Search Has Changed: Why Getting Cited by ChatGPT, Gemini and Perplexity Is Now Vital in 2026

ChatGPT reports 900M weekly users; Google's AI Overviews reach 2 billion a month. The 2026 numbers behind AI search — and what they mean for your visibility.

June 26, 2026 · 7 min

How search evolved: from Google's blue links to synthesized answers from ChatGPT, Gemini and Perplexity

In 2026, hundreds of millions of people no longer "search" the way they used to. They ask an AI a question and read the answer. ChatGPT reports 900 million users a week. Google's AI Overviews reach more than 2 billion people a month. If your brand doesn't show up in those answers, it becomes invisible to a growing share of your potential customers — no matter where you rank in the blue links.

Here's what the numbers say. And why showing up in AI search is no longer optional.

You no longer get a list of links — you get an answer. For twenty-five years, searching meant typing a query, scanning ten blue links, clicking, and comparing. The sorting was on you. Today, Google puts an AI Overview at the top of the page that sums up the answer. ChatGPT and Perplexity reply straight away, citing a handful of sources. And often, the user never has to click at all.

This isn't a minor tweak. Google has launched an AI Mode: a search where you talk to the AI, like a conversation. It has already passed 1 billion users a month (a figure Google announced at Google I/O 2026). And the way people search is genuinely changing: in AI Mode, the average query is three times longer than a classic search. People aren't typing keywords anymore. They're asking real questions.

How you get seen has changed. Yesterday, you had to rank high in the links. Today, you have to get cited in the answer. Those are two different things — and plenty of sites that rank well on Google are nowhere to be found in AI answers.

Billions. This isn't a niche for tech enthusiasts anymore. The big AI engines now have as many users as the largest platforms on the web.

AI engineReported usageSource
ChatGPT900M weekly users (Feb 2026)OpenAI
Google AI Overviews2B+ monthly users, 200+ countries (Jul 2025)Google
Google AI Mode1B+ monthly users (May 2026)Google
Gemini (app)750M monthly users (Feb 2026)Google
Perplexity~780M queries/month (2025, per Perplexity)Perplexity

And it's not just about user counts. It's how we behave that's shifting. According to Bain & Company, about 80% of search users rely on AI-written summaries at least 40% of the time. On the shopping side, Adobe found that 39% of US consumers had already used generative AI to shop online, and 53% planned to within the year.

Zero-Click Is Exploding: People Read the Answer Instead of Clicking

Most Google searches now send no click to any website at all. According to the SparkToro / Similarweb study, 68% of Google searches in the US (early 2026) end without a single click to another site. That's 7.5 points higher than in 2024.

The effect of AI answers is clear. When an AI Overview shows up, clicks on the top result drop by around 60%. And zero-click searches climb to 83% on those queries.

Your trusty #1 spot can now be buried under an AI answer that responds in the user's place. Ranking well is no longer enough to win attention. The answer itself has to cite you.

(One caveat: this zero-click figure covers searches made from a browser. It doesn't count the Google mobile app. So the real rate is probably even higher.)

What This Changes for Your Traffic — and Why It's Not All Bad

You lose clicks on one side, but new traffic arrives on the other — and it's worth more. Adobe Analytics measured a jump of more than 1,200% in traffic coming from generative-AI sources to US retail sites (February 2025 versus July 2024). And that traffic is higher quality: +12% pages viewed per visit and −23% fewer people bouncing straight off compared with other sources.

Let's be honest about the scale. AI traffic is still small — a few percent of total traffic. But it's growing exactly the way AI Overviews did two years ago: very fast, from almost nothing. Back in 2024, Gartner predicted a 25% drop in traditional search volume by 2026 as people shift to chatbots. That's a forecast, not a certainty — read it as a direction, not a number set in stone. And that direction is no longer up for debate.

Not in the AI Answer Means Invisible

If the AI doesn't cite you, you don't exist in the answer. This is the biggest break from classic SEO. A Google results page shows ten links: even in seventh place, you still have a shot at being seen. An AI answer cites only three to five sources. There's no "page 2." Either you're one of the few sources it keeps, or you're nowhere.

