LLalit Patel

How AI Overviews and AI Mode changed search traffic

AI search, explained intermediate Verified 2026-09-06

In short

AI Overviews answer the query on the results page, so fewer people click through. Pew Research found users clicked a result in 8 percent of visits with an AI Overview against 15 percent without. Impressions hold steady or rise while clicks fall, which is the pattern to check in your own Search Console.

Key points

  • Every serious independent study finds the same direction, fewer clicks when an AI answer is present. They disagree about size.
  • Pew measured real browsing, not lab conditions, and found clicks roughly halved when an Overview appeared.
  • Google disputes the studies and says overall click volume is stable and click quality is up. Both can be partly true.
  • The visible symptom in Search Console is impressions steady or rising while CTR falls on informational queries.
  • The response is not to quit content. It is to shift weight toward queries that still click, and toward being cited.
01

The short version

When an AI answer sits above the results, fewer people click the results. That direction is not seriously contested any more. The size is, and this page keeps the two questions separate, because the direction is what changes your strategy while the size is what everyone argues about on LinkedIn.

The other honest note up front: nobody outside Google has clean data, because AI Overview clicks are indistinguishable from ordinary Google clicks in your analytics. Everything below is measured from panels, scrapes and aggregates, each with known limitations, which is why the numbers vary and the direction does not.

02

What the published numbers say

Pew Research Center, July 2025. The most trustworthy study, because it watched real behaviour: browsing data from around 900 US adults in March 2025. When a search produced an AI Overview, users clicked a traditional result in 8 percent of visits. Without one, 15 percent. Clicks on the sources cited inside the Overview were around 1 percent of visits, and users seeing an Overview ended their browsing session more often. Read plainly: the Overview answers the question, and the citation links barely compensate.

Ahrefs, early 2025. A large-scale correlation across informational keywords found the top-ranking page's click-through rate roughly a third lower when an AI Overview was present. Correlational, but consistent with everything else.

Similarweb, 2025. For news publishers, the share of searches ending without any click rose sharply year over year as Overviews expanded, from a little over half toward seven in ten. News is the worst-hit category, which matters when you read publisher headlines about collapse: your niche is probably not news.

Google's rebuttal, August 2025. Google states that total organic click volume from Search is roughly stable year over year and that clicks from pages with AI answers are of higher quality, users who click after reading a summary bounce less. Google also disputes the methodology of the third-party studies. Note what is and is not being denied: stable total clicks across an expanding query volume is fully compatible with fewer clicks per informational query. Both sides can be describing the same elephant.

03

How to check your own site

Ten minutes in Search Console, and you are looking for a divergence, not a decline.

  1. Compare the last three months with the same period last year, on queries, not pages.
  2. Filter to informational queries, the ones phrased as questions or definitions.
  3. Look for the signature: impressions flat or up, clicks and CTR down. That gap is the answer layer absorbing clicks that impressions used to convert.
  4. Segment brand versus non-brand. Brand queries should be stable. If they are falling too, your problem is probably not AI Overviews.
  5. Write the numbers down monthly. The trend is the evidence, and you will want the baseline later.

A plain traffic fall with falling impressions is a different disease, rankings or indexing, and AI Overviews make a convenient scapegoat for it. Check before blaming.

04

What to actually do

  • Reweight, do not retreat. Queries with transactional or local intent still click. Pure definitions mostly no longer do. Spend new content effort accordingly.
  • Compete for the citation. If the answer box is where attention goes, being named inside it is the remaining visibility. The next page covers how sources get picked.
  • Build reasons to be visited that an answer cannot replace. Tools, data, templates, anything interactive. A summary cannot substitute for a calculator.
  • Watch value per visit, not just visits. Fewer, better-qualified clicks is survivable if conversion holds. Falling visits with flat conversions changes the maths of content entirely.
05

When this does not apply

If your traffic is mostly brand searches, direct, or from queries with strong commercial intent, the published declines overstate your exposure. And if your site sells to people who search transactionally, buy, price, near me, the Overview era has so far taken little from you. The pages this hits are the informational middle: definitions, explainers, list posts, the traffic that content marketing spent a decade optimising for.

06

FAQ

Do AI Overviews reduce clicks? Yes, on the queries where they appear. Pew Research measured real browsing and found users clicked a traditional result in 8 percent of visits when an AI Overview was present against 15 percent without one. Other studies differ on size but agree on direction.

What did the Pew study find about AI Overviews? Using March 2025 browsing data from around 900 US adults, Pew found clicks on traditional results roughly halved when an Overview appeared, clicks on the sources cited inside Overviews were about 1 percent of visits, and sessions ended more often after seeing one.

What does Google say about AI Overviews and traffic? Google says total organic clicks from Search are roughly stable year over year, that clicks from pages with AI answers are higher quality, and that third-party click studies use flawed methodology. Stable totals and falling clicks per informational query can both be true at once.

How do I know if AI Overviews affected my site? Look for a divergence in Search Console: impressions flat or rising while clicks and CTR fall, concentrated in informational, question-shaped queries, with brand queries stable. A decline in both impressions and clicks points to rankings or indexing instead.

Which queries lose the most traffic? Informational ones: definitions, explainers, how-tos, health and finance questions, and news. Transactional, local and navigational queries have been far less affected, because their answer is still a website rather than a paragraph.

What should I do about falling search traffic? First confirm the cause in Search Console. If it is the answer layer, shift new content toward queries that still click, compete for citations inside the answers, and build assets a summary cannot replace, such as tools and original data. Track value per visit, not only visits.

07

Sources

  • Pew Research Center, July 2025: analysis of real user browsing during March 2025
  • Ahrefs, 2025: CTR analysis of informational keywords with and without AI Overviews
  • Similarweb, 2025: zero-click share for news publishers, year over year
  • Google, The Keyword blog, August 2025: Google's response on click volume and quality

Last verified . This page is reviewed on a fixed cycle. Changes:
  • First published.