Users Haven't Stopped Searching. They've Stopped Clicking.

And what that means for every enterprise website that still thinks of itself as a destination.

‍

There's a stat that should end a lot of comfortable conversations about web strategy.

For every 1,000 searches performed on Google in the United States today, only 360 clicks reach an independent website. The open web receives 36% of the traffic that search generates - and that number keeps falling.

This isn't a traffic problem. It's a structural one.

The anatomy of the shift

The zero-click era didn't arrive overnight. Google introduced Knowledge Panels in 2012, Featured Snippets in 2014. Each moved a small slice of resolution from the destination website to the search results page itself. For most of that decade, the effect was easy to dismiss.

Then AI Overviews launched broadly in 2024, and the trajectory steepened.

By March 2025, 27.2% of all U.S. Google searches ended without a single click — up from 24.4% twelve months earlier. Organic click-through rates fell to 40.3% in the U.S. and 43.5% in the EU and UK over the same period. On mobile — where most searches now happen — 77% of queries end without a website visit. Searches that trigger an AI Overview show an average zero-click rate of 83%.

The number that crystallizes the situation: 88.1% of queries that trigger AI Overviews are informational — the exact content type that has historically driven B2B traffic, demo requests, and pipeline.

Google's search market share dropped below 90% for the first time since 2015. The query volume is still there. The clicks are not.

‍

Users are still searching. The resolution has just moved.

Here's the misread that most strategy documents make: they treat zero-click as abandonment. It isn't. Users aren't less curious or less intent-driven. They're getting answers faster, in the interface where they searched. The search engine has become the answer engine.

This matters because it changes where the competitive battle is being fought.

For a decade, the question was: Can we rank for this keyword? That question is still real, but it's no longer sufficient. The new question is: Are we the source that AI systems cite when they construct the answer?

Those two questions don't always have the same answer.

‍

The B2B implication most teams aren't discussing

Bain research has established that 80–90% of B2B buyers have a vendor shortlist formed before they begin formal research — and 90% of them ultimately purchase from that initial list. Being on the day-one shortlist is, statistically, the entire ballgame.

Zero-click search accelerates this dynamic in a specific way: it moves the shortlist formation upstream, into the AI interface, before the buyer ever lands on anyone's website. A buyer who asks ChatGPT or Perplexity "what platforms handle [category]" and receives a synthesized answer with three vendor names has already narrowed their consideration set — without a single click occurring anywhere.

This is what Bain described as the commercial consequence: organic web traffic has declined an estimated 15–25% across many sectors, and 73% of B2B websites experienced significant traffic loss between 2024 and 2025.

The brands losing aren't necessarily producing worse content. They're producing content that AI systems can't read, parse, or cite.

‍

The signal hidden in the conversion data

There's a counterintuitive finding in the data that tends to get buried beneath the zero-click alarm: the visitors who do click through from AI-generated answers convert dramatically better than traditional organic visitors.

Ahrefs analyzed their own traffic and found that AI search visitors represented 0.5% of total sessions but drove 12.1% of all signups — a 23x higher conversion rate than traditional organic search. Semrush's cross-industry analysis puts the average AI-referred conversion rate at 4.4x higher than standard organic. Microsoft's data on Copilot-powered purchase journeys shows them to be 33% shorter and 76% more likely to reach conversion.

The interpretation isn't complicated: when an AI system cites your brand in response to a buyer's query, it has already done the comparative work. The click that follows isn't curiosity — it's intent. The user arrives pre-qualified.

Volume is still skewed heavily toward traditional organic. But per-visitor economics have inverted. The question facing every enterprise marketing and product team is whether they're optimized for the traffic that used to come, or for the traffic that now converts.

‍

What this means for the website itself

This is where the conversation has to get more specific — and more uncomfortable.

Most enterprise websites are built around a model that no longer describes how buyers arrive. They're designed as persuasion surfaces: structured to move someone from vague interest to conversion over multiple sessions. That model assumes the website is where discovery happens.

It isn't anymore. Discovery happens in the AI interface. The website is now a confirmation surface — the place a pre-qualified, often already-decided buyer comes to verify and transact.

That shift has concrete implications for how a site needs to be structured internally.

When an AI system constructs a response, it doesn't rank web pages — it parses meaning. It reads semantic relationships between concepts, entities, and answers. A site that is structured for keyword density and PageRank signals looks entirely different to an LLM than a site that is structured for semantic clarity: well-defined entities, explicit relationships between products and use cases, answers that are self-contained and citable.

The technical term for this gap is the difference between syntactic search (matching strings) and semantic search (matching meaning). Most enterprise websites are built for the former. AI systems operate exclusively on the latter.

The sites being cited by AI Overviews, ChatGPT, and Perplexity share a structural characteristic: they answer questions in forms that can be directly extracted and synthesized. Their content is organized around what buyers need to understand, not around what the vendor wants to communicate. Their internal architecture allows an AI system to navigate from a concept to an authoritative answer without ambiguity.

‍

The practical upshot

The era of the website as a traffic destination is ending. Not because users have stopped searching — search volume has not declined — but because the resolution layer has moved from the website to the AI interface that precedes it.

The brands that will hold and grow commercial presence in this environment share a few characteristics. They're being cited in AI-generated answers because their content is semantically structured for extraction, not just for reading. Their websites function as callable knowledge bases, not brochure surfaces. They're not just optimized for what search engines index — they're optimized for what AI systems can understand and represent.

Users haven't stopped searching. They've stopped clicking — except when they're ready to act.

The only websites that will matter in that moment are the ones that were findable, citable, and legible to the systems that shaped the decision before the click ever happened.

‍

Sources: SparkToro (2024); Search Engine Land / Similarweb (March 2025 State of Search); Bain & Company, "Goodbye Clicks, Hello AI" (February 2025); Bain & Company, "Losing Control: How Zero-Click Search Affects B2B Marketers" (September 2025); Ahrefs, AI Search Traffic Conversions (June 2025); Semrush Zero-Click Study (2025); Microsoft Advertising, Copilot Conversion Data (August 2025); BrightEdge (February 2026).

‍