Are Google AI Overviews Killing Organic Traffic? What the Latest Data Shows

Reviewed 4 August 2026. Google AI Overviews are reducing the click opportunity for many informational searches, but the available evidence does not support the claim that they have made SEO or organic traffic worthless. The effect varies sharply by query intent, citation status and how a study defines a click or a zero-click search.

For Australian businesses, the practical response is not to abandon search. It is to stop treating every ranking and impression as equally valuable, measure AI-exposed queries separately, and create pages that give people a reason to visit after the summary.

What the latest data says

  • Pew observed a traditional-result click in 8% of Google visits with an AI summary, versus 15% without one. A cited AI Overview source received a click in only 1% of visits.
  • Ahrefs estimated that an AI Overview was associated with a 58% lower position-one CTR in December 2025 than its counterfactual forecast.
  • Seer found that AIO-present organic CTR recovered from 1.3% in December 2025 to 2.4% in February 2026, although it remained below the 3.8% rate for queries without an AIO.
  • Semrush found that zero-click behaviour decreased from 33.75% to 31.53% for a cohort of keywords after they began showing AI Overviews, contradicting a simple “AIO always causes more zero-click searches” story.
  • Google says total organic clicks to websites are broadly stable and click quality has improved, but it has not published the underlying sample or AI-Overview-specific figures needed to independently test that claim.

So, are AI Overviews killing organic traffic?

They are taking a meaningful share of clicks from many answer-seeking searches. That is the clearest conclusion supported by independent clickstream and Search Console studies. It is especially relevant to definitions, quick facts, how-to questions and comparisons that Google can summarise on the results page.

They are not eliminating the commercial value of search. High-intent product, service, local and transactional searches remain less exposed in most datasets. Even on AI Overview results, being cited can improve click-through rate relative to being absent. Search visibility can also influence later branded searches and conversions that are not credited to the original impression.

The honest answer is therefore “some traffic, for some queries”. A business built on large volumes of generic informational page views faces more risk than a local service firm whose important searches end in a call, quote request or store visit.

What the leading studies actually measured

The headline figures are not interchangeable. Pew measured what a panel member did after a Google visit. Ahrefs estimated the effect on the top organic result. Seer combined brand-level Search Console, paid and SERP data. Semrush focused on zero-click behaviour and keyword cohorts. Google discussed total ecosystem clicks. Each answers a different question.

Source and periodSampleFindingImportant limitation
Pew Research Center, March 2025 behaviour matched to April 2025 SERPs900 US adults; 68,879 unique Google searches; 12,593 returned an AI summaryTraditional result clicked in 8% of visits with an AI summary and 15% without; AI source clicked in 1%Observational US panel. Results were recreated after the browsing month, and query intent was not experimentally assigned.
Ahrefs, December 2023 versus December 2025300,000 desktop keywords: 150,000 with an AI Overview and 150,000 informational controls; aggregated Search Console dataActual position-one CTR of 1.57% versus a 3.73% counterfactual estimate, a 58% relative differenceThe two-year comparison can contain other SERP and behaviour changes. It is a modelled counterfactual, not a randomised test.
Seer Interactive, January 2025 to February 202653 brands; 5.47 million tracked queries; 2.43 billion organic impressionsAIO-present CTR rose from 1.3% in December 2025 to 2.4% in February 2026; no-AIO CTR was 3.8% in FebruaryAIO status was based on the latest SERP snapshot and applied retrospectively. One account materially influenced one segment.
Semrush and Datos, January to October 2025More than 200,000 keywords with volume above 100, including a cohort that changed from no AIO in May to AIO in OctoberWithin the changing cohort, zero-click rate fell from 33.75% to 31.53%Zero-click definitions and the expanding mix of commercial and navigational queries make this different from measuring position-one CTR.
Google, statement published August 2025Not disclosedTotal organic click volume described as relatively stable year on year, with slightly more “quality clicks”No underlying dataset, country split, query split or AI-Overview-specific CTR was published.

An original comparison of the click findings

The small multiples below use each study’s own metric and scale. They should not be read as one universal CTR benchmark. They show that reduced click-through is visible under several measurement approaches, while Seer’s early-2026 recovery makes a straight-line decline an unsafe forecast.

AI Overview click measures point in the same broad direction

SEO Summer analysis of Pew, Ahrefs and Seer figures. Different samples and metrics; compare within each panel only.

Pew: share of Google visits followed by a traditional-result click

No AI summary15%
AI summary8%

Ahrefs: position-one CTR, December 2025

Without-AIO forecast3.73%
Actual with AIO1.57%

Seer: all-query organic CTR, February 2026

AIO not shown3.82%
AIO shown2.36%

Sources: Pew Research Center, July 2025; Ahrefs, February 2026; Seer Interactive, April 2026.

Why the studies conflict

They measure different outcomes

A fall in position-one CTR is not identical to an increase in zero-click searches. A person might skip the first result but click another organic result, an ad, a map listing or a cited source. They might also reformulate the query and visit the business later. Pew’s “next action”, Ahrefs’ top-result CTR and Semrush’s zero-click classification therefore have different denominators.

AI Overviews select queries that were already less clickable

Informational questions have historically produced featured snippets, knowledge panels and other direct answers. Seer found that the query types which later showed AI Overviews already had lower CTR before the feature was attached to them. This selection effect can exaggerate the apparent impact if a study compares AIO and non-AIO keywords without a credible baseline.

