For more than two decades, search behavior was easy to describe: people typed keywords into Google, scanned a list of links, and clicked through to websites. Keywords and lists of results still matter. But it is no longer the whole story.
AI has not replaced traditional search, but it is changing user behavior around search.
It is not simply moving users from Google to ChatGPT. Instead, AI is changing what happens inside search: how people ask questions, how answers are presented, how often users click, and how brands are discovered before a traditional search visit ever happens.
StatCounter’s April 2026 global search-engine market-share data still shows Google with 90.02% of worldwide search-engine share. Bing was far behind at 5.14%, followed by Yahoo, Yandex, DuckDuckGo, and Baidu. That means any claim that Google search has already been displaced by AI is overstated. Traditional search remains the dominant infrastructure for web discovery.
But market share does not tell the full story. Google itself is becoming more AI-driven. At Google I/O 2026, the company described what it called “a new era for AI Search,” including more AI features, agentic capabilities, and a more intelligent AI-powered search box. Google framed this as the biggest upgrade to the search box in more than 25 years. In other words, AI is not only competing with search from outside. It is being embedded directly into the dominant search engine.
If AI becomes part of Google Search, then “AI search” and “traditional search” are no longer cleanly separable categories. A user may still be on Google, but the experience may feel less like a list of links and more like a synthesized answer with citations, follow-up prompts, and task-oriented assistance.
Pew Research Center data provides one of the clearest looks at this behavioral shift. Pew analyzed browsing data from 900 U.S. adults and examined 68,879 unique Google searches from March 2025. About 18% of those Google searches generated an AI summary. The percentage was much higher for longer queries: just 8% of one- or two-word searches produced an AI summary, but 53% of searches with 10 words or more did. Searches phrased as questions were especially likely to produce AI summaries.
This suggests that AI is changing not only the answer format, but also the shape of search queries. Traditional keyword search rewarded short, clipped phrases: “best CRM software,” “R package validation,” “running shoes flat feet.” AI search encourages fuller questions: “What CRM software is best for a small technical marketing agency?” or “How should a pharma team validate R packages for regulatory submissions?” That is a different interaction model.
And there’s less clicks. When users saw a Google AI summary, they clicked a traditional search-result link in only 8% of visits. When no AI summary appeared, they clicked a traditional result almost twice as often, in 15% of visits. Users clicked links inside the AI summary itself in only 1% of visits. Pew also found that users were more likely to end their browsing session after a search page with an AI summary: 26% did so, compared with 16% on pages with only traditional results.
This is the core change for marketers and publishers. AI does not need to take massive market share away from Google to reduce website traffic. It can reduce clicks while keeping the user inside the search experience. The search still happens. The website visit may not.
Similarweb’s 2026 data points in the same direction. Similarweb reported that visits to AI platforms grew 28.6% between January 2025 and January 2026 in the US, across desktop and mobile combined. But referral traffic from AI platforms to external sites was flat over the same period. More people were using AI platforms, but those platforms were not sending proportionally more visitors out to the open web.
That supports a practical conclusion: AI’s impact is not best measured only by referral traffic. If a user asks an AI system for product recommendations, vendor comparisons, medical information, software guidance, or travel ideas, the influence may happen before any click. The brand that appears in the answer may shape the shortlist. The brand that does not appear may never be considered.
Similarweb’s Market Research Panel (US, January 2026) data also suggests that AI is moving upstream in the buying process. It shows that 35% of U.S. consumers used AI at the product-discovery stage, compared with 13.6% who used search. At the evaluation stage, Similarweb reported 32.9% AI usage versus 15% search. By the final “where to buy” stage, the gap narrowed sharply.
Summary
Traditional search still matters enormously. Google remains dominant by search-engine share, and conventional search is still central to how users find and visit websites.
But AI is becoming more influential in research, comparison, explanation, and early-stage discovery. The shift is not that AI has replaced search. The shift is that AI is changing how search works and how users behave after they ask a question.
AI is turning more searches into direct answers, reducing clicks on some informational queries, and moving brand discovery into conversational environments.
The next phase of search is not just about ranking for keywords. It is about being understandable, retrievable, credible, and accurately represented in AI-generated answers.