Rapid Growth of AI‑Powered Search: Evidence and Implications (Research Findings – July 2025)

Background

Generative AI technologies have been integrated into web search engines such as Google’s AI Overviews, Google AI Mode and standalone answer engines (e.g., ChatGPT, Perplexity, Claude). These systems use large language models (LLMs) to synthesize answers from multiple sources rather than simply listing links. Analysts expected this to slow overall search activity because users might get answers directly, but recent data show the opposite—AI‑search is accelerating growth in queries and changing user behaviour more quickly than anticipated.

Evidence that AI‑Powered Search Usage Is Surging

1. Growth in queries and impressions on Google

• More queries and tokens processed: During Alphabet’s Q2 2025 earnings call, CEO Sundar Pichai reported that AI Overviews, which started rolling out broadly in mid‑2024, have increased search activity rather than reduced it. Google noted that AI Overviews drive over 10 % more queries for searches where they appearsearchenginejournal.com. Search revenue grew 12 % year‑over‑year and Pichai said both overall and commercial query volumes are upsearchenginejournal.com. Google also disclosed that it processes more than 980 trillion tokens per month, double the number reported in May 2025searchenginejournal.com.
• More impressions but fewer clicks: One year after launching AI Overviews, BrightEdge analysed Google’s data and found that total search impressions increased by over 49 %globenewswire.com. While impressions skyrocketed, click‑throughs declined by nearly 30 %globenewswire.com because users are satisfied by AI summaries and therefore click less. BrightEdge also noted that AI Overviews appear in over 11 % of queries, up 22 % since their debutglobenewswire.com. This shows that generative answers are becoming a common part of search results.
• Longer queries and deeper content: BrightEdge’s data show long, complex queries within AI Overviews increased 49 % since May 2024 and ranking‑style comparison queries decreased 60 %globenewswire.com. A report by Digital Information World summarised that long‑form queries (8 + words) surface AI summaries seven times more often than a year earlier and that sources beyond the top 100 organic results are cited more frequentlydigitalinformationworld.com. These trends indicate that users are asking more complicated, conversational questions and the AI system digs deeper into the web to answer them.
• AI Mode and Deep Search accelerate usage: At Google I/O 2025, Google announced AI Mode—a new tab that offers an end‑to‑end AI search experience. AI Mode uses a query fan‑out technique to break a question into subtopics and issue multiple queries simultaneously, allowing users to explore more of the web and receive follow‑up questionsblog.google. Google also previewed a Deep Search feature capable of issuing hundreds of searches and producing a fully cited reportblog.google. These enhancements signal an ongoing commitment to deeper, multi‑modal search experiences and will likely further increase query volumes.

2. Adoption of AI answer engines and consumer behaviour

• Consumer reliance on AI summaries: Bain & Company’s February 2025 research found that 80 % of consumers rely on AI‑generated summaries at least 40 % of the timebain.com. On traditional search engines, about 60 % of searches end without a click, meaning users get their answers directly from AI summaries and do not proceed to another sitebain.com. Bain also reported that 68 % of large‑language‑model users rely on these platforms for research and summarization, while 48 % use them for news/weather and 42 % for shopping recommendationsbain.com.
• AI search becoming everyday behaviour: In July 2025 Yext released The Rise of AI Search Archetypes report based on a survey of 2,237 consumers across the U.S., U.K., France and Germany. It found that 43 % of respondents use AI search tools like ChatGPT or Gemini daily or moreyext.com. 62 % trust AI to guide their brand decisions, putting AI on par with traditional search for brand discoveryyext.com. Nevertheless, 57 % prefer traditional engines for personal, medical or financial queriesyext.com, suggesting a hybrid search ecosystem.
• Growth among younger users and high‑earners: A 2025 Attest survey reported that 37 % of under‑40s in the U.K. and 32 % in the U.S. use AI for at least half of their searchesaskattest.com. Consumers also perceive AI search results as trustworthy—40 % find AI search more trustworthy than regular search results and 41 % trust AI results more than paid search adsaskattest.com. Among AI users, 60 % expect their AI search usage to increase in the next six monthsaskattest.com.
• Cross‑platform adoption and usage: Menlo Ventures’ State of Consumer AI report estimated that 61 % of American adults used AI in the past six months, equating to roughly 1.7–1.8 billion people globally, and nearly one‑in‑five rely on AI dailymenlovc.com. Adoption spans generations, with millennials emerging as heavy users and even 45 % of baby boomers reporting use of AI in the past six monthsmenlovc.com.
• Accelerated search activity: Comscore, which tracks generative AI search separately, found that since it began tracking in late 2023, U.S. search activity has grown by 16 %comscore.com—a sign that AI‑driven tools are expanding overall search usage.

3. Penetration of AI search across the web ecosystem

• Share of queries triggering AI summaries: Semrush’s study of 10 million keywords found that 13.14 % of all queries in March 2025 triggered AI Overviews, up from 6.49 % in January, meaning the share more than doubled in two monthssemrush.com. Informational queries were most likely to trigger AI Overviews (88.1 %), while navigational queries doubled from 0.74 % to 1.43 %semrush.com.
• AI traffic’s share of total search: Datos, a market‑intelligence firm, told the Wall Street Journal that desktop browser traffic to AI search engines (e.g., ChatGPT or Perplexity) has more than doubled since June 2024 and now accounts for 5.6 % of U.S. search trafficwarc.com. Among early adopters of LLM‑powered search, traditional engines’ share of desktop traffic fell from roughly 76 % to 61 % in one yearwarc.com.
• Forecasts for future dominance: A Semrush analysis projected that AI search could surpass traditional search visitors by 2028semrush.com. It also found that visitors coming from AI search engines are 4.4 times more valuable (higher conversion rate) than visitors from traditional organic searchsemrush.com. ChatGPT’s weekly active users grew 8 × from October 2023 to April 2025, reaching over 800 million userssemrush.com. These trends suggest that AI search adoption is accelerating toward mainstream dominance.

