The AI search shift: what's actually changing for brands, and what it costs

Across roughly 35,000 websites, AI tools sent 0.1% of total referral traffic in 2025, just below email at 0.2% Ahrefs · 2025B. The same visitors convert at about three times the rate of other channels, and up to eleven times in some publisher segments Microsoft Clarity · 2026B. AI search is small in volume and large in influence. The honest macro view for brand teams in 2026 is that AI is a quality channel sitting next to a volume channel. It is growing fast and reshaping decisions before any click is recorded.

Generative search and web search return different things for the same query

Generative search and traditional web search return different things even for the same query. An empirical study from MPI-SWS found three structural differences from Google web search Kirsten et al. · Ruhr University Bochum / Max Planck Institute for Software Systems · 2025A. Generative engines pull from a broader source pool. They mix retrieved pages with internal model knowledge. They surface different concept sets for the same query. The set of pages that can earn visibility widens. Classical ranked-list evaluation assumptions break.

The decision shape changes with the surface. Where keyword search produced ten browser tabs, AI search produces one synthesised answer with a short list of cited sources. The NBER working paper on ChatGPT usage finds around 700 million weekly active users sending roughly 18 billion messages per week by mid-2025 Chatterji et al. · NBER / OpenAI / Harvard University · 2025A. About 30% of those conversations are work-related, 70% personal. That single interface carries a structurally different shape of interaction from Google's ten-blue-link experience.

The shift is global. A controlled study of 24,000 queries across 243 countries finds two effects from exposure to AI summaries Aral et al. · MIT IDE · 2026A. Attention concentrates on a smaller set of authoritative domains. Users form more confident but less accurate beliefs on contested topics. The mechanism is consolidation of judgement into the answer itself.

The volume is small; the growth rate is what to plan around

The honest scale number is small. Across roughly 35,000 websites, AI tools sent 0.1% of referral traffic in 2025 Ahrefs · 2025B. Google sent 345 times more traffic than ChatGPT, Perplexity, and Gemini combined. The Chartbeat dataset reaches the same conclusion. AI sources account for less than 1% of publisher pageviews Nieman Lab (Chartbeat) · 2026B.

The growth rate is the part to plan around. Across 81,947 websites, average AI traffic grew about 9.7× year on year Ahrefs · 2025B. The average site's search traffic dropped about 21% over the same period. Mobile AI assistant visits reached 54.3 million unique visitors in December 2025, up 107% year on year Comscore · 2026A. ChatGPT mobile grew 84%. Microsoft Copilot more than tripled. Perplexity rose 265%.

The platform mix has stabilised around one dominant player. ChatGPT holds 76.85% of worldwide AI chatbot market share as of April 2026 Statcounter · 2026A. Google Gemini sits at 9%, Perplexity at 7.73%, Microsoft Copilot at 3.76%, Claude at 2.66%. Visibility in ChatGPT is the lever with the most reach today; a multi-platform presence is the only durable model.

The aggregate user base sits above two billion. DataReportal estimates 2.42 billion people use standalone generative AI platforms each month, with ChatGPT alone at around 1.15 billion monthly active users in April 2026 Kemp · DataReportal / We Are Social / Meltwater · 2026A. The volume is small as a share of total web traffic and large as a share of category-shaping intent.

AI is a quality channel: the lift is in per-visit value

Adobe Digital Insights documents the per-visit quality premium directly. US retail visitors from generative AI sources show 8% higher time on site, 12% more pages per visit, and a 23% lower bounce rate than visitors from other channels Adobe Digital Insights · 2025A. The same Adobe dataset measured visits from generative AI sources growing roughly 1,200% between July 2024 and February 2025 Adobe Digital Insights · 2025A.

Microsoft Clarity's analysis sharpens the conversion picture. AI-assistant visitors convert at roughly three times the rate of visitors from other channels, and up to eleven times in some publisher segments Microsoft Clarity · 2026B. The dataset is large but the publisher is a vendor with product interest. Treat the figure as Tier B with attribution: directionally correct, magnitude to be confirmed by independent replication.

The Bain Consumer Lab data covers the buying-intent end of the same picture. 30 to 45% of US consumers say they already use generative AI for product research and comparison Bain & Company · 2025A. AI accounts for up to a quarter of referral traffic at some retailers, though still less than 1% of total traffic. Consumers trust retailer-owned agents about three times more than third-party agents.

