Google’s shift toward AI-generated answers, through AI Overviews and the newer AI Mode, has accelerated a long-running trend: people search, get a summarized response, and never click through to a publisher’s site. For SEO teams, that forces a rethink of what “winning” looks like when the search results page itself becomes the destination.
At the same time, the ecosystem is pushing back. In late January 2026, the UK Competition and Markets Authority (CMA) signaled it wants stronger publisher control, clearer attribution, and more transparency around how AI Overviews are produced and ranked. That regulatory pressure arrives as researchers and analytics providers quantify just how different “classic” search behavior is compared with generative answers, suggesting SEO is evolving into something closer to Answer Engine Optimization (AEO).
1) AI Mode and AI Overviews: the product shift that fuels no-click behavior
Google has been explicit that it is expanding AI Overviews and introducing AI Mode as a new way to handle longer, more complex queries. In March 2025, Google announced the expansion of AI Overviews and introduced AI Mode as an experiment, and later removed the AI Mode waitlist for US Labs users, noting that millions were already using it. Tech coverage framed AI Mode as an extension of AI Overviews designed for multi-part questions and follow-ups.
Capability-wise, AI Mode is not just text. Google has positioned it as multimodal, combining Lens with Gemini so users can “search what you see,” and it includes links to “learn more.” By August 2025, Google said AI Mode was expanding globally, 180+ new countries/territories in English, and adding more agentic features, signaling that AI-first answers are not a temporary experiment but a strategic interface change.
For publishers and brands, the consequence is straightforward: when the SERP answers the question, the need to click decreases. AI Overviews and AI Mode compress the discovery journey into an on-page synthesis. Even when links are present, they compete with the convenience of the summary and with users’ growing comfort in treating the overview as “good enough.”
2) The data behind no-click: clicks down, sessions end sooner
Multiple datasets point in the same direction: AI summaries correlate with fewer clicks. Pew findings summarized by Ars Technica reported that click rates drop from about 15% when there is no AI Overview to about 8% when an AI Overview appears, and only around 1% of users clicked a source link within the Overview. That is not a marginal change; it is a structural shift in how visibility converts into traffic.
Pew analysis also suggests that AI Overviews can end sessions earlier. Reporting summarized by The Decoder indicated users were more likely to end their session after seeing an AI Overview (reported as 26% vs 16%). In practice, that means fewer downstream pageviews, fewer opportunities to retarget, and fewer chances to build a relationship with the reader on owned properties.
Industry trends reinforce the pattern. Search Engine Roundtable, citing Similarweb, reported “zero-click” searches rising from 56% (May 2024) to 69% (May 2025) following the AI Overviews launch. Digiday, also citing Similarweb, gave publisher examples such as CBS News, where 75% of top keywords that triggered AI Overviews had no click-through (May 2025), alongside similar figures for People and Google News.
3) Visibility isn’t disappearing, impressions rise while traffic falls
A key nuance is that AI features can increase visibility while decreasing visits. BrightEdge reported that impressions were up over 49% since AI Overviews launched, while click-throughs were down nearly 30% since May 2024. That “more views, fewer visits” pattern makes traditional SEO reporting feel contradictory, teams see growth in presence, yet revenue and leads may soften.
BrightEdge also reported that AI Overviews appear in over 11% of Google queries and that coverage has grown since debut. It further described query-shape changes: longer, complex queries associated with AI Overviews were up 49% since May 2024, while “ranking-style” queries were down 60% and comparison queries down 14%. In other words, the demand mix is shifting toward questions AI is particularly good at summarizing.
Third-party studies have attempted to estimate the magnitude of CTR loss. The Guardian cited Authoritas suggesting clickthroughs can drop “up to 80%” when AI-generated summaries appear and push links down. Even if the exact impact varies by vertical and query type, the directional lesson is consistent: rankings alone are less predictive of outcomes than they used to be.
4) SEO vs AEO: generative answers behave differently than classic search
A January 23, 2026 arXiv paper by Chen et al. quantified divergence between classic Google Search results and generative AI answers, noting differences in source domains, intent matching, and freshness. That matters because many SEO playbooks implicitly assume that “ranking signals” and “result composition” operate the same way across experiences. The research suggests they do not.
This is where the term “Answer Engine Optimization (AEO)” gains practical meaning. If AI answers prefer different sources, or prioritize explanatory clarity, consensus, or recency differently, then the goal becomes not only ranking blue links, but becoming cite-worthy for synthesis. In some cases, that may mean investing more in primary-source assets (original data, tools, definitions, and strong editorial workflows) that AI systems can confidently summarize and reference.
It also reframes content strategy around intent types that trigger AI Overviews. BrightEdge has noted especially high AI Overview coverage for healthcare informational queries (with one cited figure at 83.6%), while eCommerce was lower (18.5%). If your market is heavily “informational,” AEO-style optimization, clarity, entity precision, and verifiable sourcing, becomes a core survival skill, not a niche tactic.
