AI Overviews click-through rate: what’s actually taking your clicks
Your rankings are stable. Your impressions are growing. Yet organic clicks keep falling. If that sounds familiar, you are watching the AI Overviews click through rate problem play out in your own Search Console account.
AI Overviews are Google’s AI-generated answers that sit above the traditional blue links. Since Google expanded them widely, marketers and publishers have reported the same pattern: pages hold their position but send fewer visitors. Because of this, a common conclusion is that SEO is dying.
That conclusion moves too fast. AI Overviews are a real, measurable cause of AI Overviews traffic loss. But they are one factor among several. The size of the effect depends heavily on the type of query, the industry, and the intent behind the search. A recipe query behaves nothing like a legal services query.
This article works through what the available data actually shows, using research from Ahrefs, Pew Research Center, Semrush, and SparkToro, rather than a single scary headline. It also covers why zero-click search 2026 figures vary so much between sources, which industries are affected most, and what a business can realistically do about it.
What are Google AI Overviews?
Google AI Overviews are short, AI-generated answers that appear above the usual list of links on some search results. Google rolled them out widely in the United States in May 2024, after testing the idea the year before as Search Generative Experience. For hundreds of millions of people, seeing one is now a normal part of the search experience, not a rare feature.
An AI Overview pulls information from several web pages, then writes a new summary in plain language, usually with a few linked sources underneath. It tends to appear on questions with a clear factual answer, definitions, comparisons, how-to queries, and less often where Google judges a plain list of links already serves the search well.
People often confuse this with the older featured snippet, which pulls one exact passage from one page. An AI Overview generates new text from several sources at once. That makes it far less obvious which single page deserves credit, which is why earning a citation now takes a different skill than winning a snippet did.
AI Overviews click-through rate: what the evidence actually shows
Several research groups have measured the AI Overviews click through rate effect, and their headline numbers do not match. That is confusing if you have only read one of them.
Ahrefs analyzed 300,000 keywords and found the top-ranking page loses 34.5% of its usual clicks when an AI Overview appears. An updated Ahrefs study in 2026 put that decline at 58%. Position-one CTR on AI Overview keywords fell from about 0.073% in December 2023 to 0.016% in December 2025.
| Period | Position-one CTR (AI Overview keywords) | Change vs baseline |
| December 2023 | 0.073% | baseline |
| December 2025 | 0.016% | 58% lower |
Pew Research Center took a different approach entirely, tracking the real browsing behavior of 900 U.S. adults across 68,879 Google searches. It found people clicked a traditional result in 8% of visits with an AI summary present, against 15% without one.
Semrush’s much larger study, over 10 million keywords, adds an important wrinkle. Looking only at keywords that already had an AI Overview, it also found higher zero-click rates. However, when it tracked the same keywords before and after an Overview appeared, the zero-click rate barely moved, dipping slightly from 33.75% to 31.53%. Many of those keywords were already informational, high-zero-click searches beforehand, so the Overview was not the sole cause.
This is why every AI Overview CTR study seems to disagree: each measures a slightly different thing, keyword-level search data versus real browsing sessions, and each defines a click a little differently. The type of search matters as much as the study you read:
| Search type | Typical impact on CTR | Notes |
| Informational | Highest | AI can answer directly |
| Commercial investigation | Moderate | Users still compare |
| Local | Lower | Local Pack remains important |
| Transactional | Often lower impact | Purchase intent drives clicks |
In practice, a page answering what is X absorbs far more of this than a page selling X.
Why AI Overviews create more zero-click searches
Users get answers without visiting websites
A zero-click search ends without the person visiting any website. The searcher reads an answer right on the results page, whether it comes from an AI Overview, a featured snippet, or a quick fact box, then moves on. This is not new behavior. Zero-click searches existed before AI, for simple things like the time in Tokyo. What changed is the range of questions Google can now answer this way, and zero-click search 2026 figures show how far that range has grown.
Google answers follow-up questions directly
An AI Overview can also answer a follow-up question without sending the searcher anywhere new, updating the same summary in place as the question narrows. A single search session can now resolve several related questions without one website visit. Each of those unclicked questions still counts as an impression, part of why impressions vs clicks numbers have grown so far apart.
