If your organic traffic slid in the last year and the timing lines up with Google rolling out AI overviews, you are not imagining it and you are not alone. Since Google put an AI-generated summary at the very top of the results page, a large share of searches now get answered before anyone scrolls to a blue link. The result is a pattern SEO teams are seeing everywhere: XX This guide is a diagnostic. It walks through the nine most common, evidence-backed reasons a site loses organic traffic once AI answers enter its queries, using real numbers from Ahrefs, Pew Research Center, Amsive and others rather than speculation. It also separates the drops that AI Overviews genuinely caused from the ones people wrongly blame on them, and it ends with what actually works to recover. No panic, no doom, just a clear read on what changed, why it happened, and the specific, fundable steps that pull traffic and citations back.

Quick Answer

Organic traffic drops after AI Overviews mainly because Google now answers the query on the results page itself, so fewer people click through even when you still rank.Pew Research Center found that when an AI summary appeared, users clicked a traditional search result only 8 percent of the time, versus 15 percent when no summary was shown, and clicked a link inside the summary just 1 percent of the time. Ahrefs, analysing thousands of keywords, measured a 34.5 percent lower average click-through rate for the top-ranking page when an AI Overview was present. The nine reasons below, from lost featured-snippet clicks to query deflection toward ChatGPT and Perplexity, explain the fall. The fix is not to fight the summary but to earn citations inside it, defend the queries a summary rarely touches, and measure clicks rather than rankings, which no longer tell the true story on their own.

What actually changed with AI Overviews

Google introduced AI Overviews at its I/O event in May 2024 and expanded them to more than a hundred countries within months. An AI Overview is a generated answer that sits above the organic results, stitched together from several sources, with a handful of links tucked to the side. For a searcher it often removes the reason to click anything: the answer is right there. That single change is the engine behind almost every traffic story in this guide.

The important shift for anyone doing SEO is that the old scoreboard broke. For two decades, ranking first was close to a guarantee of traffic. With a generated summary occupying the space above you, first place is now a smaller slice of a smaller pie. That is why so many teams see a widening gap in Search Console between impressions, which hold up because you still rank, and clicks, which fall because the summary satisfied the query. Understanding the nine reasons below is the first step; if you want the strategic response in full, our guide to answer engine optimization lays it out.

A short timeline: how the rollout reshaped the SERP

The change did not arrive all at once, and the phased rollout is part of why the traffic impact confused so many teams. Google first showed the feature as the Search Generative Experience, an opt-in Labs experiment, through 2023 and early 2024. At I/O in May 2024 it graduated to a default feature for US users under the AI Overviews name, then expanded through the second half of 2024 to more than a hundred countries and, in 2025, to a far wider range of queries. Google also began layering in a more conversational AI mode alongside the standard summary.

For site owners, that gradual expansion meant the traffic effect crept in rather than crashing down on a single date, which is exactly why it is so easily mistaken for something else. A page might have been fine in March, softened in June as summaries reached its queries, and clearly down by autumn, with no obvious single-day cliff to point at. If you are trying to correlate your own decline with the rollout, do not look for one dramatic drop; look for a steady erosion of clicks on the query clusters that summaries reached in your market and your language, which arrived at different times for different regions.

The 9 reasons your organic traffic dropped

Traffic rarely falls for a single reason. Most sites are hit by several of the following at once, in different proportions depending on their query mix. Read all nine, then use the diagnostic section afterward to work out which ones are driving your own decline.

1. The answer now lives in the SERP, so nobody clicks

This is the root cause behind most of the others. When an AI Overview fully answers a question, the searcher has no reason to visit a website. Pew Research Center put a number on it in July 2025: with an AI summary present, only 8 percent of searches led to a click on a traditional result, against 15 percent without one, and a mere 1 percent of users clicked a link inside the summary itself. That is roughly half the click-through, evaporating at the top of the page before your ranking ever comes into play.

2. Your number-one ranking lost a third of its clicks

Holding the top spot no longer protects you. Ahrefs analysed a large sample of keywords and found the top-ranking page received a 34.5 percent lower click-through rate when an AI Overview was present than when it was not. Nothing about your page changed; the real estate above it did. If your most valuable head terms now trigger a summary, expect a material click loss on exactly the rankings you worked hardest to win.

