How GEO differs from SEO
| SEO | GEO | |
|---|---|---|
| Goal | Ranking high in search results | Being cited and described accurately in AI answers |
| Unit | The page | The passage: answer paragraphs that make sense on their own |
| Measure | Rankings, clicks | Citation rate, how the brand is described in answers |
| Foundation | Technical SEO, content, links | Good SEO plus clear definitions, sources and structured data |
GEO does not replace SEO. Google says no special markup is needed for its AI features and that standard SEO practice still applies. Tools such as ChatGPT, Perplexity and Claude use their own crawlers and different search indexes, so they benefit from extra structure and files.
How does AI search choose its sources?
- Google AI Overviews and Gemini: use Google’s search index and ranking systems.
- ChatGPT search: finds sources through its own search crawler (OAI-SearchBot) and a Bing-based index.
- Microsoft Copilot: relies on the Bing index; a page that isn’t in Bing won’t appear in Copilot either.
- Claude: uses Brave Search results when web search is switched on.
- Perplexity: cites sources in every answer and favours clear, structured, up-to-date pages.
A GEO checklist
- Crawlability: robots.txt should not block AI search crawlers, and content should be in the HTML without needing JavaScript.
- A clear definition up front: each page should answer its question in a 40-60 word paragraph that stands on its own.
- Question-style headings: headings phrased the way people ask (“How is … priced?”) with a short answer right underneath.
- Dates and ownership: show the publication and update dates and who prepared the content.
- Sources and data: back claims with sources and give figures with their dates.
- Structured data: schema.org markup such as Organization, Article and FAQPage.
- llms.txt: a file for AI systems listing a short summary of the site and its key pages.
- Bing and IndexNow: add the site to Bing Webmaster Tools and report new and updated pages through IndexNow.
- A consistent description: describe the brand in the same words on its site, on LinkedIn, in directories and in the press.
What does the research say?
In the study that defined generative engine optimisation (Aggarwal et al., 2023), adding credible sources, statistics and quotations to content raised visibility in AI answers by roughly 30-40%, while keyword stuffing lowered it. These results come from a specific experimental set-up and won’t promise the same figure in every sector, but the direction is clear: AI prefers content that is sourced and clearly written.
How do you measure GEO?
Pick the 15-20 questions that matter most for your brand and ask each one several times in ChatGPT, Perplexity and Gemini; AI answers change from one run to the next, so don’t rely on a single try. Record each month how often your brand is mentioned, which of your pages is cited and how you are described.
What to avoid
- Writing separate content for AI rather than for people.
- Keyword stuffing.
- Fake reviews, made-up awards or unsourced claims.
- Leaving important information only inside images or sections loaded by JavaScript.
Frequently asked questions
Are GEO and SEO the same thing?
No. SEO aims for rankings in search results; GEO aims to be cited and described accurately in AI answers. A sound SEO foundation is a precondition for GEO.
Do you have to pay for ads to appear in ChatGPT?
No. The sources in ChatGPT search are chosen from web content. Whether your page is crawlable, clearly written and trustworthy is what decides it.
Is llms.txt required?
No, and Google says it isn’t needed for its AI features. It is a low-cost aid that helps other AI tools understand a site quickly.
Sources
- Aggarwal et al. (2023), GEO: Generative Engine Optimization
- Google Search Central: AI features and your website
- OpenAI: Overview of OpenAI crawlers
- The llms.txt proposal
This guide was prepared by the dipro.uk team and last updated on 16 November 2026. Questions: [email protected].
