEO checklist. There is no single best practice; what works depends on the query type. A local question about a dentist leans on directories and reviews, while an informational question leans on clear, well-structured explanations. What I use instead is a set of diagnostic questions.

There is no single GEO best practice, so I start with four questions. Access: can AI crawlers read the site? Clarity: is the value proposition plain? Structure: are there structured data and answer blocks? Authority: what do other sites say about the business?

Is SEO dead now with AI?

No. Search Engine Optimisation is the foundation that GEO stands on. Google says the best practices for SEO continue to be relevant because its generative AI features are rooted in its core ranking and quality systems. A page has to be indexed before it can be quoted, so GEO extends SEO rather than replacing it.

What has changed is that ranking well and being cited are no longer the same result. A page can rank well in the classic results and still be missing from AI Overviews or AI Mode, and the opposite happens too. In my view GEO and SEO are not that different: the fundamentals are the same, and the techniques differ. For GEO I look for the real reason a competitor is cited and you are not. I compare the two in detail in GEO vs SEO.

Getting help with generative engine optimisation

If your business ranks on Google but rarely appears in AI answers, the gap usually has a specific cause: a blocked crawler, a missing profile on a source AI tools cite, or a page that buries its answer. I am Kaan TURK, an SEO & AI Search Consultant, and I run GEO as part of one plan for Google and AI search. You can see how I approach it on my AI search optimisation page.

Frequently Asked Questions

What does GEO stand for in SEO?

GEO stands for generative engine optimisation, not geographic targeting. It is the part of Search Engine Optimisation that aims for mentions and citations inside answers written by AI tools such as ChatGPT and Google AI Overviews. The two share the same foundation; GEO adds techniques for being quoted rather than only ranked.

How long does generative engine optimisation take?

GEO can show results within weeks for a single source fix: the dental clinic appeared in ChatGPT recommendations two weeks after its review profile was claimed and optimised. Entity and content work compounds over months, in line with SEO, and no one can promise when an AI tool will start citing a page.

Do you need llms.txt for generative engine optimisation?

No. An llms.txt file is optional and, in my view, low impact today. In my experience, crawler access, one consistent entity and mentions on sources AI tools trust matter far more for whether a business gets cited. I explain what the file does and does not do in my guide to llms.txt.

generative engine optimisationGEOAI searchChatGPTGoogle AI Overviews
Share:
Kaan TURK
Kaan TURKAbout
SEO & AI Search Consultant

15 years of SEO expertise. Former SEO Lead for Louis Vuitton, LC Waikiki and Vakko. Mathematics graduate bringing data-driven precision to search engine optimisation.

Ready to Implement These Strategies?

Get expert help implementing these SEO strategies for your business. With 15 years of experience and work with brands like Louis Vuitton, I can create a customized plan tailored to your goals.

If you want to know what is holding your site back, the first step is a discovery call. Before the call I look at your site and your competitors, and on the call I show you what I would fix first.

Book a discovery call