GEO in 2026: why SEO is changing everything (and what to do)
ChatGPT, Perplexity and AI Overviews capture 40% of searches. GEO is taking over from traditional SEO. Here's what to do right now to stay visible.

Traditional SEO is mutating
Google SEO is no longer enough to be visible in 2026. For 25 years, the logic was simple: you produce content, you optimise it for Google, you show up in the 10 blue links. That logic is shifting fundamentally, and 2026 is the tipping point.
Three numbers that prove it:
40%
of Google searches now trigger an AI Overview box at the top of the page
Wellows, 2026
400M
weekly active users on ChatGPT Search
OpenAI, Feb. 2026
x3
traffic growth in 12 months for Perplexity, dominant in technical B2B
Similarweb, 2026
Sources: Wellows, 2026 · OpenAI, February 2026.
The result: on many queries, users no longer click the blue links. They read the AI-generated answer, which pulls from multiple sources without necessarily citing yours. This is the "zero-click search", and it is already rewriting the rules.
What this means for your traffic: if your site isn't cited in AI answers, you lose visitors even if you rank #1 on Google. A commercial query satisfied by an AI Overview = a lost click on your site.
SEO vs GEO: the practical difference
SEO optimises so that a search engine lists your link. GEO optimises so that a generative engine cites your content in its answer. These are two different goals that require two complementary strategies.
SEO classique
Google, Bing
GEO
ChatGPT, Perplexity, Claude, Gemini, AI Overviews
SEO classique
Liens bleus + meta description
GEO
Réponse synthétisée + citations sources
SEO classique
Mots-clés, backlinks, performance
GEO
Données structurées, citations, autorité d'auteur, factualité
SEO classique
Position SERP, clics
GEO
Taux de citation par les LLM, trafic referral
SEO classique
6-12 mois pour ranker
GEO
3-6 mois sur les niches techniques
Ancien monde
SEO classique
Nouveau standard
GEO
Good news: GEO doesn't replace SEO, it extends it. A site that ranks well on Google is more likely to be cited by AI. But GEO adds specific levers that traditional SEO doesn't cover. This is exactly what we put into practice with our SEO & GEO service.
The 7 pillars of GEO in 2026
Here are the 7 levers to activate, in order. Each one is detailed below.
Structured and extractable content
Schema.org structured data
Author authority (E-E-A-T)
Citations and external sources
Freshness and updates
Explicit AI bot permissions
llms.txt file
1. Structured and extractable content
LLMs extract content in self-contained chunks of 130-167 words under each heading. If your paragraphs contain pronouns that refer back to earlier sections, they become unusable.
Three practical rules:
- Answer at the start of the paragraph. If the question is "How much does X cost?", begin with "X costs between Y and Z." That is what the AI cites.
- Frame your subheadings as questions ("How long does it take to...", "Why choose..."). LLMs match these structures to conversational intents.
- Add explicit FAQs at the bottom of the article. Q&A format = the format AI engines prefer.
2. Schema.org structured data
JSON-LD markup has become a major ranking factor in 2026, no longer optional. LLMs directly use schemas to identify content type and authorship.
Implement these first:
Articlewithauthor,datePublished,dateModifiedFAQPagefor each Q&A sectionPersonfor the author (withsameAspointing to LinkedIn, X)Organizationfor your companyBreadcrumbListfor navigation
3. Author authority (E-E-A-T)
Google and LLMs converge on the same criterion: content signed by an identifiable human scores higher than anonymous content. E-E-A-T 2026 (Experience, Expertise, Authoritativeness, Trustworthiness) is shifting from "information retrieval" to "entity verification". Google and AI engines want to link every piece of content to a traceable real person.
Four concrete actions:
- Sign every article with a real person's name (never "The X Team")
- Create a dedicated author page with bio, photo, and LinkedIn links. See the Odyssee author page as an example
- Add a complete
Personschema withsameAs(LinkedIn is mandatory) - Link the author to your organisation via
worksFor
"The Person schema + sameAs pointing to the author's social profiles is the combination that maximises AI Overview citations. We observe +36% citation rate with a complete schema vs a partial one." Source: study by Searchscale AI, 2026
4. Citations and external sources
LLMs give +40% more credit to content that cites primary sources. When you make a factual claim (figure, study, statistic), link to the original source, not a third-party rephrasing.