3. Informational and top-of-funnel content took the first hit

These summaries appear far more often on informational queries, the how-to, what-is and comparison searches, than on transactional ones, because those are the questions a summary answers cleanly. If your traffic leaned on definitional blog posts and beginner guides, that is precisely the content AI Overviews cannibalise first. Sites whose organic mix is heavy on top-of-funnel explainer content have seen the steepest declines, while pages tied to buying intent have held up better.

4. Featured snippets and People Also Ask got absorbed

Before them, position-zero featured snippets and the People Also Ask box sent real clicks to the pages they quoted. The generated answer now occupies that same space and fold those answers into the generated summary. The click that used to follow a featured snippet frequently no longer happens, because the expanded AI answer satisfies the follow-up questions in place. If a chunk of your traffic came from owning snippets, that channel has quietly narrowed.

5. Even branded and navigational queries leaked clicks

It is tempting to assume brand searches are safe, but the data is less comforting. Amsive, in its ongoing analysis of AI Overviews and organic CTR, found click-through declines that extended well beyond generic informational terms, with some query sets losing a large share of clicks once a summary appeared. When Google can summarise who you are and what you do, a slice of people who searched your brand get their answer without ever reaching your site.

6. Zero-click search was already the baseline, and this accelerated it

The summaries did not invent the zero-click search; they poured fuel on a fire that was already burning. SparkToro‘s 2024 study found that for every 1,000 US Google searches, only about 360 clicks went to the open web, the rest ending in no click or a click back into Google’s own properties. AI Overviews push that ratio further in the same direction. If your traffic model assumed most searches end in a visit, the ground had already shifted before the summaries scaled.

7. You are cited in the overview but not clicked

There is a cruel version of this problem: your page is one of the sources the AI Overview draws from, yet you gain almost no traffic from it. Being quoted is not the same as being clicked, and with only around 1 percent of users following links inside summaries, citation visibility rarely converts into sessions. This is why measuring rankings alone is now misleading. You can be more visible than ever inside AI answers and still watch sessions fall, which is exactly why you have to track your brand’s AI citations as a distinct metric.

8. Search demand is deflecting to ChatGPT, Perplexity and AI mode

Some of the queries you used to win are no longer happening on Google at all. A growing share of research questions now start inside ChatGPT, Perplexity or Google’s own AI mode. Gartner predicted that traditional search engine volume would drop 25 percent by 2026 as users shift to AI chatbots and virtual agents. When the query migrates off the search engine, the ranking that used to capture it simply has nothing to capture. Winning that demand back means being visible inside those engines, the core of generative engine optimization.

9. Publishers and content sites saw referral traffic contract

The pattern is visible at the macro level too. Analyses from Similarweb and reporting across the publishing industry have tracked falling search referral traffic to news and content sites as these summaries expanded, with some publishers reporting double-digit declines in Google referrals. If you run a media, affiliate or content-heavy site, part of your drop is this structural contraction in how much traffic Google sends to the open web, not anything specific to your individual pages.

Notice how tightly these interlock. The SERP answers the query (reason 1), which strips clicks from your top ranking (reason 2), hits your informational pages hardest (reason 3), absorbs your snippets (reason 4), and compounds a zero-click trend that was already underway (reason 6). Diagnosing your own decline is mostly a question of working out the mix.

What the data really says, in plain numbers

It is worth pulling the headline figures together, because the individual studies measure slightly different things and the combined picture is stronger than any one of them. Treat these as directional evidence rather than precise laws, since methodology and query samples differ, but the direction is unmistakable.

• Click-through on the top result:Ahrefs measured a 34.5 percent lower average CTR for the number-one page when a summary was present. That is the single most-cited figure, and it isolates the effect on the ranking that used to be safest.
• Overall click behaviour:Pew Research Center found result clicks fell from 15 percent to 8 percent of searches when a summary appeared, and that only 1 percent of users clicked a link inside the summary. This measures real user sessions, not just modelled CTR curves.
• Organic CTR erosion:Amsive and Seer Interactive have both published ongoing analyses showing meaningful CTR declines across query types once a summary is shown, including on some queries teams assumed were insulated.
• The zero-click baseline:SparkToro found only around 360 of every 1,000 US Google searches led to a click to the open web even before summaries scaled, so the new feature is accelerating a trend, not starting one.
• The demand shift:Gartner projected a 25 percent fall in traditional search volume by 2026 as users move to AI chatbots and agents.

Put together, these say the same thing from five angles: the same query now produces fewer clicks to the open web, the effect is largest on the highest rankings and the most informational queries, and part of the demand is leaving Google altogether. That is the shape of the problem you are diagnosing.