A counterweight finding belongs in the same section as the main claim. An Ahrefs quality study found AI visitors view four pages on average, 1.2 fewer than search visitors and 1.5 fewer than the typical visitor Ahrefs · 2025B. Sessions are longer in elapsed time but shallower in depth. AI visitors bounce 4.1% more often than search visitors. The reconciliation is straightforward. Adobe and Microsoft Clarity measure the high-intent commercial subset. Ahrefs measures the average across all AI-referred traffic. Both findings are true; they describe different slices.

Europe is on a longer runway with a steeper slope

AI search adoption in Europe is behind the US and on a defined trajectory. BCG's European retail dataset finds AI search visits grew from 4% of organic visits in early 2024 to 8% in early 2025. The projection reaches 25% by the end of 2026 and overtakes organic in 2028 Boston Consulting Group · 2025A. LLM referral traffic to leading European retailers is up more than 2,000% in fashion, nearly 1,200% in luxury, and almost 7,500% in specialty retail.

The implication for DACH and UK brands is precise. Absolute volume is lower than in the US today. Per-visitor value is the same. The runway is longer; the slope is steeper than US growth rates suggest. The longer runway exists to be used as preparation time.

Clicks are being absorbed, not redistributed

When a Google search result page includes an AI summary, users click on a traditional link in roughly 8% of visits. Without an AI summary, the figure is roughly 15%. Citation clicks inside the AI summary itself run at about 1% of visits Pew Research Center · 2025A. The Pew finding is the cleanest single number on the click side of AI search.

The Bain figure on the consumer side lines up directly. Roughly 80% of consumers now rely on zero-click results, AI summaries, or assistant answers for at least 40% of their search needs Bain & Company · 2025A. AI search use has reduced average organic click-through rates by 15% to 25%. The compression is consistent across measurement frames.

The publisher impact is independently measured. A four-axis academic audit of Google AI Overviews found activation on 13.7% of queries overall and 64.7% of question-form queries arXiv · 2026A. AI Overviews cite domains more credible on average than co-displayed first-page results, but nearly 30% of cited domains do not appear in those results at all. A diff-in-diff study using non-AI-Overview Wikipedias as controls found pageviews to AI-summary-triggering Wikipedia articles fell measurably more than control articles after AI Overviews rolled out arXiv · 2026A. Estimated traffic loss attributable to AI summaries runs in the single-digit to low double-digit percentages on affected article sets.

The answer absorbs the click. The decision happens inside the AI surface, and the brand presence required to compete sits at the citation layer rather than the destination page.

What this means for brand teams

McKinsey projects that AI-powered search will mediate roughly $750 billion in US consumer revenue by 2028 McKinsey & Company · 2025A. The forecast carries the standard consultancy caveat: the magnitude is a modelled projection rather than a measured value. The direction is firmer than the number. BCG's Consumer AI Disruption Index finds 67% of senior marketing leaders expect a high level of AI-driven disruption to their vertical's consumer journey Boston Consulting Group · 2026A. Travel, retail, and news are most exposed; financial services, fintech, and media are most protected.

Three implications follow directly from the evidence above.

First, raw traffic share is the wrong metric for the next 18 months. AI sends roughly one part per thousand of total referrals. Reporting against that number reads as if nothing is happening even where pipeline impact is real.

Second, per-visitor value and presence rate are the right metrics. Per-visitor value comes from the conversion and engagement data above. Microsoft Clarity reports roughly 3× the rate of other channels; Bain reports up to a quarter of referral traffic at some retailers. Presence rate comes from the citation layer: how often a brand appears, by name, in AI answers to commercial-intent prompts in its category.

Third, visible budget allocation may take 12 to 18 months to show pipeline impact. The patience curve is real, and it is consistent with the underlying growth curve rather than a reason to defer. Comscore mobile data shows 107% year-on-year growth in unique visitors to AI assistants in December 2025 Comscore · 2026A. Deloitte forecasts 72% of adults in developed markets will have generated a search overview by mid-2026, versus 61% who will have used a standalone AI tool Deloitte · 2025A. The slope is too steep to wait out.