Search sessions are becoming conversations
Search increasingly looks less like scanning ten links and more like a short exchange with the results page itself, a shift Google’s AI Mode pushes further with a chat-style interface. It would be a mistake to assume every search will look like this soon. Adoption varies by age, device, and query type, and plenty of searches still end in a click. But the direction is clear enough that AI search behavior deserves ongoing attention, not a one-time reaction.
The data behind AI Overviews and organic traffic
Search Console is the best free tool for seeing this in your own account. Reading it well, though, means separating three numbers that used to move together.
Impressions count every time a page showed up in a result, whether an AI Overview pushed it below the fold or not. CTR is the share of those impressions that became a real click. Average position shows roughly where a page sat, not what sat above it. Comparing impressions vs clicks side by side, rather than watching total traffic alone, is usually the fastest way to tell a ranking problem from a click behavior problem. Our full breakdown of this exact Search Console method covers the step-by-step version.
SparkToro, working with Similarweb clickstream data, found 68% of U.S. Google searches in early 2026 ended without any click at all, up from about 60% two years earlier. That figure covers every kind of zero-click outcome, not only AI Overviews. Google’s own Search Central documentation takes a calmer position than most third-party research. It states plainly that AI Overviews follow the same core SEO practices as classic search, with no extra technical requirement for citation beyond being properly indexed.
None of this research is perfect, and every AI Overview CTR study has a limitation worth naming. Panel studies like Pew’s involve under a thousand participants. Keyword studies like Ahrefs’ and Semrush’s rely on aggregated Search Console data that Google itself calls an estimate. What nobody disputes is the direction: clicks that used to land on a website are increasingly staying inside Google.
Which industries are being affected the most?
AI Overviews traffic loss is not spread evenly. Whether a business feels this shift depends heavily on how many of its searches are informational versus transactional.
Healthcare and finance sit at the high end, since both generate huge volumes of what is and how does questions that an AI Overview is built to answer directly. SaaS and legal sit in the middle, since buyers still compare options carefully before deciding. Ecommerce tends to see a smaller direct hit, since a shopper who wants a specific product still has to reach a page that can sell it to them. Local services land somewhere in between, since the Local Pack, not the AI Overview, usually still owns the top of the page for near me searches.
| Industry | Expected impact | Why |
| Healthcare | High | Informational queries |
| Finance | High | Definitions and explanations |
| SaaS | Medium | Comparison searches |
| Legal | Medium | Trust still matters |
| Ecommerce | Lower | Shopping intent |
| Local services | Mixed | Local Pack influence |
A lower-impact industry cannot ignore this either. Instead, the response looks different, which the next two sections cover.
Why rankings alone no longer tell the whole story
Rankings, traffic, and conversions used to move closely enough together that tracking rankings alone was a reasonable shortcut. That shortcut no longer holds.
A page can rank in the same spot for months while search visibility, meaning how often people actually register that listing, quietly changes underneath it. Traffic tells a related but separate story, since organic CTR decline concentrates in informational searches, so a business can lose sessions on a blog post while its commercial pages barely notice. Conversions matter more than either number alone. A shopper who sees an AI Overview, does not click, and searches your brand name directly a day later still becomes a customer, an assisted conversion that never credits the original page.
A full picture now needs brand search volume and engagement metrics alongside the traditional trio of rankings, traffic, and clicks, measured across the full customer journey rather than one click.
How businesses should respond instead of panicking
Create original content AI cannot replace
Content that keeps earning clicks despite AI Overviews tends to share one trait: it says something a summary cannot fully compress. First-hand experience is hard to fake. A real case study with actual numbers, a genuine customer story, or original research from your own data gives a reader a reason to click through. Proprietary data especially cannot be summarized, since a model has nothing to summarize until your page exists. Generic explainer content, the kind restating a definition covered elsewhere, is exactly what gets replaced. It needs a genuine angle or a specific number a summary cannot borrow.
Optimize for AI search, not only rankings
Ranking well is still necessary. It is no longer sufficient alone. Being cited inside an AI Overview or a chat answer depends on a related skill often called AI search optimization. Entity optimization means making it unmistakably clear who you are and how you connect to related topics. Structured data gives machines a direct way to read your content instead of guessing. Clear answers near the top of a page, rather than a long introduction first, are easier for an AI system to lift and cite. Topical authority across a subject, not one lone post, feeds directly into whether a model trusts your site enough to cite it. AI Search (GEO/AEO) for [Industry] is the specific discipline built around this, usually the fastest path for a business that has done solid traditional SEO but is not yet showing up inside AI-generated answers.
