AI’s Impact on Website Traffic, by the Numbers

The relationship between search engines and websites has changed drastically over the past few years. Historically, websites published content and Google sent visitors. That exchange powered the internet’s economy for two decades.

AI has rewritten the deal. Google’s AI Overviews now answer questions directly on the search page, and standalone tools like ChatGPT handle queries that would have sent someone to a website a few years ago. The result is a measurable shift in where traffic goes, how much of it arrives, and which sites benefit.

This page compiles the most recent and most reliable data.

Key Stats: AI’s Impact on Website Traffic

• Users click a traditional search result 8% of the time when a Google AI summary appears, versus 15% when one doesn’t appear.
• 68% of US Google searches ended without a click in 2026, up from about 60% in 2024.
• Global publisher Google search traffic dropped roughly one-third in 2025.
• Small publishers (sites with less than 10,000 daily pageviews) have averaged a 60% traffic drop over the past two years.
• AI referral traffic grew 527% in five months, but still accounts for roughly 1% or less of total traffic for most sites.
• AI traffic converts 31% better than non-AI traffic.
• 95% of ChatGPT users still also use Google.

The Click-Through Collapse

The most rigorous study on AI’s effect on search clicks came from the Pew Research Center in July 2025. Rather than relying on surveys or third-party traffic estimates, Pew tracked the actual browsing behavior of 900 US adults who agreed to install monitoring software. The dataset covers roughly 68,900 Google searches conducted during March 2025.

The findings were stark. When an AI-generated summary appeared on the search results page, users clicked on a traditional search result just 8% of the time. When no summary appeared, they clicked 15% of the time. Based on this sample, the presence of an AI Overview cuts the chance of a click almost in half.

Google does include links to certain sources inside AI Overviews, but those links tend to attract even fewer clicks. Users clicked on a source cited within the summary just 1% of the time. Google added those citation links partly in response to publisher complaints, but in practice, almost nobody uses them.

At the time of the study, about 18% of all Google searches produced an AI summary. That number has grown since. Longer queries are far more likely to trigger one (53% of searches with 10 or more words produced a summary, compared to 8% of one- or two-word searches), and questions starting with “who,” “what,” “when,” or “why” triggered summaries 60% of the time.

This aligns with a broader acceleration of zero-click search. SparkToro’s June 2026 analysis, using Similarweb clickstream data, found that 68% of US Google searches in early 2026 ended without a click to any external website. That’s up from about 60% in 2024 and roughly 50% when SparkToro first measured the phenomenon in 2019. To put it more concretely, for every 1,000 Google searches in the US, only about 276 clicks now reach the open web.

AI Overviews didn’t create zero-click search. Google has been answering queries directly on its results page since Knowledge Panels launched in 2012 and Featured Snippets in 2014. But AI has accelerated the trend significantly.

High-Profile Casualties

The aggregate statistics matter, but the impact becomes clearer when you look at specific companies that have publicly reported the damage.

Chegg

Chegg is the clearest example of a company whose business was directly undermined by AI search. The online learning platform reported in its Q4 2024 earnings filing that non-subscriber traffic had fallen 49% year-over-year by January 2025, after an 8% decline in Q2 2024. Total revenue dropped 14% for the full year, subscribers fell 21% to 3.6 million, and the company posted a net loss of $837 million.

Chegg filed an antitrust lawsuit against Google in February 2025, alleging that AI Overviews were displaying answers sourced from Chegg’s content while keeping users on Google’s own page. By October 2025, the company had cut 45% of its workforce, roughly 388 jobs, as part of a restructuring expected to save $110 million.

This is a particularly useful case study because Chegg’s content is exactly what AI Overviews handle well. Homework questions, study guides, and step-by-step explanations are the types of queries where an AI-generated summary can satisfy the user without a click. Companies whose content answers specific, self-contained questions are the most exposed.

Wikipedia

In October 2025, the Wikimedia Foundation reported that human pageviews across all Wikipedia language editions had declined about 8% year-over-year. The Foundation attributed the drop to generative AI and social video platforms.

Wikipedia’s content is used to train virtually all major large language models, and its information powers many of the AI-generated answers that now keep users from visiting the site. As the Foundation put it, search engines are “increasingly using generative AI to provide answers directly to searchers, often based on Wikipedia content.”

Business Insider

Business Insider laid off 21% of its staff in May 2025, with CEO Barbara Peng citing the need to “be structured to endure extreme traffic drops outside of our control.” About 70% of the company’s business had some degree of traffic sensitivity, making it especially vulnerable.

Similarweb data reported by AdExchanger showed Business Insider’s organic search traffic fell 55% between April 2022 and April 2025. The company shut down most of its e-commerce content, a category built on search traffic, and pivoted toward live events and subscription journalism.

The Publisher-Wide Decline

The casualties above aren’t outliers. The traffic decline is broad-based and well-documented across the publishing industry.

Digital Content Next (DCN), which represents roughly 40 premium publishers including the New York Times, Condé Nast, and Vox, surveyed 19 member companies about their Google search referral traffic during May and June 2025. The median year-over-year decline was 10% overall, with news brands down 7% and non-news brands down 14%.

The picture gets worse at a larger scale. Chartbeat data published in the Reuters Institute’s 2026 trends report tracked more than 2,500 publisher websites globally and found Google search traffic down roughly one-third in 2025.