How to tell which reasons are hitting you

Before you react, confirm what you are actually looking at. A drop that lines up with the summaries behaves differently from a drop caused by a core update or a technical problem, and the fix depends entirely on which one it is.

Start in Google Search Console and compare impressions against clicks over the period. The signature of an AI Overviews impact is a widening gap: impressions flat or rising while clicks fall, which means you still rank but fewer people click. Then segment by query type. If the decline concentrates on informational, how-to and definition queries while transactional and branded terms hold, that points squarely at reasons three and four above. Check whether your affected queries actually show an AI Overview by searching a sample of them; if most trigger a summary, the diagnosis is confirmed. Finally, overlay the timeline of Google’s confirmed search updates against your drop. If the fall aligns with a documented core update rather than the AI Overviews expansion in your market, you may be misattributing the cause, which is the subject of the honest section below.

The impressions-clicks gap, explained

If you take one habit from this guide, make it this: stop reading average position as your health metric and start reading the ratio of clicks to impressions. That ratio is where the whole story hides. When a summary answers a query in place, your listing still gets shown, so impressions are logged, but the click never happens, so your click-through rate quietly sinks. A page can look perfectly healthy on a rankings dashboard while it is steadily losing the sessions that actually matter to the business.

In Search Console, pull a twelve-month view and plot clicks and impressions on the same chart. On a page or query where an answer box arrived, you will typically see the two lines, which used to move together, visibly separate: impressions flat or climbing while clicks bend downward. Segment that view by page type and you can usually watch it happen first on your explainer and how-to content, then spread. The value of naming the pattern is that it stops you from chasing the wrong fix. You are not being outranked, so more link building or another on-page rewrite of an already number-one page will not bring the clicks back. The lever that works is changing which queries you compete for and how you show up in the answer itself. This is also the metric to bring to leadership, because it reframes the conversation from an alarming, unexplained traffic loss into a specific, understood shift with a defined response, which is a far easier case to fund and a far more honest one to make than a vague warning that the traffic is simply down.

What actually works to recover traffic

The instinct to fight the summary is a losing one. The teams recovering fastest have stopped trying to claw back the click that Google now intercepts and started competing for something else: the citation, the high-intent query, and the visibility inside AI engines. Here is where the effort pays.

First, optimise to be the source the AI Overview quotes. That means clear, extractable answers near the top of the page, strong structured data that engines can lift cleanly, and the E-E-A-T signals that make a model trust you as a source. Being cited will not fully replace the lost click, but citation is the new visibility, and the sites that win citations are the ones that stay in the conversation. Second, defend the queries AI Overviews rarely touch: high-intent, transactional and bottom-of-funnel searches where the searcher wants to act, not just learn. Those clicks are far more resilient and far more valuable. Third, extend beyond Google entirely, because a real share of demand has moved to AI engines; making your content legible to them, including publishing a clean llms.txt file, is cheap forward-looking hygiene. Fourth, and non-negotiable, change what you measure. If your scoreboard is still rankings and sessions alone, you are blind to citation visibility; you need to track AI citations and watch that trend alongside clicks. The full toolkit for doing this sits in our roundup of the best AEO and GEO tools, and it is the day-to-day work of our AEO team.

Four mistakes teams make when they react to the drop

The panic response usually does more harm than the decline itself. These are the errors we see most often when a team first realises its clicks are falling.

• Blaming a single cause. Assuming the entire drop is one thing, whether a summary or a core update, leads to a fix aimed at the wrong problem. Almost every real decline is a blend, and the split between lost rankings and lost clicks tells you where to spend effort.
• Deleting or de-indexing content in a hurry. Ripping down informational pages because they lost clicks can strip the very authority and topical depth that earns citations and supports your money pages. Prune with evidence, not fear.
• Chasing the click that is gone. Pouring resources into recovering traffic on queries a summary now fully answers is often a poor bet. Redirect that energy toward high-intent queries and toward being the cited source instead.
• Not changing the report. Continuing to judge success on sessions and average position alone, with no measure of citation visibility, means you cannot even see the new game you are being asked to play. Fix the scoreboard before you judge the team.

The honest part: not every drop is AI Overviews

It would be convenient to blame every lost visitor on AI Overviews, but doing so leads to the wrong fixes. Several other forces run on the same timeline and get misattributed constantly. Google shipped multiple broad core updates across the same period, any of which can reshuffle rankings independently of AI Overviews. Seasonality moves traffic in ways that have nothing to do with the SERP. Keyword cannibalization, where several of your own pages compete for one query, can bleed clicks quietly. And ordinary technical decay, from slow pages to lost backlinks, still matters.

