AI Overviews Cost 61% of Clicks: Make the Loss Visible

What AI Overviews do to your organic clicks

AI Overviews answer the query directly at the top of the results page, synthesized from several sources. For many informational questions, that removes the reason to click any result at all. Agency Seer Interactive analyzed 3,119 informational queries across 42 organizations with roughly 25.1 million organic impressions between June 2024 and September 2025 (Seer Interactive, 09/2025). The result: when an AI Overview appears, the organic click-through rate falls by 61%, and the paid rate by as much as 68% (Seer Interactive, 09/2025). In absolute terms that is a drop from 1.76% to 0.61% - roughly one in three to four former clicks simply does not happen.

An independent analysis by the Pew Research Center confirms the pattern from the user side: people who saw an AI summary clicked through to a website only 8% of the time, versus 15% without a summary (Pew Research, 07/2025). The source link inside the AI Overview itself was clicked just 1% of the time (Pew Research, 07/2025). And sessions ended outright more often: 26% of searches with an AI summary ended the browsing session, compared with 16% without (Pew Research, 07/2025).

The numbers: how hard AI Overviews push down CTR

The effect is far from a niche phenomenon. AI Overviews now appear on more than 20% of all Google searches (Datos/SparkToro, 2026), and 83% of searches with an AI Overview end without a single click on a result (Datos/SparkToro, 2026). Overall, in the first months of 2026 around 68% of all US searches ended without a click, up from 60% in 2024 (Datos/SparkToro, 2026) - a level that consultancy Bain also describes with roughly 60% click-less searches for the classic case (Bain, 02/2025).

There is an important counter-movement: pages cited within the AI Overview see a 35% higher organic click-through rate and a 91% higher paid one than non-cited competitors (Seer Interactive, 09/2025). So the question is not only how much traffic you lose, but whether your content appears in the AI answer at all. Both can only be judged if you look at visibility and real visits together - the piece on AI search and GEO referral traffic goes deeper on the visibility side.

Why classic ranking checks hide the problem

Conventional SEO reporting asks: where do we rank? That question falls short now that the AI answer is wedged between ranking and visit. A page can sit stably at position one for months and still lose half its traffic, because the answer is already given at the top. Teams that only monitor positions see green while revenue from search shrinks. Four signals separate a zero-click loss from a genuine ranking collapse:

Making the zero-click effect visible in cookieless analytics

The loss only becomes visible when you place two data sources side by side: the search position per page and the actual, cookielessly measured page visit. When organic entries to a page collapse while its position holds, that is the fingerprint of an AI Overview. Cookieless analytics measure that entry without a consent banner and without device storage, so no consent refusal skews the data - an advantage that decides whether the numbers are reliable for up to every second visitor.

• Organic entries per landing page over time - if the value falls while the position stays put, an AI answer is the likely cause.
• Ratio of impressions to visits per URL - a widening gap reveals intercepted clicks.
• Share of ended sessions after the search entry, as a proxy for zero-click pressure (Pew Research, 07/2025).
• Referral from AI assistants as the counter-movement, so you do not confuse the new channel with direct traffic.
• Page types compared - guide and definition pages typically lose more than transactional pages.

The decisive step is attribution: it is not total traffic that counts, but which individual page loses how much. Only then can you steer deliberately instead of writing off search across the board. When the drop is detected automatically and alerted, you react in days rather than months.

Which pages lose the most

Not every page is affected equally. AI Overviews appear disproportionately for informational questions - 'what is ...', 'how does ... work', 'symptoms of ...'. Precisely the guide and knowledge content that many companies built as a reach engine now supplies the AI answer and loses the click. Seer's analysis explicitly focused on informational queries (Seer Interactive, 09/2025), and the Pew data likewise show the sharpest click loss for explanatory searches (Pew Research, 07/2025). Transactional pages - product, price, contact - are so far less affected, because here users seek an action rather than just an answer.

In practice that means: sort your pages by intent. Pure knowledge pages, FAQ answers and definition articles compete directly with the AI answer and need a deliberate decision - either optimize consistently for citability or tie the topic more strongly to an action that an AI summary cannot replace. Comparison, configuration and purchase pages keep their value because they lead to a step the user has to take themselves. This per-page classification is the basis of any sensible countermeasure - without it you optimize in the dark.

Telling zero-click apart from a genuine ranking loss

When a page's traffic falls, two causes are possible: either the position has dropped - through a search-engine update or stronger competition - or the position holds but the AI answer intercepts the click. The difference is decisive, because the two cases demand entirely different measures. The dividing line is the search position: if it stays stable while clicks fall, everything points to a zero-click loss. If it slips at the same time, it is a classic ranking problem. Check this not in aggregate but query by query - often the drop affects only the informational terms, while transactional searches keep delivering clicks.

• Position held, clicks fallen - the typical sign of an intercepting AI Overview.
• Position fallen, clicks fallen - more likely a ranking or competitive problem than a zero-click effect.
• Only individual, informational queries affected - matches the observed pattern of AI Overviews (Seer Interactive, 09/2025).

A baseline helps: record the organic entry per page from the period before AI Overviews appeared for the relevant terms, and compare today's value against it. A year-over-year comparison rules out seasonal swings, so a real break is not confused with the usual summer lull. Anyone who keeps that baseline running continuously spots not only the one-off drop but also the slow erosion over many weeks - the true hallmark of the zero-click trend.

From loss to counter-strategy

You cannot switch off the zero-click effect, but you can shape your content so it gets cited in the AI answer and wins the remaining click. If you also want to control which AI crawlers may read your content at all, the piece on llms.txt and AI crawlers covers the basics. Since cited pages achieve a 35% higher click-through rate (Seer Interactive, 09/2025), working on citability pays off twice.

1. Measure the loss per page: put search position and cookielessly measured entry side by side and flag pages with falling clicks at a stable position.
2. Improve citability: clear definitions, structured data and precise answer sections raise the chance of appearing as a source in the AI Overview.
3. Track AI visibility: check regularly whether and how your brand appears in AI answers - via AI visibility.
4. Value the new AI referral channel: recognize visits from AI assistants separately and assign them value instead of losing them in direct traffic.
5. Monitor and report the trend: document the development per page over weeks in reports and status pages, so measures become measurable.

Measuring and monitoring zero-click with XICTRAQ

AI Overviews do not change your rankings - they change what a ranking is worth. That is why zero-click monitoring belongs in the same toolbox as position monitoring: position, impression and real visit together - only then does the intercepted click become visible.

XICTRAQ measures organic entries per page cookielessly and without a consent banner, puts them next to the search position and flags when clicks break away at a stable position. So you see on the affected day which content is lost to the AI answer - and with AI visibility you keep track of whether your brand appears in the answers at all.

Because the measurement is cookieless and never touches device storage, it stays reliable even when visitors reject a consent banner - unlike a measurement that only counts after someone clicks 'Accept'. For the zero-click question that matters twice over: you want to see the full organic entry, not just the consented share of it. Otherwise two losses overlap - the one from the AI answer and the one from rejected consent banners - and can no longer be told apart cleanly.

Start with a free trial and check which of your pages have been quietly losing clicks since AI Overviews took hold. Or book a personal demo to see zero-click analysis and alerting in action.