Google’s AI summaries have changed what happens after a search. When an AI Overview appears at the top of the results page, many users get their answer without clicking any link below it. The question for every business that depends on organic traffic is how much this costs in practice.
Several independent studies have measured that cost, and they agree on direction even when the exact size differs. Pew Research Center found that users clicked on a result 8% of the time when an AI summary appeared, compared with 15% without one. That is a relative reduction of about 47%. Ahrefs reported that AI Overviews reduce clicks by 34.5%. DMG Media’s worst-case figure put click-through drops as high as 89%. The range is wide, and the lesson is not that every page loses nearly half its traffic. The lesson is that the effect is large enough to plan around.
Why the Loss Is Uneven
The drop does not fall evenly across queries. Informational searches, where a user wants a short answer to a clear question, are the most exposed, because the summary can satisfy them completely. Searches where the user needs to compare options, make a purchase, or read something in depth are less likely to end at the summary.
That means a business should not judge its SEO by total traffic alone. A page that answers a simple question may lose clicks while still being valuable. A page that supports a considered decision may hold its traffic. Separating your pages by the kind of question they answer is the first useful step.
Visibility Has a New Meaning
Lower clicks do not mean lower visibility. Being cited or referenced inside an AI answer is a different kind of exposure. The brand is seen even when the user does not visit the site, and that exposure can shape later searches and recommendations.
Getting cited depends on how content is built. Answers that appear clearly near the top of a page are easier for systems to extract. Concrete data backed by a named source gives an AI system something it can attribute. Structured headings that match real questions help the system map content to a query. None of this replaces good content, but it changes what good content needs to look like.
What to Check on Your Own Site
Start with three checks. First, confirm that your key pages do not block the crawlers that feed AI systems in your robots.txt file. A rule written for one crawler can accidentally exclude others. Second, look at the opening of your most important pages. If the direct answer sits three paragraphs down, consider moving it up. Third, verify that any FAQ content uses correctly implemented schema markup, so the question and answer pairs are machine-readable.
Then measure. Track which pages still receive clicks and which have lost them, and note whether those losses cluster around informational queries. Use that pattern to decide where to invest in restructuring and where to leave a page alone.
Plan Around the Data, Not the Headline
Google changes how AI features work over time, so no single percentage will stay accurate. What holds up is the underlying pattern: AI summaries reduce clicks, the reduction varies by query type, and visibility now includes being cited. A business that tracks its own pages against that pattern will make better decisions than one that reacts to a single statistic.
If you do not yet know how your site appears in AI answers, dpanell’s free AI Search Visibility Audit gives you a starting point. It shows the gaps worth fixing first, so you can focus effort where the data says it matters most.
