Nobody's Leaving Google. But They Search Differently.

The "Google is dying" story is tempting. And wrong.

How many times in the past month have you heard that "Google is finished"? That AI is taking over? That SEO is dead? Probably more times than you'd like. And every time, someone on the client's team sends a Slack message with an alarmist article and asks: "What are we doing about this?"

Short answer: we're not panicking. The longer answer needs data. And the data tells a story far more interesting than any catchy LinkedIn headline.

A recent article on Search Engine Journal pieces together research from Similarweb, Bocconi University, Pew Research, and Google, reaching a conclusion we see every day in practice: people aren't leaving Google. But the way they search has fundamentally changed. And if you don't understand how, you risk optimizing for a behavior that no longer exists.

Three metrics moving in different directions

Let's start with the number that reassures everyone: 95% of ChatGPT users also use Google. The data comes from Similarweb, covering a nine-month period (September 2025 to May 2026). At first glance, everything looks fine. Google is still the giant in the room.

But here's what that number hides. A Bocconi University study, based on Comscore clickstream data covering 4,394 households, shows that those adopting ChatGPT reduce their Google searches by 9.4% immediately. After 20 weeks of use? The drop reaches 17%. These aren't users who leave. They're users who stay but search less.

And there's a third dimension, the most critical one for anyone who depends on organic traffic. An experiment by Wang et al. on 1,100 Chrome users showed that under AI Mode conditions, click-through rates to external websites drop by 18.8 percentage points. To put that in context: academic sites lose 32.8% of their referrals. Reference sites lose 26.5%. News sites, 13.4%.

So we have three metrics, each moving independently. Audience is stable. Search volume is declining. Click-through rates are falling even faster. If you only look at the first one, you relax. If you combine all three, you see a completely different picture. And if you're a marketer, that's exactly the picture you need to present to your client.

Not all search traffic bleeds equally

The useful part of this data is that it shows exactly where the shift is happening. It's not a uniform decline. It's a selective migration.

Informational queries ("how does X work", "what does Y mean", "why Z") are migrating to AI the fastest. And that makes perfect sense. When you want a quick, contextual, personalized answer, a chatbot is more efficient than 10 blue links followed by 10 cookie pop-ups.

But transactional and navigational queries stay on Google. Nobody asks ChatGPT for "directions to the nearest store" or "bank account login." Pew Research confirms this: 60% of users read AI summaries at the top of search results, but they continue using Google for concrete actions. Search Console's AI reports show the same split: Google sees AI traffic, but not the clicks that went missing.

For a business, this means something very specific. Your product pages, locations, and campaign landing pages will continue receiving traffic from Google. Probably at similar levels. But your blog posts, guides, and educational content? That's where the pressure builds. Not because people can't find you, but because they get the answer directly inside the AI interface, without ever clicking through to your site. And the 15.3% of AI Mode users who reported difficulty reaching specific websites? That's not a bug. That's the new default.

What to do (without throwing SEO in the trash)

The first wrong reaction we see from clients: "If AI is taking over informational queries, we'll stop investing in content." It's a tempting conclusion. And completely wrong. Your content is the source from which AI learns and formulates answers. ChatGPT reads you, even if it doesn't cite you. If you disappear from the content landscape, you disappear from AI answers too.

Here's what we do at difrnt. for clients in this context:

We monitor referral traffic by content category, not just as a total. If product pages are stable but the blog is declining, it's not a technical SEO problem. It's a user behavior shift. And the answer isn't to write more content, but to write differently: with your own perspective, with data, with direct answers to specific questions.

We strengthen content that answers specific questions. Research shows that AI engines cite sources that provide clear answers, with concrete data and original perspective. Vague, generic paragraphs don't get cited. Well-written FAQs with precise answers do. Studies with proprietary data do. Argued opinions do.

We build visibility on two fronts simultaneously. Classic SEO for transactional and navigational queries, where Google remains dominant. GEO (Generative Engine Optimization) for informational queries migrating to AI. It's not either/or. It's both/and. The two strategies feed each other: good GEO content also improves SEO, and vice versa.

We track AI reports in Search Console. Google recently launched global reports showing how your site appears in AI Overviews. With 2.5 billion monthly active users on AI Overviews, this data is no longer optional. It's essential.

The real story isn't "Google vs AI." It's "Google and AI."

Pew Research numbers show that 49% of U.S. adults have used a chatbot at least once, and 42% use them to find information. Google reports 2.5 billion monthly active AI Overview users and 1 billion AI Mode users in the first year. The numbers are real. The change is real. But it's not the change that alarmist headlines describe.

The "Google is dying" narrative is just as inaccurate as "everything is fine." Reality sits somewhere in the middle, and that's where your strategy needs to be too. The data from Bocconi, Pew, and Google's own reports all point in the same direction: a gradual, selective behavioral shift, not a sudden platform death. People aren't leaving Google. But they use Google and AI for different things. Those who understand this split and adapt their content for both channels have a clear advantage over the competition.

Those who ignore the data and cling to a single narrative (whichever one it is) lose ground. Not because they chose wrong, but because they chose not to look.