5) Measurement in an AI SERP: what you can and can’t track
Google’s messaging to site owners has been consistent on fundamentals: Search Central says SEO best practices remain relevant for AI Overviews and AI Mode, with no additional requirements and no special schema needed. Practically, that means teams should not abandon technical hygiene, crawlability, structured organization, and high-quality content just because the interface changed.
But measurement has to evolve. Google has stated that AI Overviews/AI Mode traffic is included in Search Console Performance under the “Web” search type, and it encourages measuring beyond clicks, such as conversions and time on site. That guidance implicitly admits the new reality: clicks are a weaker proxy for value when the SERP can satisfy intent without a visit.
There is also a reporting constraint: Search Engine Land reported (May 26, 2025) that Google confirmed AI Mode data will appear in Search Console, but it cannot be broken out separately, similar to AI Overviews. So, while you can watch trends in impressions, clicks, and conversions overall, isolating the precise impact of AI Mode vs classic results will require inference, controlled experiments, and triangulation with rank tracking, on-site behavior, and paid media signals.
6) Publisher pressure and regulation: the UK CMA targets the “no-click” impact
On January 28, 2026, The Guardian reported that the UK regulator is targeting the “no-click” impact, with the CMA proposing that publishers should be able to opt out of Google AI Overviews without disappearing from Search. That is a pivotal point: today, publishers can face a painful trade-off, accept being summarized (and potentially losing traffic) or opt out and risk losing visibility.
The Financial Times covered the CMA’s proposed conduct rules the same day, emphasizing more transparency, attribution, and publisher control over AI Overviews, with a consultation ending on February 25, 2026. The message is that market power plus AI summarization creates a bargaining imbalance, and that attribution, choice, and clarity may need to be mandated rather than requested.
Associated Press reporting echoed the CMA’s view that AI Overviews reduce traffic, and highlighted recommendations including opt-out mechanisms, better citation/attribution, and ranking transparency. Whether or not these proposals become enforceable rules, they signal that “no-click” is no longer just an SEO concern; it is a competition-policy and media-sustainability issue.
7) Quality, trust, and the “confident authority” risk in AI answers
No-click answers amplify the cost of errors because users may never reach the original sources that contain nuance, caveats, or updates. A January 24, 2026 investigation reported by The Guardian highlighted harmful inaccuracies in Google AI Overviews for health queries and described the “confident authority” risk, where the tone of a summary can sound definitive even when it is wrong.
That concern scales with reach. The same reporting noted that AI Overviews reportedly reach around 2 billion users monthly across 200+ countries. When that many people are exposed to a synthesized answer, the accuracy of citations, the faithfulness of summarization, and the visibility of source links become public-interest issues, not merely UX design choices.
For SEO and content teams, this changes the optimization target: it is not only “be discoverable,” but “be safely summarizable.” Content that is well-sourced, clearly structured, and explicit about uncertainties is more likely to be interpreted correctly by both users and models. In sensitive verticals like health and finance, editorial rigor becomes an optimization lever because it reduces the chance your brand is associated with a flawed or misleading synthesis.
8) Practical adaptation: from ranking to brand demand and multi-channel resilience
SEO teams are already seeing the need to diversify. Seer Interactive summarized observed organic CTR declines when AI Overviews were present and also reported paid CTR declines, recommending testing new channels and increasing brand presence in LLMs and AI Overviews. The takeaway is not “SEO is dead,” but “SEO alone can’t carry the full acquisition plan in an AI-first SERP.”
At the macro level, Digiday cited Similarweb showing US organic search referrals falling, for example, from roughly 2.3B visits in July 2024 to roughly 1.8B in June 2025, arguing that AI may drive more traffic in some cases, but not enough to offset the rise in zero-click behavior. In that environment, resilience comes from building demand that begins outside the SERP: email, direct traffic, social, partnerships, apps, and community.
Within SEO itself, the tactical focus shifts toward outcomes that still matter in a no-click world: winning citations, owning branded queries, being the entity AI systems reference, and capturing the users who do click with fast paths to conversion. That means prioritizing pages that answer complex intent (where users still want depth), strengthening internal linking to move visitors quickly to next steps, and treating “impression share + brand recall” as a legitimate KPI alongside sessions.
SEO is adapting because the interface has changed: AI Mode and AI Overviews are training users to accept a synthesized answer as the endpoint. The evidence, from Pew’s click-rate drops to Similarweb’s zero-click growth and BrightEdge’s “impressions up, clicks down” pattern, shows that traffic is no longer the default reward for visibility.
The next phase is a mix of strategy, measurement discipline, and governance. Google says classic best practices still apply, but reporting limitations mean teams must get smarter about inference and outcome-based metrics. Meanwhile, the CMA’s January 2026 proposals, publisher opt-outs without disappearing from Search, improved attribution, and transparency, suggest the market may impose new rules on how no-click answers are generated and credited. Until then, the most durable SEO playbook is one that earns trust, wins citations, and builds demand beyond a single click.
