05/16/2026
According to SparkToro/Datos, 2024-2025, 60% of all Google searches end without a click. Since the rollout of AI Overviews, the tech outlet Digital Trends has seen a 97% drop in traffic from Google. Even Google’s AI Mode, which replaces the traditional search page with a conversational interface, sees 93% of queries end without a single outbound click.
Let that sink in. Ninety-three percent. The user asks Google a question. Google’s AI Mode answers it. The search is over.
This three-act play is what the SEO industry calls “zero click search.” While most of the surrounding buzz has been predictably focused on traffic loss, ranking strategy, and optimizing for citations in AI-generated responses, there may be a more interesting question lurking beneath: what happens to authority and credibility when Google stops ranking sources and starts synthesizing them?
If AI systems are trained to surface the most credible, expert, and trustworthy information, then the question isn’t just who loses traffic. It’s who gets to be considered as an authority in the first place.
The answer, I think, looks very different from what the doomsday narrative suggests.
To understand what’s really at stake, it helps to understand how we got here.
Google’s AI Overviews—the generated summaries that now appear above traditional search results—began rolling out broadly in 2024. By March 2025, they were triggering on roughly 13% of all queries—double the rate from just two months earlier. As of late 2025, AI Overviews appear on nearly every informational search.
The expansion wasn’t gradual. It was a takeover.
AI Mode, which was launched in the U.S. by Google in 2025, takes it a step further. Whereas AI Overviews sit above a traditional results page, AI mode replaces the results page entirely. There are no blue links below the fold to scroll to. There’s no ranked list of sources to evaluate. The interface is conversational—the user asks, the AI responds. No click required.
This is a meaningful structural shift. The traditional Google search results page was built on a specific logic: Google surfaces the most relevant sources, the user evaluates them and clicks what’s actually useful. That feedback loop is what made search rankings matter. More clicks lead to more visibility and a higher spot on Page 1.
Consider clicks as a kind of currency. They’re what made online publishing viable—every click is a potential reader, subscriber, or customer. AI Mode, however, short-circuits that feedback loop. The synthesis happens before the user ever sees a source. Yes, Google points to its sources, but by the time a website can earn a click, the question has already been answered.
The traffic consequences are real and measurable. According to Ahrefs research from December 2025, the presence of an AI Overview correlates with a 58% reduction in click-through-rates (CTR) for top-ranking pages.
An important but less talked-about aspect of this trend is directionality. The 58% CTR drop Ahrefs documented was 34.5% in April of the same year. The zero-click share of searches, which hit 60% in 2024, now reaches 83% on queries that trigger AI Overviews—a trajectory that only steepens as AI Mode expands. Gartner projects traditional search volume will drop by 25% in 2026. The trajectory is not ambiguous.
In response, the SEO industry has largely been tactical: optimize for AI citation, build structured data, and earn mentions in the Overview rather than clicks below it. This strategy isn’t wrong—but it treats the symptom rather than examining what’s actually changing.
The deeper shift isn’t about traffic—it’s about what happens to the concept of authority when a single AI system decides, at scale, whose expertise is worth synthesizing, and whose isn’t.
The doomsday narrative exists for a reason. It isn’t wrong—but it is, I think, incomplete.
The case against zero-click is straightforward: if Google stops sending traffic, the economic model that sustains online publishing collapses. Display advertising depends on pageviews. Affiliate revenue depends on clicks. Newsletter growth depends on discovery. If 93% of searches in AI Mode end without anyone visiting a website, the entire top-of-funnel for independent publishers effectively disappears.
This isn’t hypothetical—the traffic losses are already documented and, in some cases, catastrophic. Publications that spent years building search authority by producing careful, well-sourced, ranking content have watched that equity evaporate. Not because their content got worse, but because the game changed underneath their feet.
There’s a legitimate argument that this represents a kind of theft. Publishers created the content that trained the AI systems that now answer questions in their place. Google built its search dominance on the back of the open web, and AI Mode is, in some sense, the final step in that extraction—synthesizing the web’s knowledge while severing the traffic relationship that made knowledge production economically rational. A growing coalition of publishers has already filed suit. The legal question of whether AI synthesis constitutes fair use is genuinely unsettled.
The structural concern is also real. If traffic concentrates entirely around whatever sources AI systems choose to cite, small and independent publishers face a new kind of gatekeeping—one that’s less transparent than search rankings (and harder to influence). With traditional SEO, the rules were understandable. You could study the algorithm, adapt to it, and compete. With AI citations, the selection criteria are opaque. Being authoritative in the eyes of an AI system may require resources—institutional affiliation, publishing history, clean networks—that independent writers simply don’t have.
And therein lies the deeper cultural concern. Search, for all of its flaws, preserved a kind of plurality. A well-written independent blog could outrank a major publication on the right query. A niche expert with a small audience could surface search results in their domain. The ranked list of blue links was imperfect, but it was at least a list you could see—multiple sources, multiple perspectives, multiple chances for a reader to encounter something unexpected.
AI synthesis collapses that plurality into a single answer. However well-intentioned, it’s a significant concentration of epistemic power in a system none of us fully understand (nor elected).
These are serious concerns, and they deserve to be taken seriously—not dismissed as technophobia or nostalgia for a web that was already imperfect. But here’s what the doomsday narrative assumes—and it’s an assumption that I think deserves examination—that the web we’re losing was worth saving in its current form.