4. Evidence of changing user behaviour and reduced web engagement

• Lower click‑through rates and session endings: Pew Research Center analysed the browsing habits of over 900 U.S. adults in March 2025. Users whose searches displayed an AI summary clicked any link only 8 % of the time, compared with 15 % when no AI summary was shownsearchenginejournal.com. Only 1 % of users clicked a citation link within an AI summarysearchenginejournal.com. The study found that users were more likely to end their browsing sessions when presented with AI summaries—26 % of pages with a summary versus 16 % withoutsearchenginejournal.com. Overall, two‑thirds of searches resulted in users either browsing elsewhere on Google or leaving the site entirely without visiting an external websitesearchenginejournal.com.
• Impact on publishers and SEO: BrightEdge’s data and independent analyses (e.g., Digital Information World) underscore that while AI Overviews boost impressions, they shift the value from clicks to impressions. Sources outside the top ten organic results are cited frequently—89 % of sources in AI Overviews came from beyond the top 100 organic resultsdigitalinformationworld.com. This redistribution of attention challenges traditional SEO strategies and prompts marketers to optimize content for generative engine optimization (GEO) rather than purely ranking for keywords.

Analysis: Why Is AI‑Powered Search Growing So Quickly?

1. Better user experience: AI summaries provide direct, conversational answers, saving users time. According to Google, younger users are particularly drawn to these featuressearchenginejournal.com. Features like AI Mode, Lens, and Circle to Search allow multimodal interactions, further expanding search usesearchenginejournal.com.
2. Expanding range of queries: Longer, more complex questions can now be answered effectively; query fan‑out techniques issue multiple sub‑queries to gather deeper informationblog.google. This encourages users to ask follow‑up questions, driving more total queries.
3. Trust and convenience: Surveys show that a large share of consumers trust AI search for brand discoveryyext.com and view its results as more trustworthy than organic or paid searchaskattest.com. Trust coupled with convenience leads to habitual use.
4. Integration into mainstream platforms: Google’s inclusion of AI Overviews across many search results and the rapid adoption of ChatGPT, Gemini, Perplexity and other answer engines expose billions of users to AI search capabilities. With generative tools integrated into mobile devices and browsers, switching costs are low.
5. Growing ecosystem and marketing focus: Enterprises such as BrightEdge, Semrush and Yext are launching tools for AI search optimisation, signalling an ecosystem shift. Marketers are beginning to measure “impressions” and “citations” in AI outputs rather than just clicksglobenewswire.com. Predictions that AI search visitors will surpass traditional search visitors by 2028semrush.com drive investment and adoption.

Implications

• Traditional SEO is no longer sufficient: AI summaries often cite pages beyond the top organic resultsdigitalinformationworld.com. Content must be structured for AI digestion—clear, semantically rich and answer‑focused—to be referenced by LLMs. Marketers need to optimize for AI crawlability and diversify formats (text, video, interactive)bain.com.
• Expect continued growth: With AI Overviews already appearing in ~11 % of Google queries and doubling share in monthssemrush.com, and AI search visitors projected to surpass traditional search visitors by 2028semrush.com, adoption is likely to accelerate. Search activity is also expanding; Comscore reports a 16 % growth in search since AI adoptioncomscore.com.
• Potential market disruption: Data suggests that AI search results may cannibalize some web traffic and compress the marketing funnel. Pew’s study shows that only 8 % of users click a link when an AI summary appearssearchenginejournal.com and two‑thirds of searches end without visiting any external sitesearchenginejournal.com. This shift may threaten ad‑supported business models and emphasises the need to measure influence rather than clicks.
• Hybrid search ecosystem: Despite rapid growth, many users still prefer traditional engines for sensitive topicsyext.com. This suggests that AI and traditional search will coexist. Data indicates that only about 5.6 % of desktop search traffic currently goes to AI answer engineswarc.com, far below Google’s 90 % market share. Thus, AI search is growing quickly but has not yet overtaken traditional search.

Conclusion

Recent research across multiple sources reveals that AI‑powered search is growing far more quickly than analysts anticipated. Google’s own data show that AI features like AI Overviews and AI Mode are driving double‑digit increases in queries and impressionssearchenginejournal.comglobenewswire.com. Independent studies by BrightEdge, Semrush, Bain & Company, Yext, Attest, Comscore and Pew Research Center demonstrate that consumers are rapidly adopting AI search tools, trusting their results and using them more frequently. The share of queries triggering AI summaries doubled within two monthssemrush.com and total U.S. search activity has grown by 16 % since generative AI search tracking begancomscore.com. At the same time, AI search is altering user behaviour—click‑through rates are falling sharplyglobenewswire.comsearchenginejournal.com, and many searches now end without visiting any external sitesearchenginejournal.com.

The evidence indicates that AI‑powered search is not cannibalising traditional search but rather expanding it, albeit in a way that reduces website visits. Businesses and marketers must adapt quickly—optimizing content for generative engines, measuring AI impressions, and preparing for a future where conversational, multi‑modal search is the norm. While AI search currently represents a small but rapidly growing slice of overall search trafficwarc.com, its trajectory and user adoption suggest it will play a central role in how information is discovered and consumed in the coming years.