In two years, Google's results page has changed in nature. Where users once found ten blue links, they now read an AI-written answer — the AI Overview — before even seeing the first website. Meanwhile, millions of queries no longer go through Google at all: they're asked directly to ChatGPT, Perplexity or Gemini. For any business that relies on organic search, the question is no longer “how do I rank first?” but “how do I become the source the AI cites?”.
What has actually changed
Three deep shifts explain the scale of the change. None is trivial: together, they redefine how a page wins — or loses — its visibility.
1. “Zero-click” search becomes the norm
When AI answers directly in the results page, the user no longer needs to click. Convenient for them, but it reshuffles traffic: a #1 position can now drive far fewer visits than before. The answer isn't to flee these queries, but to target the ones where the user needs to go further — comparisons, quotes, getting in touch, deep expertise.
2. Intent beats keywords
Language models understand meaning, not just words. Stuffing a page with “web agency Lyon” fifteen times is now pointless — and can backfire. What matters is fully answering an intent: what the person is actually trying to accomplish. A page that covers a topic in depth, with its nuances and edge cases, outperforms ten shallow ones.
3. Trust (E-E-A-T) becomes a hard filter
Google and generative engines look for reliable sources to cite. The E-E-A-T framework (Experience, Expertise, Authoritativeness, Trustworthiness) is no longer a bonus: it's a selection criterion. A named author, a real company with legal information and structured data, signed and dated content, credible citations — these signals decide whether the AI picks you up or ignores you.
GEO: the new layer of search
Alongside classic SEO, GEO — Generative Engine Optimization — is emerging: the craft of being cited by generative AIs. The levers overlap with SEO in part, but the logic differs: you're no longer optimising only for a ranking algorithm, but for a model that summarises and attributes information.
- Answer the question in the first paragraph — AIs extract direct answers, not introductions that circle the topic.
- Structure everything: explicit headings, lists, tables, FAQs. Well-segmented information is easier to cite.
- State verifiable facts: numbers, dates, clear definitions. Models favour what they can attribute safely.
- Strengthen your entity: Schema.org structured data, consistent presence (site, LinkedIn, directories), demonstrated expertise.
We've laid out the full method in a dedicated guide.
Read: the GEO guideThe technical foundation still decides
One truth AI hasn't changed: an engine can only cite what it can read quickly and well. Technical performance governs indexing, crawling and trust. A slow site, poorly structured or rendered only with client-side JavaScript, forfeits part of its visibility — no matter how good the content.
This is exactly where a modern foundation makes the difference: server or static rendering, clean semantic markup, controlled Core Web Vitals, native structured data. We detail the framework's impact on SEO in our Next.js article.
The best content in the world stays invisible if it lives on a site engines crawl poorly. Technique isn't the opposite of content: it's its precondition.
A 5-step method for 2026
- 01Map intents, not just keywords: group your topics by real user need.
- 02Build pillar content that covers a topic end to end, and link it to satellite articles (the cluster model).
- 03Optimise for extraction: direct answer up top, clear structure, FAQ, hard data.
- 04Consolidate the entity: E-E-A-T, structured data, trust signals, cross-platform consistency.
- 05Measure differently: track citations inside AIs and traffic quality, not just click volume.
Want to audit your visibility in the AI era and build a strategy that holds?
Talk to VD InnovatiumFrequently asked questions
Is SEO dead now that AI has arrived?
No. SEO is evolving, not disappearing. The fundamentals — useful content, a fast site, clear structure — still apply. What changes is that you must now also optimise for generative engines (GEO) that summarise and cite your content, on top of classic search engines.
What is an AI Overview?
It's the AI-generated answer Google shows at the top of the results page, above the classic links. It synthesises several sources and cites them. Appearing there requires clear, structured and trustworthy content.
How do I appear in ChatGPT or Perplexity answers?
By applying GEO principles: answer questions directly, structure content (headings, lists, FAQs), state verifiable facts, and strengthen your entity with structured data and trust signals. You also need to allow AI crawlers in your robots.txt file.
Do I have to redo all my existing content?
Rarely from scratch. The right approach is to audit what exists, identify high-potential pages, and rework them for intent and AI extraction. A few excellent pillar pieces beat a multitude of shallow pages.