What Is Generative Engine Optimization (GEO)? Definition, Origin and How It Relates to SEO

Generative engine optimization (GEO) is the work of making a website's content visible in the answers of generative engines: search systems that use language models to gather information from several sources and summarize it into one answer. Today that includes Google's AI Overviews and AI Mode and the search features of ChatGPT and Perplexity, all of which show websites in their answers. The term was introduced in a research paper in 2023. GEO does not replace search engine optimization (SEO): Google says the same SEO best practices apply to its AI features, with no additional requirements.

The glossary entry on generative engine optimization gives the short definition. This post explains where the term comes from, how it differs from SEO and what a website needs for AI search.

The term was introduced in the paper "GEO: Generative Engine Optimization" by Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan and Ameet Deshpande. It first appeared on arXiv on November 16, 2023, and was published in the proceedings of KDD '24, the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, held in Barcelona from August 25 to 29, 2024.

The authors call these systems generative engines and look at them from the side of the people who write websites. A generative engine combines several sources into one answer, and, in the paper's words, content creators have "little to no control" over when and how their content is shown. GEO is their response: a framework for optimizing website content for generative engines and for measuring visibility inside a generated answer, for example by how many words of the answer draw on a source and where its citation appears.

To test the idea, the authors built GEO-bench, a set of 10,000 queries from different domains, and a generative engine of their own: it fetches the top five Google results for a query and has the model GPT-3.5 Turbo write the answer. In that setup, three changes to a page worked best: adding citations from reliable sources, adding credible quotations and adding relevant statistics. Compared with unchanged pages, they raised visibility by 30 to 40 percent (relative) on the paper's position-adjusted word count. Keyword stuffing, a tactic the paper describes as traditionally used in SEO, brought little to no improvement. The authors also found that which methods work best varies from one subject area to another. These are results for one test setup, not a figure any website can expect.

SEO and GEO aim at different places on the screen. Classic search shows a list of websites, and the GEO paper notes that the average position in that list is a good measure of visibility there. A generative engine writes an answer and names websites as citations inside it, at different positions and at different lengths, so the paper measures visibility by how much of the answer draws on a source and how prominently it is cited.

For Google, the overlap is stated in its own documentation: "The best practices for SEO remain relevant for AI features in Google Search", and there are "no additional requirements to appear in AI Overviews or AI Mode". Google also says that you don't need new machine-readable files, AI text files or markup for these features, and no special schema.org structured data.

ChatGPT and Perplexity use crawlers of their own for search. What Coday does for classic search, from page titles to structured data and load time, is on the SEO service page.

Most of this list is ordinary SEO groundwork. Each point is named by the providers themselves:

• Crawlers are let in. Google lists crawling allowed in robots.txt, and by any CDN or hosting infrastructure, among the basics. OpenAI recommends allowing its crawler OAI-SearchBot so a site can appear in ChatGPT's search results, and Perplexity recommends the same for PerplexityBot. OpenAI's settings for search and training are independent: you can allow OAI-SearchBot and still disallow GPTBot to signal that your content should not be used to train its generative AI foundation models.
• Pages are indexed. Google only shows a page as a supporting link in AI Overviews or AI Mode if it is indexed and eligible to appear in Search with a snippet.
• The content is in the HTML. Google recommends making important content available in text form and calls server-side rendering or pre-rendering "still a great idea", partly because "not all bots can run JavaScript".
• Structured data matches the page. Google uses structured data to understand the content of a page. For its AI features it should match the visible text; no special Schema.org type is needed.
• Business details are current. Google names up-to-date Business Profile and Merchant Center information among the basics.
• Pages work well for people. Google lists a great page experience and helpful, reliable, people-first content, meaning content created primarily for people and not to manipulate search engine rankings. What load time means for visitors is explained in the post on Core Web Vitals.

An llms.txt file is optional. It is a proposal by Jeremy Howard, first published in September 2024, for a Markdown file that gives AI agents brief background information on a website and links to more detailed pages. Google says its AI features need no such file.

At Coday in Wetzlar, Germany, GEO is part of the website work. Owner Umutcan Emre Tezgel builds websites with Next.js and does the GEO work in person: on a new website as set out in the quote, on an existing one as a rework. That covers content in the delivered HTML, structured data that matches the page, a robots.txt that lets search crawlers in, business details that read the same on every page, and an llms.txt. codayweb.de has its own at codayweb.de/llms.txt.

Every project has a fixed price, agreed after a free first call, and you receive the source code at project completion. What is included and what is not is set out on the page for generative engine optimization services. Whether an AI assistant mentions your business is its own decision; nobody can promise that.

If you want to know what your website gives AI search today, book a free first call. You speak with the owner, look at what your website delivers and what is missing, and then receive a list of the changes needed and a quote with a binding fixed price.

Is GEO the same as SEO?

No, but both rest on the same groundwork. SEO aims at a place in the list of search results; GEO aims at being used and cited in an AI-generated answer. Google says that the best practices for SEO remain relevant for its AI features and that there are no additional requirements to appear in AI Overviews or AI Mode.

Who came up with the term generative engine optimization?

The term was introduced in the research paper "GEO: Generative Engine Optimization" by Pranjal Aggarwal and five co-authors. It first appeared on arXiv on November 16, 2023, and was published in the proceedings of the KDD 2024 conference, held in Barcelona in August 2024.

Do I need an llms.txt file to show up in AI search?

Not for Google. Google says you don't need new machine-readable files, AI text files or markup to appear in AI Overviews or AI Mode. llms.txt is a proposal, first published in September 2024, for a Markdown file that helps AI agents use a website.

Can I keep my content out of Google's AI Overviews?

Google points to the controls it uses for Search as a whole: nosnippet, data-nosnippet, max-snippet and noindex limit the information Google shows from your pages in Search, so they also affect the normal results. Google-Extended is a separate control for AI training and grounding in some of Google's other systems.

Can anyone guarantee that ChatGPT or Google will cite my website?

No. Google says that even a page that meets all requirements and best practices is not guaranteed to be crawled, indexed or served. OpenAI and Perplexity recommend letting their search crawlers in so a site can appear in their results. Which sources an AI answer cites is decided by the service.

All sources retrieved on October 21, 2026.

• Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan, Ameet Deshpande: "GEO: Generative Engine Optimization", KDD '24, arXiv:2311.09735 (version 3): arxiv.org/abs/2311.09735 (opens in a new tab)
• Google Search Central: AI features and your website (last updated December 10, 2025): developers.google.com/search/docs/appearance/ai-features (opens in a new tab)
• Google Search Central: Understand JavaScript SEO basics (last updated March 4, 2026): developers.google.com/search/docs/crawling-indexing/javascript/javascript-seo-basics (opens in a new tab)
• Google Search Central: Introduction to structured data markup in Google Search (last updated December 10, 2025): developers.google.com/search/docs/appearance/structured-data/intro-structured-data (opens in a new tab)
• Google Search Central: Creating helpful, reliable, people-first content (last updated October 5, 2026): developers.google.com/search/docs/fundamentals/creating-helpful-content (opens in a new tab)
• OpenAI: Overview of OpenAI Crawlers: developers.openai.com/api/docs/bots (opens in a new tab)
• Perplexity: Perplexity Crawlers: docs.perplexity.ai/docs/resources/perplexity-crawlers (opens in a new tab)
• Jeremy Howard: The /llms.txt file, v2 (published September 3, 2024, last modified August 10, 2026): llmstxt.org (opens in a new tab)