What is generative engine optimisation (GEO)?

Generative engine optimisation (GEO) is the work of making your content more likely to be used and cited when an AI assistant or AI search feature answers a question. Where SEO aims at a position in a list of links, GEO aims at being one of the sources an AI answer draws on and names.

Where the term comes from

The term was set out in a 2023 research paper, "GEO: Generative Engine Optimization", by Pranjal Aggarwal, Vishvak Murahari and colleagues, later presented at the ACM KDD conference in 2024. The authors describe generative engines as systems that retrieve sources with a search engine and then summarise them with a language model. Their concern was that website owners have little control over whether their content appears in those summaries. In their tests, changes to how content was written and presented raised its visibility in generative engine responses by up to 40%, and which changes worked depended on the subject of the question.

How GEO differs from SEO

The two overlap more than the new name suggests. AI answers are built from pages the system retrieves, and retrieval still depends on a page being crawlable, indexed and relevant. What changes is the unit of success:

• In SEO you measure where a page ranks and how often it is clicked.
• In GEO you measure whether an answer names your brand or cites your page, for the questions your buyers ask, across different assistants.

An answer can mention you without linking to you, link to you without recommending you, or recommend a competitor using your page as a source. So GEO tracking has to record more than a position: who was named, which sources were read, and how you were described.

What Google says about its own AI features

For AI Overviews and AI Mode in Google Search, Google's guidance for site owners is that no special optimisation is required and that ordinary SEO fundamentals still apply, such as making content easy to find through internal links and keeping structured data consistent with the visible text. Google also says there is no special schema.org markup needed to appear in these features. Clicks and impressions from them are counted in the Search Console Performance report under the Web search type. See AI features and your website on Google Search Central.

That guidance covers Google's features only. Assistants such as ChatGPT, Claude and Perplexity choose and cite sources in their own ways, and they do not report to your Search Console.

How to check where you stand

Start with the questions a buyer would ask before choosing a product like yours, written the way a person would type them. Put each question to the assistants your buyers use, and record for each answer whether you were named, which competitors were named, and which pages were cited. Repeat on a schedule, because answers change. Our guide How to check whether AI assistants recommend your business walks through doing this by hand.

What usually improves it

Most of what helps is ordinary good publishing, aimed at questions rather than keywords:

• Answer the question directly near the top of the page, then explain.
• Back claims with sources and specific figures that a reader, or a model, can check.
• Keep facts about your company consistent across your site, listings and review profiles, so every source an assistant reads says the same thing.
• Cover the comparisons and objections buyers raise, since those questions are where recommendations are made.

The research found that results vary by subject, so test changes on your own questions and measure them, rather than applying a checklist and assuming it worked.

Tracking it over time

Checking by hand works for a handful of questions. Once you track dozens across several assistants, it becomes a recurring job. The GX360 AI visibility tracker keeps a bank of questions for your category, asks them on a schedule, records who was named and which sources each answer read, and turns the questions where you are missing into proposed fixes.