Improve click appeal
Some lost clicks are recoverable with better basics, even without a ranking change. A specific, concrete title earns more clicks than a vague one at the same position, and the same is true of a meta description that answers the actual question rather than restating the title. A recognizable brand earns clicks almost regardless of position. Rich results, star ratings, FAQ dropdowns, clear pricing shown directly in the result, give a listing more weight on a page with more competing for attention above it than before.
Common myths about AI Overviews
AI Overviews killed SEO. They changed it. Organic search still drives real traffic for most sites, and the skill set has expanded to include AI search optimization alongside the traditional list.
Rankings do not matter anymore. Rankings still decide who is even eligible for a citation, since a page has to rank and be indexed well enough to be considered at all.
Every industry is equally affected. In practice, it is not close, as the industry table above shows.
AI Overviews always reduce traffic. In fact, pages cited as a source inside an Overview can see a real increase in clicks compared with the same position without a citation.
Publishing more AI-generated content is the solution. It tends to make things worse, since generic AI-written content is exactly what an AI Overview is best at summarizing without sending a click your way. Our guide to what still earns organic clicks goes further into what does work instead.
Frequently asked questions
Do AI Overviews reduce click-through rate?
Yes, for most searches, though the exact size depends on which study you read. Ahrefs has measured position-one click-through rate falling 58% on keywords that trigger an AI Overview. Pew Research Center found real users clicked a traditional result in 8% of visits with an AI summary present, against 15% without one. Both point the same direction: the effect concentrates in informational searches and is smaller for transactional ones.
Can AI Overviews hurt SEO?
They change what success looks like more than they eliminate it. A page can still rank well and stay properly indexed while earning fewer clicks, since the loss happens at the click stage, after the ranking already did its job. The practical response is to add AI search optimization and clear, citable answers on top of existing SEO work, not to abandon organic search.
Are zero-click searches increasing?
Yes, though this trend predates AI Overviews by years. Featured snippets, knowledge panels, and simple instant answers were already producing zero-click searches long before generative summaries existed. AI Overviews accelerated the trend by extending that shortcut to a much wider range of complex questions.
How do I track AI Overview impact?
Start in Google Search Console. Pull impressions, CTR, and average position for your top pages over the same window, rather than watching one blended traffic number. A page with rising impressions, a stable position, and a falling CTR is showing the AI Overview pattern specifically, a different problem than a genuine drop in rankings.
Which websites lose the most traffic?
Sites built around broad, definition-style content tend to lose the most, since that content is exactly what an AI Overview is built to summarize on its own. Sites built around transactional or highly specific commercial content, where the searcher must reach the page to complete an action, tend to hold up noticeably better.
Can businesses recover lost clicks?
Some of it, yes, though not all. Improving titles, meta descriptions, and eligibility for rich results recovers clicks still winnable at your current position. Earning an actual citation inside the AI Overview, rather than only ranking below it, recovers a different and often larger share.
The bottom line
AI Overviews are genuinely changing how people interact with search, and the data backs that up from several independent directions. But a falling click number does not automatically mean falling visibility or falling business performance. It often means the same visibility is now expressed through fewer clicks and more brand searches than before.
The businesses handling this well are not panicking over one traffic chart. They separate impressions, CTR, and average position before drawing conclusions. They build original content a summary genuinely cannot replace, and treat AI search optimization as an addition to their SEO work, not a replacement for it.
Traditional SEO is not dead. It decides whether a page is even eligible to be considered for a citation in the first place. Ranking well is now the entry requirement, not the finish line.
Want a clear picture of where your own pages stand, and a specific plan for earning AI citations instead of losing clicks to them? AI Search (GEO/AEO) for [Industry] is built around exactly that question, for the industries where this shift is moving fastest.
Sources
The specific figures cited above, linked in order of first use.
Ahrefs, AI Overviews reduce clicks by 34.5%, April 2025.
Ahrefs, AI Overviews reduce clicks by 58%, updated February 2026.
Pew Research Center, Google users are less likely to click on links when an AI summary appears, July 2025.
Semrush, AI Overviews study, 10 million keywords, 2025 to 2026.
SparkToro, when Google stops sending clicks, what still works, June 2026.
Google Search Central, AI features and your website, documentation.