The query mix is changing

Semrush’s analysis of more than 10 million keywords found AI Overviews settling at about 16% of queries during 2025, while their presence expanded beyond purely informational searches. A later Semrush study of more than 600,000 US desktop keywords found a 71% increase in commercial-intent SERPs with an AI Overview between November 2025 and April 2026. Transactional presence fell 5% over the same period.

That shift matters. Commercial and navigational searches are naturally more likely to end in a click than simple fact queries. Adding those terms to an AIO sample can reduce the measured zero-click rate even if the same feature still suppresses clicks on a typical how-to query.

Aggregate clicks and per-query CTR can both tell the truth

Google says people are searching more and that total organic click volume to websites is relatively stable. Independent studies say CTR falls when an AI Overview appears. Those claims are not logically incompatible. If total search volume grows while clicks per impression fall, aggregate clicks can remain stable.

The problem is verification. Google has not published the underlying numbers, and AI feature traffic is grouped into the normal “Web” search type in Search Console. Site owners cannot cleanly isolate AI Overview impressions and clicks using Google’s own report.

Which Australian businesses are most exposed?

Most published studies use US users, US keyword databases or multinational client accounts. They should be treated as directional for Australia, not as a precise Australian benchmark. Your exposure depends more on search intent and page type than on a single global percentage.

ExposureTypical searchesWhat to do
HigherDefinitions, simple how-to questions, broad comparisons, basic health or finance explanations, commodity informational contentAdd original evidence, tools, examples, expert judgement and a useful next step that cannot fit in a summary.
MixedProduct research, software comparisons, professional-service education, travel planningProtect comparison and category pages, build brand authority, and monitor whether citations lead to later branded demand.
Lower, not zero“Near me”, urgent services, quote requests, bookings, stock availability and specific purchase searchesKeep local, service and conversion pages accurate and prominent. Track Maps, calls and leads as well as website clicks.

What businesses should change now

1. Segment performance by intent, not just page

In Search Console, separate informational questions, comparisons, branded terms, local terms and transactional searches. Compare clicks, impressions, CTR and average position year on year. A CTR fall with stable ranking and rising impressions is a different problem from a ranking loss.

2. Audit the queries most likely to trigger an AI Overview

Manually review the highest-impression questions and comparisons in your market, using a consistent Australian location and device. Record whether an AI Overview appears, whether your brand is cited, and which competitors or publishers are used as sources. Repeat the sample monthly because results change.

3. Give the searcher a reason to click

Generic summaries are easiest to replace. Pages are more defensible when they contain first-party data, Australian benchmarks, interactive calculators, detailed pricing logic, original images, genuine case evidence, expert trade-offs or a service the reader can act on. This is also consistent with Google’s official advice to create helpful, reliable, people-first content.

4. Treat citation visibility and website traffic as separate outcomes

Seer’s 2026 study found that an AI Overview citation delivered 120% more organic clicks per impression than being uncited on an AIO result, but still underperformed a result without an AI Overview by 38%. Earning citations can protect visibility, but it should not be reported as equivalent to a traditional number-one ranking.

5. Measure the business result

Track qualified leads, bookings, calls, assisted conversions and branded search demand. Session volume from broad informational content may decline while the remaining visits become more deliberate. Google calls these “quality clicks”, but each business should test that claim against its own conversion data.

You can compare the channel trade-offs in our guide to SEO versus Google Ads.

A practical forecasting adjustment

Do not apply the Ahrefs 58% estimate to every keyword in an SEO forecast. It describes a modelled average for position-one desktop keywords in a specific informational sample. A more defensible plan uses scenarios.

ScenarioIllustrative CTR adjustmentUse when
Lower exposure0% to 15% below your recent non-AIO baselineLocal, branded or transactional terms where an AIO rarely appears
Moderate exposure15% to 40% below baselineCommercial research and detailed informational searches with mixed SERP layouts
Higher exposure40% to 60% below baselineSimple questions and comparisons that consistently show a prominent AI Overview

These bands are planning assumptions, not universal benchmarks. Replace them with your own observed deltas as soon as you have enough impressions. SEO Summer’s SEO ROI calculator can help translate revised traffic expectations into a commercial range.

Limitations of the available evidence

  • Most public datasets are US-heavy and may not match Australian SERPs.
  • AI Overview presence changes by device, location, user and date.
  • Observational studies cannot fully separate the feature’s effect from query intent, ranking changes, other SERP features and broader behaviour shifts.
  • Search Console does not provide an AI Overview filter, so third parties must infer presence and citation status.
  • CTR does not capture brand exposure, later searches, offline conversions or click quality.
  • Very large impression counts do not automatically make a sample representative. Query and account concentration still matter.

The bottom line

AI Overviews are not killing all organic traffic, but they are changing which searches produce a website visit. The most credible evidence points to meaningful click suppression for answer-seeking queries, with a less severe and more volatile effect for commercial and transactional searches.

Businesses should keep investing in SEO services when search contributes to discovery and sales, but the strategy needs to move beyond publishing commodity answers and reporting rankings alone. The winning pages will be the ones Google can cite and people still need to visit, which is the focus of our AI SEO services. Organisations that need evidence-led production can use our SEO copywriting service.

i 3 Table Of Content

SEO Summer is a digital agency based in Melbourne Australia. We've helped 100+ clients reach the first page of Google Search. Get in touch to see how we can help you.

Our Marketing Services

Social Media Marketing
Email Outreach

Written By Victor Volynski

Vic is the Founder and Lead SEO at SEO Summer. He has over 10 years experience in building brands through digital marketing. Vic has worked with businesses in all niches, all the way from local businesses to international brands.

Vic focuses on the cutting edge of SEO, employing the most effective techniques in the space to achieve the best results for clients.