Is AI Search Overhyped? What 260 Billion Clickstream Sessions Show (2026 Data)

Is AI search overhyped? The data says yes, overwhelmingly. 70% of websites get less than 2% of their traffic from ChatGPT. Thirty-eight percent get under half a percent. That's not a revolution. That's a rounding error.

These numbers come from a LinkedIn poll by Lily Ray: 1,316 responses from SEOs and marketers who actually track this stuff. Not predictions. Not vibes. Real referral data from real sites.

Meanwhile, your LinkedIn feed is full of people screaming that Google is dead, AI search is eating everything, and you need to pivot your entire strategy right now or get left behind.

Someone's wrong. Let's look at what the data actually says.

The hype narrative vs. reality

The story goes like this: ChatGPT and Perplexity are replacing Google. Traditional search is dying. If you're not optimizing for AI answers, you're invisible.

It's a compelling story. It's also mostly wrong.

Google maintained 90%+ search market share through 2025 and into 2026. That number barely moved. Google Gemini's monthly active users actually surged 30% in Q4 2025. Google isn't losing to AI search, it's absorbing it.

ChatGPT mobile usage? It dropped 22% by January 2026 after peaking in September 2025. The initial novelty wore off. People tried it, used it for a while, and many went back to their old habits.

This doesn't mean AI search is irrelevant. It means the narrative is wildly disconnected from the numbers. For a deeper look at what metrics actually tell the truth, see our breakdown of AEO metrics that matter in 2026.

What 260 billion clickstream sessions show

Eli Schwartz at ProductLedSEO did something most AI search commentators haven't: he looked at actual data. Not a sample of 500 users. Not a survey about intent. He analyzed 260 billion+ clickstream sessions.

His finding kills the zero-sum narrative: users run parallel journeys between Google and ChatGPT. They're not replacing one with the other. They're using both.

Someone researching project management software might ask ChatGPT for a comparison, then Google specific products to read reviews, then go back to ChatGPT to ask follow-up questions, then Google pricing pages.

There's no cannibalization happening. Google searches aren't declining because people use ChatGPT. The two channels serve different functions in the same buyer journey.

This is a fundamentally different picture than "AI search is replacing traditional search." It's additive, not substitutive.

The traffic truth: most sites get almost nothing from AI search

Back to Lily Ray's poll. Here's the breakdown of what sites actually receive from ChatGPT referral traffic:

• 38% of sites get 0.0–0.5% of traffic from ChatGPT
• 70% of sites get less than 2% total
• A tiny minority see meaningful volume

If you're rebuilding your content strategy around a channel that sends less than 2% of your traffic, you need to seriously question your priorities.

As Lily Ray put it: "Every single URL surfaced in an LLM response is pulled from a live search index." The content that shows up in AI answers is the same content that ranks well in traditional search. You don't need a separate "AI search strategy." You need good content that's authoritative enough to rank. Period.

The sites that do get meaningful AI search traffic? They already had strong domain authority, comprehensive content, and solid SEO fundamentals. AI search didn't create new winners. It rewarded the same winners.

When AI search actually matters

None of this means you should ignore AI search entirely. There are specific scenarios where it's genuinely changing behavior:

Research queries

When someone's trying to understand a category ("what's the difference between CDP and DMP" or "how does intent data work"), they're increasingly asking an AI first. If your brand shows up in that answer, you've planted a seed early in the journey.

Complex decisions

B2B buyers comparing multiple vendors sometimes prefer an AI-synthesized comparison over reading six separate G2 reviews. If your product gets mentioned favorably in these responses, it matters.

Technical how-tos

Developers and technical buyers have adopted AI search faster than any other segment. If you sell to engineers, AI search visibility is more than a rounding error for you.

Long-tail conversational queries

Questions phrased naturally ("what's the best way to handle inbound leads when my SDR team is overwhelmed") tend to go to AI. These are high-intent, specific queries where a mention carries weight.

The pattern: AI search matters most for early-stage research and complex questions. It matters least for transactional and navigational queries, which still overwhelmingly go through Google.

Where to actually invest your time

If you're a marketing leader deciding where to allocate budget and attention, here's the honest answer:

1. Don't abandon SEO. Google still sends 90%+ of search traffic. The fundamentals haven't changed. Your organic search program is still your biggest traffic driver by a massive margin.

2. Make your content genuinely authoritative. AI models pull from the same index as Google. Content that ranks well in search will get cited in AI answers. You don't need to optimize separately. You need to be the best source on your topics. We cover how to build that authority in E-E-A-T for AI search.

3. Monitor, don't overreact. Track your AI search referral traffic. Look at what queries send visitors from ChatGPT and Perplexity. If it's growing, pay attention. If it's under 1%, don't restructure your team around it.

4. Focus on conversion, not just visibility. The real question isn't whether someone finds you through AI search or Google. It's what happens when they land on your site. A visitor from ChatGPT who hits a generic landing page and bounces is worth nothing regardless of the referral source.

5. Invest in the experience layer. Whether traffic comes from Google, ChatGPT, Perplexity, or a LinkedIn post, what matters is what greets them. A static form that says "talk to sales" loses to an intelligent experience that actually helps the buyer in real time.

The real shift isn't where traffic comes from

Here's what the AI search hype crowd gets backwards: the channel doesn't matter nearly as much as the experience.

Buyers are getting smarter. They've used ChatGPT. They've had conversations with AI that actually understood their questions. And now they show up on your website and get... a form. A chatbot that asks "what department are you looking for?" A 48-hour response time on a demo request.

The bar for buyer experience has permanently changed. Not because AI search sends you traffic (it mostly doesn't) but because AI has reset expectations for how interactions should work. The rise of agentic commerce is accelerating this shift even further.

That's where the real investment should go. Not into chasing a channel that sends 70% of sites less than 2% of traffic. Into making sure that when buyers arrive, from any source, they get an experience that matches what they've come to expect from AI-native interactions.

An AI sales agent that can have a real conversation, qualify interest, answer product questions, and route to the right person instantly isn't an "AI search strategy." It's a "every-channel strategy." And it works whether the visitor came from Google, ChatGPT, a paid ad, or a referral link.

Stop chasing the AI search hype. Start fixing the experience gap. That's where the actual revenue is.