State of AI Search 2026: The Numbers That Actually Hold Up

Which AI search numbers actually hold up in 2026?

A number "holds up" here if it traces to a named publisher, a stated sample and a specific date, and if the publisher's own framing is preserved rather than flattened into a punchier headline. By that bar, five figures are worth repeating as of October 2026: Pew Research's click-rate drop when an AI summary appears, Ahrefs' measurement of how rarely AI citations match Google's top 10, Seer Interactive's organic and paid click-through data for AI Overview queries, Fractl's survey of how marketers actually talk about this work internally, and SEOFOMO's practitioner survey on budget and tracking adoption. Each is detailed below with what it measured, what it didn't, and where it conflicts with another study measuring something that sounds similar but isn't.

What did each study actually sample?

That last row is included deliberately, as a contrast case: it's a real, dated figure, but it comes from a vendor analysis republished in a roundup with no disclosed sample. Use it with that caveat attached, every time.

Where do studies disagree, and why?

How many teams track AI visibility at all: 22% or 92%? Both numbers are real, and they disagree because they sample different populations in different ways. The 22% figure (Averi analysis, via getpassionfruit.com, March 2026) has no disclosed sample and reads like a broad market estimate. SEOFOMO's 92% (up from 78% in 2025) comes from 171 self-selected search professionals who follow a practitioner newsletter closely enough to answer its survey — a population that already skews toward people who track this kind of thing for a living. Neither number describes "the average company." Quote each with its sample, and don't average them.

Is AI Overview CTR down 61% or up 85% from its floor? Seer's own study reports both, because it's tracking a trend over time, not a single snapshot. Organic CTR on AI Overview queries fell from a June 2024 baseline to a floor of 0.61% around September 2025 — a steep decline — but had partially recovered by early 2026 in Seer's continuing coverage of the same dataset. Reporting only the floor, or only the recovery, tells a one-sided story about a number that moves.

Does AI search just reflect Google's rankings? No, and here the studies agree, which is itself worth noting: Ahrefs found only about 12% of AI-cited URLs rank in Google's top 10 for the same query overall, with Perplexity a clear outlier at 28.6% and ChatGPT, Gemini and Copilot all in the 6–8% range. About 80% of AI citations don't rank anywhere in Google's results for that query at all. SEO rank and AI citation are correlated but far from identical signals — a page can win one without the other.

What does each number mean, and not mean?

• Pew's 8% vs. 15% is a per-visit click rate on Google search results, for US adults, measured in March 2025. It does not mean "AI search cut clicks by 47% everywhere" — it's one platform's summary feature, one country, one month, and it measures clicks, not what the AI answer actually said about any brand.
• Seer's CTR figures come from Seer's own client base (42 organizations), mostly informational and educational queries. The authors explicitly say they "cannot definitively prove that citation causes higher CTR" — stronger brands may simply be more likely to get cited in the first place, so the 35%/91% citation advantage is a correlation, not a proven lever.
• Ahrefs' 12% overlap measures URL-level ranking overlap for long-tail queries specifically. It says nothing about short head-term queries or about branded searches, which may behave differently.
• Fractl's 81% still say "SEO" is about internal vocabulary, not about whether anyone is doing the work. A team can call it SEO and still be running AI-specific tactics; the number measures a label, not a practice.
• SEOFOMO's 92% monitoring is self-reported by survey respondents, with no independent verification of what "monitoring" means to each of them — it could range from a dashboard subscription to an occasional manual check.

Where two numbers look like they're measuring the same "zero-click" story but aren't, see Zero-Click in the AI Era for the longer breakdown of Pew vs. Ahrefs vs. Seer specifically.

What's changed since the last check?

Two trend lines moved in 2026, per the studies above. SEOFOMO's tracking-adoption figure rose from 78% (2025 survey) to 92% (2026 survey) among its own practitioner sample — a real increase, though still within one self-selected population, not a market-wide claim. And Fractl's ContentGrip-reported consumer trust figure dropped sharply: the share of consumers who find AI search more helpful than traditional search fell from 82% in 2025 to 54% in 2026, alongside AI search usage itself continuing to rise. Rising usage with falling trust is the single most interesting tension in this batch of numbers, and it's worth watching whether it holds in the next quarterly survey round.

Two adjacent posts round this out: which content types AI engines actually cite most digs into a separate, equally contested set of vendor studies, and first-party citation reports vs. third-party tools covers what Bing and Google now report directly, which is a different data source from any survey in this list. For the developments that landed after this round of studies, see this month's AI search briefing.

FAQ

What's the single most reliable number in this list? Pew Research's 8%-vs-15% click rate. It's a primary source, a disclosed sample (900 US adults, 68,879 searches), a stated date, and it's been independently cited by nearly every other outlet covering this topic without contradiction.

Why do tracking-adoption surveys disagree so much (22% vs. 92%)? Different populations and different disclosure standards. The higher figure comes from a self-selected sample of search professionals already engaged enough to answer a practitioner survey; the lower figure has no disclosed sample and reads as a broader market estimate from a single vendor's analysis.

Does a falling AI Overview CTR mean SEO is dead? No single CTR number supports that claim. Seer's own data shows the CTR floor partially recovering over time, and the authors stop short of claiming AI Overviews caused the decline rather than correlating with it. See is SEO dead because of AI search for the fuller case either way.

How often should these numbers get re-checked? Quarterly, at minimum, and immediately after any of the named publishers (Pew, Ahrefs, Seer, SEOFOMO) releases an update — several of these are running studies, not one-off reports.

What to do next

1. Before quoting any AI search statistic internally, trace it to this table or open the original source yourself — don't repeat a number you found in a second roundup.
2. When two numbers seem to conflict, check whether they're actually measuring the same thing (sample, date range, population) before picking a "winner."
3. Set a quarterly reminder to re-check the numbers you rely on most, since several of the sources above (Seer, SEOFOMO) update the same study over time.

ContextOwl sells one-time AI visibility reports. The sample report shows what a real, dated, sourced snapshot of one brand's AI visibility looks like.