GEO: the new SEO for the ChatGPT era (and why it matters now)
7 min · November 3, 2026 · @solennstories
What actually changes when more and more people ask AI directly instead of searching on Google — and how to adapt your content to that reality.
What is GEO (Generative Engine Optimization) and how does it differ from SEO?
With classic SEO, the user sees a link in position 3 and clicks, or not. With GEO (Generative Engine Optimization), the AI directly cites a brand, a statistic or a definition in its answer — the user consumes the information without ever clicking, but associates the knowledge with the cited source.
Why are people turning to ChatGPT instead of Google?
A study by OpenAI's economic research team and Harvard economist David Deming, published by the NBER in September 2025, shows that "practical guidance", "seeking information" and "writing" together account for nearly 80% of ChatGPT conversations (NBER) — users now treat AI as a full-fledged search tool, not just a creative assistant. Gartner, for its part, predicted in February 2024 a 25% drop in traditional search volume by 2026 in favour of AI chatbots (Gartner) — a forecast that remains debated.
How does a brand get cited by ChatGPT and AI engines?
Which factors decide a citation: Reddit, listicles, query length
AI engines have measurable preferences: Reddit ranks first or second among cited domains on all five platforms studied, and about one citation in ten comes from a self-promotional "listicle" (Peec AI, analysis of 30 million cited sources, 31 March 2026). Query length also matters: according to the same study, the share of queries that trigger an AI Overview rises from 64.6% for a 2-word query to 89.1% for an 11-to-15-word query.
How to adapt your content to GEO: self-contained blocks, clear answers, sources
Structuring text into self-contained, directly citable blocks (a clear answer, a sourced figure, a table) is in line with what Google recommends for its AI features: content written for a human audience, well organised into paragraphs, sections and headings, and grounded in real expertise (Google Search Central). Structural clarity beats volume of text.
What does a brand risk when AI never cites it?
Content that's never cited as a source simply doesn't exist in the "cognitive map" the AI builds on a topic during its synthesis phase — whatever the content's real quality otherwise.
Comparison table — classic SEO vs. GEO
| Classic SEO | GEO | |
|---|---|---|
| Ultimate goal | Click on a link | Citation in a generated answer |
| Format that performs | Long, well-linked pages | Self-contained, sourced answer blocks |
| Sources favoured | Domain authority, links | Multi-source consensus (Reddit, forums, reviews) |
| Measure of success | Position, traffic | Citation rate, Share of Model |
What do research and field experience say about GEO?
Where GEO comes from: the Princeton study (up to +40% visibility)
The term GEO comes from a study by researchers (IIT Delhi, Princeton and other institutions), published in 2023 and then accepted at the KDD 2024 conference: it shows that suitable optimisation can increase a piece of content's visibility in generative engine answers by up to 40% (Aggarwal et al., arXiv). It's the most solid basis available to date, even though the field is still young.
Cited or ignored content: why optimise for long, precise queries
Beyond the self-contained block structure already mentioned, query length and precision play an underestimated role: according to Peec AI, the share of queries that trigger an AI Overview rises from 64.6% for a two-word query to 89.1% for an 11-to-15-word query. Concretely, this means content optimised for long, precise queries — the ones actually asked by someone looking for a detailed answer — has mechanically more chance of being picked up by generative engines than content built around generic keywords.
Is GEO a technical box to tick once? No: tracking is continuous
Many still think GEO is a technical box to tick once and for all. It isn't: market shares between ChatGPT, Perplexity and Gemini shift from one quarter to the next, and content cited today may no longer be in six months without any mistake having been made in between. Regular tracking isn't a comfort option, it's a condition for keeping the visibility you've earned.
Is GEO just a fad? Why start now
Classic SEO took fifteen years to settle into recognised practices. GEO is in its first two years of documented existence. The vocabulary, the tools and even the reference metrics are still evolving from one month to the next. That's not a reason to ignore it — it's the opposite: it's the moment when getting started costs the least, before the practice becomes standardised and competitors catch up with the lead that's still possible to take today.
Keep reading: structuring content, measuring AI visibility, tools
→ GEO: how to structure content so AI cites it — the concrete method, rule by rule.
→ SEO for beginners: how you actually get found on Google — the foundation GEO builds on.
→ GEO and AI visibility, my services — the overview page on the topic.
Frequently asked questions
- Is classic SEO still useful in the era of generative AI?
Yes: a good part of the signals feeding AI answers (authority, technical structure, quality content) rest on the same foundations as good classic SEO. Google also states that basic SEO best practices remain the foundation of its AI features (Google Search Central).
- How do I know if my industry is already affected by GEO?
An industry is affected by GEO as soon as its potential customers ask an AI assistant information-seeking or comparison questions before buying. That's already the case for 94% of B2B buyers in 2026 (Forrester).
- What is GEO (Generative Engine Optimization)?
GEO, or Generative Engine Optimization, is the optimisation of content to be cited in the answers of generative engines such as ChatGPT, Perplexity or Google's AI Overviews. Where SEO aims for a click on a link, GEO aims for the citation of a brand, a figure or a definition.
- What is the difference between SEO and GEO?
SEO seeks to place a page in the results to earn a click; GEO seeks to get a brand cited in a generated answer. SEO relies on domain authority and links, GEO on self-contained, sourced answer blocks, and is measured by citation rate and Share of Model.
- Is GEO proven to work?
A study by researchers (IIT Delhi, Princeton and other institutions), accepted at KDD 2024, shows that suitable optimisation can increase a piece of content's visibility in generative answers by up to 40% (Aggarwal et al., arXiv). The field is still young and methods evolve fast.
Got a project in mind ?
Influencer collaboration, UGC production or strategic support: write to me at hello@solennstories.com.
Get in touchrelated articles
GEO: how to structure content so ChatGPT and other AIs cite it
Answer first, self-contained blocks, questions as subheadings, sources and a named author: the concrete method I use to make content easy for an AI to cite.
7 min · November 6, 2026
SEOLink building audit: how to evaluate a website's backlinks
How I run a link-building audit step by step, which metrics actually matter and which ones don't — based on hotel-industry client work.
7 min · October 30, 2026
SEOSEO tools compared: which one fits your needs
Semrush, Ahrefs, or both? An honest comparison based on what I actually use, depending on client size and project type.
8 min · October 27, 2026
