GEO, AI SEO and AEO: what they are and what we measure
GEO (generative engine optimization) is the work of getting a website read and quoted by ChatGPT, Perplexity, Gemini or Google AI Overviews. This page explains the terms, what can be measured, and what nobody can promise.
Updated 2026-09-15
What is GEO?
GEO (generative engine optimization) is preparing a website and its content so that generative AI engines (ChatGPT, Perplexity, Gemini, Claude, Google AI Overviews) can read it, understand it and quote it in an answer. The market calls the same work “AI SEO”, “AEO” (answer engine optimization) and “LLMO”. The differences between the terms are marketing; the method is the same everywhere.
In practice GEO has two layers. The first is site readiness: whether the page can be fetched, whether it carries structured data, whether the text is written so it can be quoted as an answer. The second is the outcome: whether an AI engine, asked in a buyer's words, mentions your business and in what context.
How does GEO differ from SEO?
SEO optimises a page for a list of search results: position, click, visit. GEO optimises a page for an answer: the AI engine reads several sources and composes one text in which your business is either mentioned or not. There is no ranked list, so the idea of “first place” loses its meaning.
The foundation is shared: indexability, clear structure, facts that can be verified. The emphasis differs. GEO cares more about answer-shaped text (a question in the heading, the answer in the first paragraph), structured data carrying real facts, and access for AI crawlers, which many sites block without knowing it.
What can be measured in GEO?
Readiness is measured with seven checks and a 0–100 score: structured data, meta tags, llms.txt, feeds and sitemap, answerability of the content, access for AI agents, and an informational AI transparency block. Weights and the formula are published in the methodology. The score is deterministic: the same page on the same day gives the same number.
Mentions are measured by putting 20 buyer questions to four engines through their official APIs and recording whether the business is mentioned, with a quote. We repeat monthly and show the change together with a noise band, so random variation is not read as an event. Our own curve is public, including the zeros.
What can GEO not promise?
Nobody who measures honestly promises rankings or mentions in AI answers: answers differ even between two identical queries on the same day, and the engines' source selection is not public. So we show observed facts and a work list, and give no guarantees.
We also do not work around the AI platforms' terms of use. Where there is no API (Google AI Overviews), measurement is manual, without automated collection, and the report says so next to every such number.
Where do Lithuanian websites stand?
Corpus run of 2026-09-13: 180 Lithuanian business homepages measured, median 41 out of 100. 20 sites score at least 60, 24 score below 30. The full list is public in the index, and every number can be repeated with the same scanner.
Where to start?
The first step is to see your own number: the free scanner shows the score, the seven checks and your place among the country's sites in a few seconds, with no sign-up.
The second step is the audit (€590): what AI answers about your business to 20 questions, with quotes and a work list. The third is monthly monitoring (€189/month), which shows whether the work changes the outcome.
Common questions about GEO
Are GEO, AEO and AI SEO the same thing?
- Essentially yes: all three describe the work of getting AI engines to read and quote a website. GEO stresses generative engines, AEO answer engines, and “AI SEO” is the generic market term. The method and the measurement are the same for all three.
Do I need an llms.txt file?
- The file does no harm and takes half an hour, but studies show AI systems do not currently read it. It changes the score by 10 points out of 100; mentions, so far, not at all. Sources with quotes are published in the methodology.
How much does a GEO audit cost?
- The free scan shows the score and the checks. The audit costs €590 and covers 20 questions to four engines with quotes and a work list. Monthly monitoring with the tracker costs €189 per month and can be switched on later.
How quickly do AI answers change?
- Each engine refreshes its sources at its own pace, and answers vary even between two identical queries. So we judge change in monthly steps and only above the noise band. Our own curve shows what that looks like in practice, including months at zero.