AI Search Visibility Is Not the Same as Traffic

On July 17, 2026, Nick Fox, Google's SVP of Knowledge & Information, posted on X that Google's AI search features now send "billions of clicks to websites every week." The claim moved through trade coverage inside a day — ppc.land, Search Engine Land, Search Engine Roundtable, and Search Engine Journal all ran it. Read next to the click-through rate data published the same season, the story gets more complicated. Ahrefs, and a separate randomized study of Chrome users, both show organic CTR falling wherever an AI Overview appears on the results page. AI search visibility — showing up inside an AI-generated answer — is no longer the same thing as traffic to your site. Teams still scoring generative engine optimization work by clicks are scoring the wrong game.

What Google actually claimed about AI search visibility

Nick Fox's post was a scale claim, not a rate claim. "Billions of clicks" sounds enormous, and at Google's volume it probably is technically true — the company has been building toward this framing for a while, and the timing was not an accident. It landed the same week independent CTR research was making the rounds in trade press, and it gave publishers and SEOs a bigger, friendlier number to anchor on instead. Part of the underlying plumbing is Preferred Sources, the system that decides which pages AI Overviews and AI Mode pull from and cite when they generate an answer. Per the same coverage cycle, Preferred Sources expanded into AI Overviews on May 27, 2026, and now spans more than 345,000 unique sources. A bigger pool of eligible sources plausibly means a bigger raw number of outbound clicks in absolute terms, even while the odds of any single ranking page getting clicked keep falling year over year.

That distinction matters because a total is not a rate, and Google's statement gave the industry a total with no denominator. Billions of clicks across a search engine that handles trillions of queries a year tells you almost nothing about whether your page, ranking where it ranks, is more or less likely to get clicked than it was two years ago. It is the kind of number a company controlling the narrative reaches for when the rate story is unfavorable. That is not an accusation of dishonesty — Fox never claimed CTR was flat or rising, and Google has never disputed Ahrefs' methodology publicly. It is a reminder to read the claim for exactly what it says, and no further, before you let it change how you brief a client or a board on what AI search visibility is actually doing to organic performance this quarter.

“Google's AI search features are sending billions of clicks to websites every week. — Nick Fox, SVP of Knowledge & Information, Google, in a post on X, July 17, 2026”

What the independent data shows about AI search visibility

Ahrefs has been tracking this specific question since 2025, and the trend line only moves one direction. In an April 2025 study, Ahrefs found that pages ranking #1 saw a 34.5% lower CTR when an AI Overview appeared on the results page. Ahrefs revisited the question in February 2026 using a fresh analysis of 300,000 keywords pulled from Google Search Console data, comparing December 2023 to December 2025. The updated figure: a 58% CTR reduction for #1 rankings when an AI Overview is present. The decline did not level off. It got worse, and it got worse fast.

Ahrefs' numbers come from Google Search Console data, which means they are correlational — real, large-sample, and directionally trustworthy, but not proof that the AI Overview itself caused every point of decline rather than some other concurrent shift in how people search over two years. The 39.8% figure reported in the same July 17 coverage cycle closes that gap. It comes from a randomized study of 1,065 Chrome users, structured to isolate the AI Overview as the variable and measure its causal effect on whether users clicked an outbound organic result, rather than just observing what happened in aggregate across millions of unrelated queries. A controlled study landing in roughly the same range as a two-year correlational trend is the strongest version of confirmation you can ask for in this field — two different methods, two different data sources, one answer. Anyone still arguing AI Overviews are CTR-neutral has to explain both studies away, not just one.

Why clicks are the wrong scoreboard for AI search visibility

For twenty years, a click was the unit everyone agreed on. Rank higher, get more clicks, get more revenue — the chain was clean enough that entire reporting stacks got built around it. That chain breaks the moment an AI Overview answers the query directly and the user never scrolls to the results. Your page can rank #1, satisfy the query completely, and still register zero clicks, because the AI Overview already gave the user what they needed using your content as raw material. Google and ChatGPT citation patterns have already started shifting toward the bottom of the page as engines restructure how — and whether — they hand off to the source at all. Treating that page as a failure because CTR dropped misreads what happened. It got cited. It just did not get clicked.

The clearest evidence that citation and organic ranking have split into separate systems comes from Ahrefs' ongoing "1 Billion Data Points" research led by Tim Soulo: 28.3% of ChatGPT's most-cited pages have zero Google organic visibility at all. Nearly three in ten pages that ChatGPT trusts enough to cite would not show up if you ran the same query on Google. That is not a rounding error — it means a meaningful share of what actually earns AI citations is invisible to the tools most teams still use to measure search performance. If your reporting stack only reads Google Search Console, you are structurally blind to over a quarter of the pages doing the work that matters most for generative engine optimization.

What to track instead

None of this means clicks stop mattering. It means clicks stop being sufficient. A page that gets cited in an AI Overview or a ChatGPT answer without a click still did something valuable — it put your brand name, your framing, and your specific claims in front of a buyer at the exact moment they were evaluating options. That is closer to a branded impression than a wasted opportunity, and it needs its own line item in reporting, separate from CTR. The teams getting this right in 2026 are running two parallel scoreboards: one for organic click volume and conversion, and one for citation frequency, citation share of voice, and which specific pages engines pull from when they answer questions in your category.

Start narrow. Pull your twenty highest-traffic pages from Google Search Console and cross them against whatever citation tracking you have running for ChatGPT, Perplexity, and Google AI Mode. Sort by the gap between citation frequency and click volume. The pages with high citations and falling clicks are not underperforming — they are doing exactly what an AI-cited page does now, and your reporting needs to say so in the next board deck instead of flagging them as a CTR problem. If your mention-rate dashboard is still lying to you by treating every mention as equal, fix that before you fix anything else. And if leadership asks why organic traffic looks flat while AI citations climb, that is exactly the conversation attributing pipeline to organic and AI work is built to have — with numbers, not vibes. Pair that reporting rebuild with a reporting and analytics setup that actually separates the two metrics instead of averaging them into something meaningless. Do this before your next quarterly review, not after someone in the room quotes Google's billions-of-clicks number back at you and asks why your traffic chart disagrees. You want the citation data on the same slide, with the same rigor, ready to answer that exact question instead of getting caught flat-footed by a headline number with no denominator.

See where you are cited today

A free snapshot audit of your rankings and AI citations before we ever talk.