Is SEO Dead? What the Data on AI Search and Clicks Shows

Are people really clicking less?

Yes, on the specific queries where an AI summary appears — this is the most directly measured part of the question. Pew Research Center tracked 900 US adults' real browsing behavior across 68,879 Google searches (March 2025), of which 12,593 produced an AI summary. When an AI summary appeared, people clicked a traditional search result link in 8% of visits. When it didn't, they clicked in 15% of visits. Clicks on a link inside the AI summary itself happened in just 1% of visits. People were also more likely to end their search session entirely after seeing a summary page (26%) than after a page without one (16%).

That's a real, directly measured gap — roughly half the click rate on AI-summary queries versus non-summary ones, in a representative panel, not a self-reported survey. It's also narrower in scope than "AI search killed clicks" suggests: Pew found AI summaries appeared on 18% of all Google searches in the sample, concentrated on longer, more conversational queries (question-format and ten-plus-word searches produced summaries far more often than one- or two-word searches).

Who is affected and who is not?

Unevenly, and the newest longitudinal data suggests the first wave of drops wasn't the whole story. Seer Interactive tracked 42 client organizations across 2.43 billion impressions and 5.47 million queries from June 2024 through February 2026. Organic CTR on AI-Overview queries fell from a 1.76% baseline in June 2024 to a low of 0.61% in September 2025 — a 65% drop at the floor — then recovered to 2.4% by February 2026, still roughly 37% below the 3.8% CTR on comparable non-AI-Overview queries in the same period. Paid CTR followed a similar pattern, dropping further (68%) before partially recovering.

The more actionable finding in that data: being cited inside the AI Overview itself carries a real advantage over not being cited. Seer found cited brands earned 35% higher organic CTR and 91% higher paid CTR than uncited competitors on the same query. That reframes the question from "are AI Overviews bad for clicks" to "what happens to brands that are cited inside them versus brands that are absent" — and the second group is doing meaningfully worse than the first, not just worse than the pre-AI baseline.

Separately, Ahrefs' August 2025 study (15,000 long-tail queries) found only about 12% overlap between links cited by ChatGPT, Gemini and Copilot and Google's own top 10 for the same query — with Perplexity a notable outlier at 28.6%. Roughly 80% of AI citations didn't rank anywhere in Google's top 100 at all. That's a different signal than the click-rate data: it says ranking well on Google doesn't reliably predict getting cited by most AI engines, which is a separate risk from the click-rate drop on Google's own AI Overviews specifically.

What does Google say about SEO and AI features?

Directly, in its own published guidance: "In short, yes! The best practices for SEO continue to be relevant because our generative AI features on Google Search are rooted in our core Search ranking and quality systems." The same guide explicitly debunks several tactics marketed as "necessary for AI visibility" — special markup, breaking content into tiny chunks, llms.txt files (which it says Google Search doesn't use and which "neither harm nor help" rankings) — and instead recommends the same fundamentals SEO has asked for all along: crawlable, well-structured, satisfying content with "a unique point of view."

That statement is about Google's own AI features specifically. It says nothing about ChatGPT, Perplexity or Claude, which don't run on Google's ranking systems and haven't published an equivalent statement. The Ahrefs overlap data above is a reasonable proxy for why: if ranking well on Google only weakly predicts citation on other engines, Google's "SEO still applies" guidance doesn't automatically transfer to them.

What do forecasts claim and how reliable are they?

One widely repeated forecast deserves more scrutiny than it usually gets. In February 2024, Gartner published a press release attributing to VP Analyst Alan Antin the prediction that "traditional search engine volume will drop 25% by 2026," with "search marketing losing market share to AI chatbots and other virtual agents." It's a real, dated, named prediction, not an unsourced number — unlike some of the figures (like the frequently cited "$67.4 billion hallucination cost") that circulate without a traceable origin at all.

But it is a forecast, made nearly two years before its own target date, not a measured result. Reporting closer to that target date has questioned whether the 25% figure materialized at the scale predicted, though we haven't independently verified a specific current market-share number from a primary analytics source strong enough to state one here [VERIFY]. The honest read: treat the Gartner figure as "a 2024 prediction from a named analyst, now past its own checkpoint date, without confirmed independent data validating it" — not as a settled fact about 2026 search volume.

What should a team measure now?

Add to rankings, don't replace them. The data above supports three things being true at once: SEO fundamentals still matter (Google says so directly, and the mechanism — ranking and retrieval feeding generation — explains why), clicks really do fall on the specific queries where an AI summary shows up, and getting cited inside that summary meaningfully offsets the click loss compared to being absent from it.

That means the KPI stack needs an addition, not a swap:

For the queries where you already rank, start checking whether you're also named and cited in the AI answer sitting above those rankings — how to measure AI visibility covers the four metrics to track alongside your existing SEO reporting, and how AI citations compare to Google's top 10 covers why a top ranking doesn't guarantee a citation.

FAQ

Did Pew's study measure clicks or traffic? Clicks per visit, from real browsing panel data — not self-reported traffic or aggregate site analytics. 900 US adults, 68,879 searches, March 2025 data, published 22 July 2025.

Does an AI Overview always reduce my click-through rate? Not evenly. Seer Interactive's longitudinal data shows the steepest drop was in September 2025, with partial recovery by February 2026, and brands cited inside the AI Overview had substantially higher CTR than uncited competitors on the same queries.

Is the "25% drop in search volume by 2026" claim confirmed? It's a real prediction, from Gartner in February 2024, correctly attributed and dated. Whether it was confirmed by 2026 at that scale isn't something we can verify from a primary dataset at time of writing — treat it as an unconfirmed forecast, not a settled number.

Does ranking #1 on Google mean I'll be cited by ChatGPT or other AI engines? No. Ahrefs found only about 12% overlap between AI citations and Google's own top 10, varying by engine (Perplexity closer to 29%, ChatGPT closer to 8%). A Google ranking doesn't reliably predict citation elsewhere.

What to do next

1. Check whether your best-ranking pages for buyer-intent queries are also being cited in the AI answers that sit above those rankings, not just assume they are.
2. Split your CTR reporting by AI-summary-present vs. absent queries where your analytics allow it, rather than reading one blended average.
3. Keep your SEO roadmap. Add AI-answer visibility as a second, separately tracked line — see how to measure it.

Check one real query with the AI Citation Snapshot to see whether you're named in a live AI answer today. Free, one run.