Winning format: according to [a 2025 Carnegie Mellon study](url). Descriptive anchor + name + date + institution. LLMs favour this format to assess reliability.
Snowball effect: sites that are cited repeatedly become reference authorities that LLMs cite in turn.
5. Freshness and updates
AI engines heavily weight recent content. An article published 3 years ago carries less weight than one updated this week, even if the latter is shorter. Perplexity has a particularly strong recency bias: fresh content gets cited within 1-2 weeks vs 4-8 weeks for traditional SEO (source: GenOptima, 2026).
Best practice:
- Update your strategic articles every 7-14 days
- Set
dateModifiedin the frontmatter and theArticleJSON-LD - Visibly display "Updated on " at the top of the article
- Never fake freshness: if you bump
dateModifiedwithout a real edit, Google detects it and penalises you
6. Explicit AI bot permissions
AI crawlers look for their specific user-agent in robots.txt. If you only allow * without naming the bots, some refuse to crawl by default since the NYT vs OpenAI case.
Add to your robots.txt:
User-Agent: GPTBot
Allow: /
User-Agent: OAI-SearchBot
Allow: /
User-Agent: ClaudeBot
Allow: /
User-Agent: PerplexityBot
Allow: /
User-Agent: Google-Extended
Allow: /
Also add Applebot-Extended, MistralAI-User, Meta-ExternalAgent, cohere-ai, Bytespider depending on your coverage ambitions.
7. llms.txt at the root
Emerging standard (2025-2026) adopted by Anthropic, Mistral, OpenAI Search. A plain text file at the root of your site that guides LLMs to your canonical content.
Minimal format:
# My Company
> Short description of what you do.
## Main Pages
- [Home](https://example.com/)
- [Services](https://example.com/services)
- [Blog](https://example.com/blog)
## Authors
- [First Last - Role](https://example.com/authors/slug)
Place it at https://your-domain.com/llms.txt. It is read by all modern AI crawlers. See llmstxt.org for the official spec.
How to know if you appear in ChatGPT, Perplexity, etc.
GEO measurement today relies on 3 complementary channels. No single tool gives you the full picture; you need to cross-reference:
- Direct manual testing: type your target queries into ChatGPT (with web search enabled), Perplexity and Gemini. Is your site cited in the sources? Repeat the test every 2 weeks and keep a Notion tracker.
- Bing Webmaster Tools, AI Performance tab: Bing tracks how often your site is cited in Copilot and ChatGPT Search (which use the Bing index).
- Google Search Console: since 2025, GSC shows when your site appears in AI Overviews (filter "Search appearance > AI Overviews").
Practical tip: create a Notion page with 10-15 key questions your prospects ask. Submit them monthly to the 4 major AI engines (ChatGPT, Claude, Gemini, Perplexity) and calculate your Mention Rate (% of answers that cite your brand) + Citation Rate (% that include a clickable link). This is the most reliable measure of GEO positioning in 2026.
The GEO roadmap: where to start
If you're starting today, follow this order of attack for results within 6 months. The goal is not to do everything at once. It's to chain short steps that reinforce each other.
- 1
Technical site audit
1-2 daysrobots.txtwith AI bots,llms.txt, schema markupOrganization+Articleon key pages. - 2
Author page + JSON-LD Person
1 dayIdentify a real human as the signatory of all your editorial content. See our approach on the Odyssee author page. - 3
Restructure your existing content
1-2 weeksAdd TL;DRs, FAQs, question-format subheadings, self-contained chunks. - 4
Create 5 to 10 pillar pieces
1-2 monthsLong (1,500+ words), structured, with primary source citations, on your core topics. - 5
Measure and iterate
ongoingDashboard tracking AI citations, strategic content updates every 90 days.
How long before you see results? Expect 3 to 6 months to start appearing in AI answers for niche queries, 9 to 12 months for competitive queries. This is faster than traditional SEO on under-covered technical topics.