The web we’re losing was largely built on a lie. Not a malicious one—but a structural distortion.
The content economy that Google created rewarded volume, keyword density, and backlink acquisition over general expertise. It produced a web bloated with articles written not to inform readers but to satisfy algorithms—2,000-word pieces that answered simple questions, listicles optimized for featured snippets, “comprehensive guides,” authored by writers with no expertise in their subject.
The lie was that ranking equaled authority. It didn’t. It equaled optimization.
Consider what actually populated Page 1 of Google for most informational queries over the last decade. Largely: sponsored links, content farms, media companies with large SEO teams, and affiliate sites with significant domain authority accumulated via link building rather than genuine knowledge. The independent expert—the academic, the practitioner, the deeply niche writer—was frequently outranked by a faster, better-optimized competitor with shallower knowledge and a larger content budget.
This is the version of the web that the doomsday narrative mourns. Here’s the contrarian proposition: if AI systems are genuinely trained to surface credible, expert, and trustworthy sources, then the zero-click future may actually disadvantage the wrong people. The content mills, the affiliate farms, the keyword-stuffed “ultimate guides” are precisely the sources AI systems are designed to look past. They’re optimized for old ranking logic, not for the epistemic criteria that determine AI citation.
What AI systems appear to reward looks more like: demonstrated expertise in a specific domain, original analysis and perspective, credentialed authorship, consistency of voice and argument over time, and content that answers questions other sources haven’t addressed as precisely or well. Remember that English composition gen ed you took in undergrad? Now, it’s all relevant.
The implication is significant. In the traditional SEO era, a credentialed expert with a small publication competed against organizations with dedicated teams, link-building budgets, and keyword research infrastructure. The expert often lost—not because their knowledge was inferior, but because their optimization was. In a zero-click world where AI systems handle synthesis, that optimization advantage largely disappears. What remains is the question of whose knowledge is actually worth citing.
It doesn’t make the transition painless or fair. Traffic losses are real, and the economic models that sustained even good publishers are genuinely under pressure. But the destination—a web in which authority is determined more by actual expertise than SEO budget—is not obviously worse than where we’ve been.
Seer Interactive Research found that brands and sources cited inside AI Overviews see a 35% higher CTR than sources appearing below an Overview that aren’t cited. AI-referred traffic converts at substantially higher rates than standard organic traffic. The readers who click through from an AI citation are not casual browsers. They’ve already received a synthesized answer and want more—which means they’re arriving with higher intent, deeper interest, and more reason to stay.
In other words, fewer clicks but better ones. For a publication built on genuine expertise and a distinctive voice, that’s a trade worth understanding.
So, what does this actually mean for small and independent publications?
The honest answer is it depends entirely on what kind of small publication you are.
If your publication was primarily built on search traffic—high-volume informational content, broad keyword targeting, content designed to rank rather than argue—the zero-click future is genuinely threatening. That model depended on Google sending readers to you before they knew you existed. If Google stops doing that, the discovery mechanism disappears.
But if your publication is built on specific, demonstrable experience—a distinctive voice, a consistent argument, a defined perspective on a defined subject—the calculus looks different. AI systems cite sources for a reason. They’re not pulling randomly from the web; they’re pulling from sources that have established they know what they’re talking about through credible, consistent content.
This is a meaningful distinction that often gets glossed over. The doomsday narrative treats “small publisher” as a monolithic category. Not to state the obvious, but: a content farm producing 50 SEO articles a month and a credentialed writer publishing two or three deeply considered essays a week are not facing the same future.
The content farm is facing genuine extinction. The serious writer is facing something more like a renegotiation: publish less, but with more depth and specificity. Build authority in a defined domain rather than chasing volume across broad topics. Write the kind of content that answers questions other sources haven’t answered as precisely—the kind of content an AI system has a reason to cite because no adequate substitute exists. Develop a direct relationship with readers through email and subscription, so that discovery doesn’t depend entirely on search in the first place.
None of this is new advice. It’s what serious editors and publishers have always said. What’s new is that the SEO era briefly made it possible to ignore that advice—to substitute optimization for expertise and still get traffic. That window is closing.
There’s also a long-term possibility (though speculative): as AI systems mature and citations become more important than ranking, the publications most likely to be consistently cited are those with the clearest editorial identity and the most recognizable expertise. Brand recognition may matter more in an AI-moderated web than it ever did in a search-ranked one. AI systems learn from patterns. A publication that consistently produces credible, specific, well-argued content in a defined domain is a pattern worth learning.
That’s not a guarantee—but it’s a better bet than optimizing for a ranking system that’s being dismantled in real time.
There’s a version of this story that ends badly.
Traffic collapses, economic models fail, independent publishing becomes economically irrational, and the open web consolidates into a handful of AI systems trained on content they no longer reward.
That version is plausible. It might even be likely for publishers who built their entire model on search discovery and never developed a reason for readers to seek them out directly.
But there’s another version—one in which zero-click transition functions less like an extinction event and more like a filter, by clearing out the optimization-first content while creating more room for the kind of writing that was always harder to game. Writing that shows evidence of knowledge. Writing that has perspective. Writing that a reader would seek out by name rather than stumble across in a search query.
That version requires something the SEO era never required: that you are who you say you are. Not optimized for authority, not structured to “appear” credible. Actually expert, consistent, and worth reading at length.
The last click may be coming. But what it filters out, and what it leaves standing, is still being written.