AI assistants aren’t killing search engines. Not yet, anyway. A new study from Search Engine Land analyzing global search data finds that ChatGPT, Perplexity, and other AI tools still account for a tiny fraction of total search volume compared to Google. But the trajectory matters more than the current snapshot, and that trajectory should make anyone in the search industry pay close attention.
The study, conducted by Rand Fishkin and SparkToro using data from Datos (a Semrush company), examined web traffic patterns across the US and globally to measure how much search activity AI assistants are actually capturing. The headline finding: as of late 2024, AI-powered tools represent roughly 1-2% of Google’s total search volume in the US. That’s it.
One to two percent. Against a backdrop of breathless predictions about the death of traditional search, those numbers are almost anticlimactic.
But here’s where it gets more interesting. The data shows that while AI assistants haven’t yet made a serious dent in Google’s dominance, their growth rate is steep. ChatGPT’s web traffic has been climbing consistently, and Perplexity AI — which positions itself more explicitly as a search replacement — has seen rapid adoption among early adopters and knowledge workers. The growth curves are real even if the absolute numbers remain modest. And Google’s own internal concerns, evidenced by the rushed deployment of AI Overviews (formerly Search Generative Experience), suggest the company doesn’t view this as a trivial competitive threat.
So what does this mean for SEO professionals and marketers?
First, don’t panic. Google still processes an estimated 8.5 billion searches per day, according to data frequently cited by Internet Live Stats. The installed base is enormous, the habit deeply ingrained. People don’t abandon default behaviors quickly, especially when those behaviors are embedded in browsers, phones, and operating systems. Google’s distribution advantages remain staggering — it’s the default search engine on Chrome, Safari, Android, and most other platforms that matter.
Second, the composition of queries shifting to AI assistants matters more than the raw volume. Fishkin’s analysis suggests that informational queries — the kind where someone wants a quick, synthesized answer rather than a list of links — are disproportionately moving toward AI tools. These are precisely the queries that have historically driven top-of-funnel traffic for publishers and content marketers. If you’re a site that depends on “what is” or “how to” traffic, the erosion may be hitting you harder than aggregate numbers suggest.
This is the real story. Not that AI is replacing search wholesale, but that it’s selectively siphoning off specific query types that many businesses have built their content strategies around.
The study also highlights something often overlooked in the AI-versus-search discourse: most people still don’t know what Perplexity is. Consumer awareness of AI search alternatives remains low outside of tech-savvy demographics. ChatGPT has broader name recognition, but most users still treat it as a chatbot or writing tool rather than a search replacement. The mental model hasn’t shifted for mainstream users. That could change — these things often tip faster than expected once they reach a critical threshold — but we’re not there yet.
Google’s response has been telling. The company rolled out AI Overviews across its main search results in May 2024, essentially embedding AI-generated summaries directly into the search experience. The move was defensive, designed to keep users within Google’s orbit even as AI-generated answers become expected. Early reception was rocky — AI Overviews generated some embarrassingly wrong answers that went viral — but Google has continued iterating. As Google’s own blog detailed, the company views AI integration as fundamental to search’s future, not optional.
This creates a paradox for publishers. Google adding AI summaries to search results means fewer clicks to external websites, even when users stay on Google. So whether traffic goes to ChatGPT or stays on Google but gets answered by an AI Overview, the outcome for content creators is similar: fewer visits.
Fewer visits. Same amount of effort producing content. That math doesn’t work long-term.
Fishkin, who has been tracking these trends through SparkToro’s research, has been vocal about what he calls the “zero-click” problem intensifying. His data has consistently shown that a growing percentage of Google searches end without the user clicking any result — a trend that predates AI but is accelerating because of it. The AI assistant competition is in some ways a secondary concern compared to Google’s own cannibalization of organic click-through rates.
For advertisers, the implications are more nuanced. Search ads still work because they capture high-intent commercial queries — the kind where someone is ready to buy, compare, or sign up. AI assistants haven’t meaningfully disrupted this category yet, partly because their monetization models are still immature and partly because commercial queries require trust signals that AI tools haven’t fully established. When someone searches for “best CRM for small business,” they still largely end up on Google clicking ads or organic results. That behavior has proven sticky.
But the advertising moat isn’t permanent either. Perplexity has started experimenting with sponsored answers, and OpenAI is reportedly exploring ad-supported models for ChatGPT. Microsoft has already integrated ads into Bing’s AI-powered chat experience. The pieces are moving into place for AI assistants to compete not just for informational queries but eventually for commercial ones too.
The international picture adds another dimension. Google’s dominance varies significantly by market. In regions where Google holds 95%+ market share, AI alternatives face an even steeper climb. But in markets where alternatives like Baidu, Yandex, or Naver already fragment search behavior, AI tools may find adoption easier because users are already accustomed to switching between platforms. The Search Engine Land study focuses primarily on US data, so global extrapolation requires caution.
What should practitioners actually do with this information? A few things. Diversify traffic sources — if your business depends entirely on Google organic traffic, you’re increasingly exposed regardless of whether the threat comes from AI assistants or Google’s own AI features. Build direct audience relationships through email, communities, and owned channels. And start paying attention to how your content appears (or doesn’t) in AI-generated responses, because that’s an emerging distribution channel whether you optimize for it or not.
The data is clear: AI assistants aren’t replacing Google search today. But they’re growing fast, reshaping user expectations, and forcing Google into defensive moves that themselves reduce the value of organic search for publishers. The disruption isn’t a single dramatic event. It’s a slow compression happening from multiple directions simultaneously.
And slow compressions, historically, are the ones that catch industries off guard.
WebProNews is an iEntry Publication