The discipline is to separate the click problem from the ranking problem. If your rankings themselves dropped, that is a content, authority or technical issue and AI Overviews are probably not the culprit. If your rankings held but your clicks fell, that is the AI Overviews signature. Most sites have a bit of both, and honest diagnosis, rather than a single scapegoat, is what leads to a fix that works. A traffic drop after AI Overviews is real, but so are the boring, fixable problems it is often confused with, and the discipline of strengthening your fundamentals helps against all of them at once.

Who is hit hardest, and who is relatively safe

The damage is not evenly spread, and knowing where you sit tells you how urgent this is. The sites bleeding the most are content and media publishers whose model depends on top-of-funnel informational traffic, affiliate sites built on how-to and comparison queries, and any brand whose organic strategy was mostly definitional blog content. Those are exactly the queries AI Overviews answer best, so the click loss is steep.

Relatively more protected are businesses whose traffic clusters around high-intent, transactional and local queries, where the searcher wants to buy, book or contact rather than simply read. A search for a specific product, a local service or a pricing page is far less likely to be resolved by a summary, so those clicks survive. This is why diversifying toward bottom-of-funnel intent, and toward channels beyond Google, is the structural answer rather than a tactical one. It is also why pairing resilient SEO with a deliberate presence inside AI engines is becoming the default shape of a modern program, and why understanding which sites AI answers cite most is now part of the strategy.

Building a program that survives the next shift

The uncomfortable truth is that summaries are not the last change of this kind; they are one step in search becoming more answer-led and less link-led. A program built only to rank blue links is fragile by design. A program built to be the trusted, citable authority on its topic, wherever that topic is answered, is far more durable, because the same signals that earn a citation in a Google summary also earn one in ChatGPT or Perplexity.

Practically, that resilience rests on three habits. Publish genuinely original, experience-backed content that a summary has to quote rather than paraphrase away, since thin, derivative pages are the easiest for an answer engine to replace. Keep your machine-readable layers clean and current, from structured data to a maintained llms.txt, so every engine can parse you without guessing. And measure across surfaces, watching not just Google clicks but where and how often you appear in AI answers, because the brands that treat citation as a first-class metric are the ones that adapt before the next shift instead of after it. Done together, these turn a defensive scramble into a durable position, which is the whole point of a modern GEO and AEO strategy.

Frequently asked questions

Did AI Overviews really cause my organic traffic to drop?

If your rankings held steady but your clicks fell, and your affected queries now show an AI summary, then AI Overviews are the likely cause. Pew Research found clicks roughly halved when a summary appeared, and Ahrefs measured a 34.5 percent lower CTR for the top result. If your rankings themselves dropped, the cause is more likely a core update or a technical issue.

How much traffic do AI Overviews take away?

It varies by query mix, but the studies are consistent in direction. Ahrefs found a 34.5 percent lower click-through rate for the top-ranking page when an AI Overview was present, and Pew Research found overall click rates dropping from 15 percent to 8 percent. Informational and top-of-funnel content loses the most.

Which pages lose the most traffic to AI Overviews?

Informational content: how-to guides, definitions, comparisons and beginner explainers. These queries trigger AI Overviews most often and are the easiest for a summary to answer completely. Transactional, branded and local-intent pages are more resilient.

Can I recover traffic lost to AI Overviews?

You rarely recover the exact click Google now intercepts, but you can recover value by becoming the source the summary cites, defending high-intent queries, and expanding into AI engines like ChatGPT and Perplexity. The goal shifts from winning the click to winning the citation and the resilient, buying-intent traffic.

Is being cited in an AI Overview worth anything if nobody clicks?

Only about 1 percent of users click links inside summaries, so citation rarely drives direct sessions. Its value is brand visibility and trust: appearing as a trusted source shapes buyer perception and feeds the wider AI ecosystem. It is worth pursuing, but it should be measured as citation visibility, not as clicks.

How do I know if my drop is AI Overviews or a Google core update?

Compare impressions and clicks in Search Console. A widening gap, impressions flat while clicks fall, is the AI Overviews signature because you still rank but fewer people click. If rankings themselves dropped and the timing matches a confirmed core update, that is a different problem with a different fix.