What is GEO (Generative Engine Optimization)?

Where does the term GEO come from?

The term Generative Engine Optimization was proposed in 2023 by researchers from Princeton University and IIT Delhi, in the paper “GEO: Generative Engine Optimization”, accepted at the KDD 2024 conference. The authors define “generative engines” as systems that answer by writing a text from several sources, instead of showing a list of links.

Their main finding: some changes to a page increase its visibility in generated answers. According to the study, citing sources, adding quotations and adding statistics produce a relative improvement of 30 to 40%, and up to 41% for the best methods on their benchmark. Keyword stuffing, inherited from old-school SEO, brings nothing and even slightly lowers visibility.

What is the difference between GEO, AEO and SEO?

• SEO aims to rank your pages in Google’s and other search engines’ results.
• AEO aims to make your pages the source cited in an AI assistant’s answer. See our guide What is AEO?
• GEO targets how generative engines talk about your brand: whether they mention it, how they describe it and whether they recommend it over your competitors.

The three follow on from each other: SEO makes your pages findable, AEO makes them citable, GEO makes your brand correctly understood.

Why can an AI get your brand wrong?

A generative engine assembles what it finds. If information about your business is scarce, contradictory or close to that of another business, it can get things wrong: wrong activity, wrong city, an old phone number, or confusion with a namesake.

We saw this with our own brand: asked in October 2026, Perplexity sometimes confused Houdour with a foreign business with a similar name. The fix was to publish a clear statement on our website and profiles. See Is Houdour related to Houdour Digital?

How do you improve what AI engines say about your brand?

1. Publish a factual identity page: official name, other spellings, activity, area, founder, contact details, and what you are not if a namesake exists.
2. Keep the same information everywhere: website, LinkedIn, Google Business Profile, directories. Links between these profiles (the sameAs property in structured data) help machines connect them.
3. Add structured data that reflects the visible text. Google recommends that structured data match the text shown on the page, and states that no special markup is needed for its AI features.
4. Back your pages with sourced facts, as the GEO study recommends: figures, quotations and links to sources.
5. Get mentioned by others: press, directories, comparisons, partners. A generative engine trusts information more when several sources confirm it.
6. Correct errors at the source rather than in the AI: an outdated directory or old profile can keep false information alive for months.

How do you measure GEO?

GEO is measured with a fixed panel of questions, asked regularly to the same assistants:

• questions about your brand (“what is [your business]?”): is the description accurate?
• questions about your category (“which agency should I choose in Casablanca?”): are you mentioned, and in what position?
• the same questions in each language, because answers in French, Arabic and English often differ.

Repeat each question several times: answers vary from one try to the next, and only an average over a set of questions is reliable.