AI engines pick sources that are:

  • well organized — clear headings, paragraphs that answer one specific question;
  • easy to quote — a sentence that stands on its own, that you can lift off the page without losing its meaning;
  • trustworthy and sourced — numbers, sources named explicitly, an identified author. These are the E-E-A-T signals: they show a page's experience, expertise, authority and trustworthiness;
  • machine-readable — data laid out in a format bots understand (JSON-LD) and, above all, a page that's open to AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended). If you block them, or they can't read your page, you're out of the running before you even start.

SEO Is No Longer Enough: Enter GEO

Ranking well on Google is no longer enough to get cited by AI. A page can sit third on a search and be completely absent from ChatGPT's answer to the same question. They're two unrelated things. That's exactly what GEO (Generative Engine Optimization) is about: shaping content so AI engines cite it, not just so classic engines rank it.

SEO is still the foundation, and it's non-negotiable: if Google can't find your page, or it's badly organized, nothing works. But GEO adds criteria SEO never looks at: passages you can quote as-is, well-attributed sources, data laid out for machines, and a page open to AI crawlers. For the full breakdown of the 46 criteria that matter, read the complete guide to GEO in 2026.

How to Tell If You're Visible — and How to Get There

Start by measuring, then fix what AI engines look at on your pages. In practice, three steps:

  1. Test. Ask ChatGPT, Perplexity and Google (AI Mode) the key questions in your market. Note who gets cited. If your competitors show up and you don't, the diagnosis is made.
  2. Audit. The manual test shows the problem; the audit shows the cause. You need to look page by page at what counts: how the content is organized, the sources, the data laid out for machines, the access left open to crawlers.
  3. Fix. Rewrite the key passages into easy-to-quote answers, add sources and machine-readable data, open the door to AI crawlers. Then you measure again.

That's exactly what LightSpot does. Our agents analyze your pages against 46 SEO + GEO criteria. They explain, in plain language, why AI engines aren't citing you and what to fix first. Run a free audit to see where you stand.

The Key Takeaways

Search has moved from blue links to ready-made answers. And people have followed: billions of them ask ChatGPT, Gemini, Perplexity and AI Overviews their questions every month. The result: most searches no longer end in a click. Visibility no longer depends on your ranking, but on the answer itself. Ranking well is no longer enough — you have to be quotable by AI. It's new ground, still lightly occupied, where getting ahead today costs far less than catching up tomorrow.

FAQ

Is Google Search going to disappear?
No, but it's changing. Google has put AI at the heart of its engine. AI Overviews already reach more than 2 billion users a month, and AI Mode has passed a billion. People are searching as much as ever. But more and more searches end without any click to a website: you read the answer right there on the page. The goal is no longer just to rank well. You have to get cited in the answer.
Should you drop SEO for GEO?
No. SEO is still the foundation. Content that Google can find, that's well organized and trustworthy, serves both. But ranking well on Google is no longer enough to get cited by ChatGPT or Perplexity. These engines pick their sources on other criteria: sentences you can quote as-is, sources named explicitly, data laid out for machines, and access left open to AI crawlers. GEO complements SEO — it doesn't replace it.
How do I know if ChatGPT or Perplexity is citing my brand?
The simplest test: ask ChatGPT, Perplexity and Google (AI Mode) the questions your potential customers ask in your market. Look at which sources get cited. If your competitors show up and you don't, you have an AI visibility problem. A GEO audit goes further: it analyzes the exact elements these engines read on your pages.
Is AI traffic worth chasing while it's still small?
Yes, for two reasons. First, how fast it's rising: Adobe measured a jump of more than 1,200% in traffic from generative AI to retail sites in just a few months. Second, quality: these visitors view more pages and leave the site less quickly. It's still a small source of traffic. But it's growing fast and delivering good results. This is the moment when getting ahead costs the least.
Which AI engines should I target first?
The three that get used most: ChatGPT (900M weekly users), Google's tools (AI Overviews and AI Mode, several billion users) and Perplexity. The good news: the criteria that make you citable are almost the same everywhere. Working on one improves your visibility on the others.
Portrait de Nicolas Meridjen, Fondateur de LightSpot.ai — outil d'audit de visibilité IA (46 critères SEO + GEO)

Nicolas Meridjen

Fondateur de LightSpot.ai — outil d'audit de visibilité IA (46 critères SEO + GEO)

Je construis LightSpot.ai et j'analyse comment les moteurs de recherche IA (ChatGPT, Perplexity, Google AI Overviews) choisissent les sources qu'ils citent. J'écris sur le GEO et le SEO à partir de données d'audit réelles.

On the same topic