The pain isn’t evenly distributed. According to Chartbeat data cited by Neil Patel, small publishers with fewer than 10,000 daily page views saw 60% declines in search referral traffic over two years. Medium publishers (up to 100,000 daily page views) saw 47% declines. Large publishers saw 22%. In general, smaller sites have been hit harder.

AI Referrals Are Growing, but Not Enough

AI platforms are starting to send measurable traffic to websites. The growth rate is striking, but the volume is low.

The Previsible AI Traffic Report, which tracked more than 400 sites across 12 industries, found that sessions from AI platforms (ChatGPT, Perplexity, Claude, Gemini, Copilot) grew 527% between January and May 2025. That sounds great until you see the raw numbers. Sessions went from about 17,000 to 107,000 across the entire panel of 400 sites. It’s fast growth from a tiny starting point.

That’s the consistent finding across every study on AI referral traffic. SE Ranking put AI referrals at 0.32% of total website traffic in 2026. Ahrefs measured 0.17%. The exact number varies by methodology, but it’s clear that AI sends some traffic, and it’s growing, although it’s a fraction of what Google sends and nowhere near enough to replace what’s being lost from traditional search.

The quality detail is more encouraging. Adobe Digital Insights tracked over one trillion visits to US retail sites during the 2025 holiday season and found that AI-referred traffic converted 31% better than non-AI traffic. Those visitors were 33% less likely to bounce and spent 45% longer on the sites they visited.

AI Doesn’t Always Cite the Sources You’d Expect

One of the less obvious shifts is in which sites AI actually pulls from when generating answers. If you’ve spent years building your site’s search rankings, this finding is important: ranking #1 on Google doesn’t guarantee you’ll be cited in the AI-generated answer on that same search page.

The data on this is evolving and, frankly, messy. Different studies use different methodologies and arrive at different numbers. But they all point in the same direction.

BrightEdge tracked the overlap between AI Overview citations and organic top-10 rankings over 16 months and found it grew from about 32% in May 2024 to 54% by September 2025. So at peak overlap, nearly half of all AI Overview citations came from pages that weren’t ranking in the top 10 for that query. A separate BrightEdge analysis in February 2026 put top-10 overlap at just 17%.

Ahrefs’ study of roughly 863,000 search results found that only 38% of AI Overview citations came from top-10 pages, with about 36% coming from pages outside the top 100 entirely.

The practical takeaway is that Google’s AI system assembles answers from a wider pool of sources than traditional search rankings. It breaks queries into sub-questions, pulls from pages that answer those specific fragments well, and frequently cites YouTube videos, forums, and government sites that don’t show up on page one of the traditional results. A site can rank first for a query and still be absent from the AI Overview on that same search.

The Other Side: Search Isn’t Dead

The data above paints a challenging picture, and none of it should be dismissed. But don’t ignore the counterweight data.

The most important context comes from Datos and SparkToro’s clickstream research, which found that 95% of ChatGPT users still also use Google, while only about 14% of Google users go to ChatGPT. AI search, at least for now, is mostly additive. People are using AI tools for some queries and Google for others, not abandoning one for the other.

Research from Graphite.io published in March 2026 adds useful perspective. The study found that total information discovery (combining traditional search engines and AI platforms) grew 26% worldwide from 2024 to 2025. Google’s absolute search volume hasn’t fallen. The pie got bigger, and AI captured most of the new growth. When you filter AI sessions to only search-like prompts (people actually asking questions, as opposed to coding, writing, or other tasks), AI’s true search-competing footprint drops to about 28% of global search volume and 17% in the US.

Google has pushed back on the narrative directly. In an August 2025 blog post, VP and Head of Search Liz Reid wrote that “total organic click volume from Google Search to websites has been relatively stable year-over-year” and that Google was “sending slightly more quality clicks to websites than a year ago.” She described third-party decline reports as “often based on flawed methodologies, isolated examples, or traffic changes that occurred prior to the roll out of AI features.”

Google hasn’t shared auditable data to support that claim, and every major independent study has found the opposite. But Google is the platform in question, and any honest accounting of this topic should note what they’ve said and that it conflicts with the independent data.

What Website Owners Should Actually Do

The data points to a few practical responses that apply across industries.

The first is measurement. If you’re still evaluating your online presence primarily by organic search traffic, your dashboard will look like a slow-motion disaster even if your business is healthy. Track branded search volume, AI citations, direct traffic, email subscribers, and conversion rates alongside referral traffic. A site that gets fewer visits but converts better from AI referrals and maintains strong direct traffic isn’t necessarily in trouble. A site that’s losing traffic with no compensating channel growth is.

Second, the data on which sources AI cites is instructive. The sites that tend to appear in AI-generated answers share some common traits. They present information in a clear, structured format that AI systems can parse easily. They include original research, specific data points, or first-hand expertise rather than repackaged information from other sources. They tend to have strong topical authority in their specific area rather than publishing broadly across unrelated subjects.

Third, the sites most vulnerable to AI traffic losses are those that answered simple, self-contained questions. Chegg, for example. If your content provides a factual answer that can be summarized in a paragraph, an AI system can satisfy that query without sending anyone to your site. Content that involves complex analysis, original data, interactive tools, personal expertise, or community discussion is harder for AI to replace.

Finally, the AI referral channel is small but growing fast, and the visitors it sends are genuinely valuable. Positioning your site to be cited by AI platforms (through original research, structured content, and clear expertise signals) is becoming a parallel strategy to traditional SEO. Not a replacement for SEO, but an additional channel that compounds over time.

FAQ