The ongoing rollout of Google’s AI Overviews has prompted intense debate among publishers, SEOs and advertisers. Critics claim the feature is eroding referral traffic from search, while Google insists the opposite is true, arguing that its AI-driven changes are generating more searches and better-quality clicks. The latest exchange between independent researchers and Google’s own data team highlights the complexity of measuring the real impact.

Google’s Official Position

In a blog post from Liz Reid, VP and Head of Google Search, the company doubled down on its position that AI in Search is not causing widespread traffic collapse. According to Google, total organic click volume to websites from Google Search has been “relatively stable” year-on-year, with a slight increase in what it calls “quality clicks”, those where users do not quickly return to the search results.

The company attributes this to a shift in behaviour. AI Overviews, and now the newly launched AI Mode, allow users to ask longer, more complex questions. Google says these richer queries often result in more links being displayed on the results page, expanding opportunities for websites to be clicked.

Google concedes that for certain quick-answer queries, like “when is the next full moon”, users may be satisfied without clicking through. However, it argues that this has been the case with other answer-box features such as the Knowledge Graph and live sports scores. For the majority of queries, particularly those involving deeper research, product discovery or nuanced information, users still click through, and these clicks tend to be more engaged.

The Pushback From External Studies

Independent research paints a more pessimistic picture. A recent Pew Research study, based on search history data from 900 users, found that when an AI summary appeared in search results, users clicked a traditional link in just 8% of sessions. In cases without an AI summary, that figure nearly doubled to 15%.

While Google dismissed the Pew findings as based on too small a sample size and “flawed methodologies”, it is not the only dataset showing a downturn in organic click-throughs.

• SEMrush (May 2025) analysed tens of millions of US search queries and found a steady rise in zero-click searches since AI Overviews began wider testing in late 2024. For informational queries in categories such as health, news and “how-to”, click-through rates to external sites fell between 18% and 29%. Although discussion-based and review-oriented queries saw some uplift, echoing Google’s claim about forums gaining visibility, SEMrush’s sample showed a net decline for publishers overall.

• Ahrefs (June 2025) tracked 50,000 keywords over a 12-month period, reporting that SERPs with AI Overviews saw, on average, a 23% lower CTR for the top three organic results compared with similar queries without AI blocks. Sites in position one, traditionally the biggest winners, were hit hardest, losing over 40% of expected clicks in some cases. Ahrefs concluded that Google’s aggregate “stability” masks a redistribution effect where a small number of sites win big, but the majority lose.

• SparkToro and Datos (April 2025) analysed 60 million US searches and estimated that nearly 70% now end without a click to an external website, the highest rate on record. They identified AI Overviews and other on-SERP answer formats as the primary drivers, with news publishers, recipe blogs and reference sites most negatively impacted. Their view is that “net clicks from Google are no longer a growth channel for the average site”, urging marketers to focus on brand demand and platform diversification.

These findings, from three of the industry’s most trusted data sources, reinforce the idea that while some sites benefit from AI-driven search changes, many others are experiencing meaningful declines in referral volume.

Winners and Losers in the New Search Landscape

Google acknowledges that while overall traffic levels are steady, traffic patterns are shifting. Sites featuring forums, videos, podcasts and authentic first-person content are seeing gains. Reddit, for example, has become more visible in search results, with user-generated threads frequently linked within AI Overviews.

This reflects a broader trend in user preference for experiential and community-driven content. In-depth reviews, unique perspectives and original analysis are more likely to capture clicks from AI-informed searchers who want to go beyond the initial overview.

For marketers and publishers, the implication is clear. Traditional, surface-level SEO content may be less resilient in the age of AI Overviews. To remain competitive, sites will need to invest in richer, differentiated assets that complement, rather than compete with, Google’s summarised answers.

Strategic Implications for Digital Marketers

1. Optimise for Deeper Engagement
   With “quality clicks” becoming the priority, marketers should focus on content that invites exploration. Detailed comparisons, expert commentary, original research and multimedia formats are more likely to capture user interest post-overview.

2. Diversify Traffic Sources
   Sole reliance on Google Search is increasingly risky. Expanding presence on YouTube, TikTok, LinkedIn, Reddit and niche forums can hedge against fluctuations in search-driven traffic.

3. Leverage Structured Data and Rich Snippets
   AI Overviews pull from sources that are well-structured and easy to parse. Using schema markup, clear headings and authoritative citations can increase the chances of inclusion and visibility.

4. Track and Segment Search Traffic More Precisely
   Aggregate click metrics may mask changes in query types. Splitting performance reporting by informational, navigational and transactional queries can reveal which segments are growing or declining.

5. Experiment With AI-Friendly Content Types
   Podcasts, interactive tools, and community Q&A sections can appeal to the formats that AI Overviews seem to prioritise.

The Unanswered Questions

Google maintains that only 20% of desktop searches currently display AI Overviews, meaning the full impact is yet to be felt. If coverage expands significantly, the redistribution effect could intensify, amplifying both the wins and losses.

There is also the question of transparency. While Google shares directional trends, it has not provided detailed public data to validate its claims. Without independent verification, the debate over AI’s net impact on the open web will remain unresolved.

For now, the safest assumption for senior marketers is that the search environment is in flux. The “relatively stable” traffic figures cited by Google may hold true in aggregate, but that stability is underpinned by major shifts in where, how and why users click.

The Bigger Picture

Google insists it “cares passionately” about the web ecosystem and has built AI experiences to highlight, not replace, the web. By training models to cite sources prominently and link to relevant sites, the company argues it is preserving the value exchange between search and publishers.

Whether that is an accurate portrayal or a strategic reassurance, the reality is that AI-driven search is changing the way discovery works. The brands and publishers who adapt fastest, by aligning with evolving user behaviour and diversifying their visibility, will be best placed to thrive in what Google describes as “the most expansionary moment for the web”.

For digital leaders, the takeaway is not to fear AI Overviews, but to plan for them. The shift is already here, and waiting for consensus on the data will be too late for those unwilling to evolve.