What no longer works in 2026
Five SEO tactics that paid off yesterday but have zero impact today. Continuing to invest in them is wasting time that should go towards GEO.
Ce qui ne sert plus
- Keyword stuffing: penalised by Google + ignored by LLMs since 2023.
- 500-word articles written in 30 minutes: too short to be extractable by AI.
- Buying backlinks: penalty risk + zero effect on AI citations.
- Raw AI content without human editing: -40 to -60% traffic on mass-AI sites during 2025 core updates.
- Zero-value "song lyrics" pages: AI Overviews actively demote them.
À faire à la place
- Content depth: 1,500+ words, cover a topic fully instead of 3 shallow articles.
- Author authority: real human byline, detailed bio, Person schema with sameAs LinkedIn.
- Verifiable external citations: dated primary sources (studies, institutions, official reports).
- Extractable structure: self-contained chunks of 130-167 words under each heading, explicit FAQ.
- Real freshness: substantive updates every 90 days on strategic content.
See the official Google rule on core updates ranking that broke the raw-AI strategy in 2025.
The real paradigm shift
GEO isn't just a new SEO technique. It's a shift in editorial posture. Before, you wrote to please an algorithm. Now, you write so that an AI recommends you to a human. These are two different things.
Who does an AI recommend? The sources it deems reliable, factual, up-to-date, and signed by an identified human. If you hit that standard, you're in the top 10% of sites in 2026. If you miss it, your competitor takes it, and your prospects end up with them.
To go further, also see:
- Our SEO & GEO service: the audit + 7-day action plan
- Our showcase websites expertise: where SEO/GEO is built in from the design stage
- Our maintenance pricing: which includes monthly SEO/GEO monitoring
Questions fréquentes
Is traditional SEO dead in 2026?
No, but it is evolving. Google remains the dominant search engine and traditional SEO is still relevant. GEO doesn't replace SEO, it extends it: you now need to optimise for generative engines on top of traditional search engines. Good practices reinforce each other.
What is the difference between SEO and GEO?
SEO (Search Engine Optimization) optimises so that a search engine lists your link in its results. GEO (Generative Engine Optimization) optimises so that a generative engine (ChatGPT, Perplexity, Gemini, AI Overviews) cites your content in its answer. SEO targets clicks, GEO targets citations.
How do you appear in ChatGPT answers?
ChatGPT Search uses the Bing index. Five steps: (1) register your site on Bing Webmaster Tools, (2) explicitly allow the OAI-SearchBot bot in robots.txt, (3) structure your content with clear subheadings, FAQs, and Article JSON-LD, (4) create an llms.txt file at the root of your site, (5) publish quality content signed by an identified author with a Person schema.
How long before you see GEO results?
On under-covered technical niche queries: 3 to 6 months to start being cited by AI. On competitive queries: 9 to 12 months minimum. GEO is generally faster than traditional SEO because technical niches lack quality content signed by human experts.
Should you abandon backlinks for GEO?
No, but their role is changing. Editorial backlinks (cited spontaneously in articles) remain an authority signal for both Google and LLMs. However, bought or low-quality backlinks have zero effect on GEO and risk a Google penalty. Focus on creating content so good it gets cited naturally.
What is the llms.txt file?
It is a plain text file placed at the root of your site (e.g. https://example.com/llms.txt) that guides AI crawlers to your canonical content. Simple Markdown format with a site description, main pages, and authors. An emerging standard since 2025, adopted by Anthropic, OpenAI Search, Mistral and Perplexity. See the official spec at llmstxt.org.
How do you measure your presence in AI engines?
Three complementary channels: (1) monthly manual tests on ChatGPT, Claude, Perplexity, Gemini with a list of 10-15 key queries; (2) Bing Webmaster Tools "AI Performance" tab tracking Copilot/ChatGPT Search citations; (3) Google Search Console filter "Search appearance > AI Overviews". Calculate your Mention Rate (% of answers citing your brand) and Citation Rate (% including a clickable link).
Shall we discuss your project?
30 minutes to scope your needs and tell you exactly what we would do in your position.
Start a project→



