Key points
- Generative engine optimisation (GEO) is the work of making a business, product or page more likely to be found, cited and described accurately in answers written by AI search tools.
- The term comes from a 2023 research paper by Pranjal Aggarwal and colleagues. Bing now uses it too, describing GEO as the practice of understanding how content participates in AI-driven experiences.
- Google says there are no extra technical requirements for its AI features: a page must be indexed and eligible to show a snippet. Good GEO rests on the same foundations as good SEO.
- For most UK businesses, the work is practical: measure what AI says, fix the website pages and profiles it reads, earn fair reviews, and measure again.
- Nobody can promise that an AI tool will name a business. Honest GEO reports counts across repeated runs, with dates and methods stated.
A short definition
Generative engine optimisation, usually shortened to GEO, is the work of making a business, product or page more likely to appear in answers written by AI tools such as Google AI Mode, Google AI Overviews, ChatGPT search and Microsoft Copilot, and more likely to be described accurately when it does.
A traditional search result is a list of links. An AI answer is a paragraph or two that names a few businesses, says something about each one and cites the pages it relied on. GEO is about those names, those descriptions and those citations.
Other names for the same work include AI search optimisation, answer engine optimisation (AEO) and LLM optimisation. We use GEO and AI search visibility because they're the terms most people search for.
Where the term comes from
The phrase was introduced in a research paper titled "GEO: Generative Engine Optimization", first posted to arXiv in November 2023 by Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan and Ameet Deshpande. The authors built a benchmark of queries and sources, tested ways of rewriting web content, and reported that some methods raised a source's visibility in generated answers by up to 40% in their tests. They also found that what worked varied by subject area.
That figure describes what worked on a research benchmark in 2023. The AI tools have changed a great deal since, and results for a particular business will differ.
The term has since moved from research into industry use. In February 2026 Microsoft's Bing search blog described GEO as "the practice of understanding how content participates in AI-driven experiences", and Bing launched reporting in Bing Webmaster Tools that shows how often a site is cited in Copilot and other AI answers.
How AI search tools build an answer
The details differ by platform, but the published documentation describes a common pattern.
- Google AI Overviews and AI Mode draw on Google's search index. Google says both may use a technique it calls query fan-out, issuing multiple related searches across subtopics and data sources to build a response. When AI Mode launched in the UK in July 2025, Google said early users were asking questions 2 to 3 times the length of traditional searches.
- ChatGPT search searches the web when a question would benefit from current information and shows citations to the pages it used. OpenAI's help pages say it sometimes works with third-party search providers. OpenAI's own crawler for search, OAI-SearchBot, collects pages so they can appear in ChatGPT search answers.
- Microsoft Copilot and Bing rely on what Microsoft calls grounding: connecting the AI model to current information from the web so its answers can cite sources.
In every case, the AI can only name and describe a business using what it can find and read. That's the practical basis of GEO.
How GEO relates to SEO
They overlap heavily. Google's guidance on AI features says there are "no additional technical requirements" beyond being indexed and eligible to appear with a snippet, and that you don't need new machine-readable files, AI text files or special markup. The usual SEO basics apply: crawlable pages, useful content in text, internal links, and structured data that matches what's visible on the page.
The differences are in what you measure and what you fix:
- What you measure. SEO tracks rankings and clicks. GEO tracks whether AI answers name you, how often across repeated runs, what they say, and which sources they cite.
- What the questions look like. People ask AI longer questions with more detail, such as a town, a budget, a worry or a deadline. Pages that answer those specifics give AI something to quote.
- Which sources matter. AI answers about local businesses lean heavily on the business's own website and Google Business Profile. Accurate profiles and listings matter as much as the website.
- What counts as an error. An AI answer can state a wrong price or the wrong opening hours as fact. Finding and correcting those at the source is part of GEO.
What GEO work involves for a business
For a typical UK business, the work falls into 4 stages. This is the method we use, and you can run a basic version yourself.
- Measure. Write the questions your customers ask, in their words. Ask each one several times in the AI tools they use, logged out, and record who's named, what's said and which sources are cited. Our guide to measuring AI search visibility sets out the method in full.
- Fix what AI reads. Clear pages for each service with prices as text, a team page, answers to common questions, and a Google Business Profile that matches your website. The 20-point checklist covers the usual gaps.
- Earn fair reviews and accurate listings. Ask every customer for a review in the same way, with no incentives, and correct wrong details on directories and other sites that answers cite.
- Measure again. Same questions, same method, and a report that states counts before and after, including when nothing changed.
What GEO can't do
No one controls what an AI tool says. Answers vary between runs, between users and between days, and the platforms change how they work without notice. Any provider who guarantees that an AI will name your business is promising something they can't deliver.
Some tactics sold as GEO also break platform rules or UK law. Hidden text written for AI crawlers, fake or incentivised reviews, and asking only happy customers for reviews all fall into that category. Since 6 April 2025, the Digital Markets, Competition and Consumers Act 2024 has banned fake reviews and concealed incentivised reviews in the UK.
A note on llms.txt
llms.txt is a proposal, published by Jeremy Howard in September 2024, for a Markdown file at the root of a website that summarises the site for AI tools. It's cheap to add, and it remains a proposal. Google's guidance says you don't need AI text files to appear in its AI features. We publish one on this site, and we don't present it to clients as something that changes results.
Where to start
Type the question your best customer would ask into Google AI Mode and ChatGPT, logged out, and note who's named. Do it again on a different day. If you're not named, or you're described wrongly, the guide to how AI chooses businesses explains what's likely to be missing.
If you'd rather we did it, start with a free AI snapshot, and the AI visibility audit gives larger firms the full picture. We work across every industry.
Sources
- Aggarwal et al., GEO: Generative Engine Optimization (arXiv:2311.09735)
- Google Search Central: AI features and your website
- Google: Introducing AI Mode in the UK (28 July 2025)
- Microsoft Bing blog: grounding on the AI web (12 February 2026)
- OpenAI: Overview of OpenAI crawlers
- OpenAI Help Center: ChatGPT search
- llms.txt proposal
All sources checked on 12 October 2026. Platform documentation changes often; if something here is out of date, tell us at hello@inplaincite.com.