What A New AI Mode Study Actually Tells Us (And What It Doesn't)

A new usability study on Google's AI Mode just landed, and if you're in the C-suite trying to figure out what this means for your business, I've got good news and bad news.

The good news? We finally have real behavioral data on how people actually use AI Mode—not just click data or speculation, but actual screen recordings and think-aloud sessions from 37 participants completing 250 search tasks.

The bad news? The findings are... let's just say they're not exactly comforting if you've built your customer acquisition strategy around traditional search traffic.

Let me walk you through what Kevin Indig and Eric Van Buskirk's team at Clickstream Solutions actually found in their study published on Search Engine Journal, and more importantly, what you should be doing about it right now.

The Core Finding That Changes Everything

Here's the headline number: the median number of external clicks per task was zero.

Not declining. Not reduced. Zero.

In roughly three-quarters of the total user sessions, users never left the AI Mode pane, and 88% of users' first interactions were with the AI-generated text. People spent between 52-77 seconds per task actually engaging with the AI Mode output—reading, scrolling, making decisions.

Then they just... left. Without clicking your website.

The study classified over half the tasks as "skimmed quickly"—users glance at the AI summary, form an opinion, and move on. The study confirms that participants often shared out loud that they'll "go to the seller's page," or "find the product on Amazon/ebay" for product searches, but even those intentions didn't always translate to clicks within the AI Mode environment.

But Let's Talk About What This Actually Means

Before we go too far down the rabbit hole, let's add some context here.

First, AI Mode usage in the US currently sits at just over 1% according to Similarweb data. We're not talking about mass adoption yet. Google has been crystal clear that AI Mode is a "proving ground" for what search might become—Sundar Pichai himself said on the Lex Fridman podcast that as features work, they'll migrate them to the main page. But that's aspirational, not current reality.

Second, and this is crucial: Google makes over $200 billion annually from search ads. They're not going to scale something that generates zero clicks without figuring out how to monetize it first. The question isn't whether AI Mode will take over—it's how Google threads the needle between user experience and their business model.

My take? This evolution takes years, not quarters. And the companies that win won't be the ones who panic and overhaul everything overnight.

The B2B Implications Nobody's Talking About

Here's where it gets interesting for B2B companies, and frankly, more concerning than the e-commerce implications.

The study included comparison tasks like "Ramp vs Brex" which had very few clicks (≤6 total across all tasks). Think about what that means: someone researching your B2B SaaS product against competitors is forming their entire opinion without ever visiting your carefully crafted comparison pages, case studies, or product pages.

For comparison prompts, the study found that review sites dominate exits—83% of exits for Ramp vs Brex went to review sites. Not to the company websites. Not to your thought leadership content. To third-party review sites.

Your entire mid-funnel content strategy—all those comparison guides, buying guides, "why choose us" pages that nurture enterprise prospects through a 6-month sales cycle—is getting absorbed into the AI interface.

And here's the kicker: the study found that comparison tasks show the lowest exit share. People are making brand judgments and shortlist decisions entirely within AI Mode.

For B2B, where the buying journey is already complex and involves multiple stakeholders, this is a fundamental shift. Your prospects are doing their initial research and forming opinions before you even know they exist.

What The Study Reveals About Site Types And Visibility

One of the most practical findings is how AI Mode matches site types to user intent. The researchers categorized sites into brands, marketplaces, review sites, publishers, and platforms, then tracked which appeared for different query types.

The pattern is clear: brands beat marketplaces when users know what product they want, marketplaces are preferred when options are broad or generic, review sites appear for comparisons, and opinions highlight Reddit and publishers.

For the B2B world, this means:

• If someone is comparing your product to a competitor, you're competing with NerdWallet, G2, and Capterra for that citation

• If someone is researching your brand reputation, you're competing with publishers and community sites like Reddit

• Direct product queries might surface your site, but only if AI Mode decides you're the authoritative brand source

As the study notes: "Sites can't and won't be visible for all types of queries in a topic anymore; you'll need to filter your strategy by the intent that aligns with your site type".

The Part That Should Actually Concern You

Here's what keeps me up at night about these findings: it's not just about traffic loss. It's about losing control of your narrative during the crucial consideration phase.

The study found that even when comparing products—whether software or physical goods—users barely clicked out. In plain terms, AI Mode is eating up all top-of-funnel and middle-of-funnel interactions. Users discover products, compare features, and form opinions entirely within the AI interface.

For B2B companies with long, consultative sales cycles, this is existential. You've built your entire go-to-market around being present during the research phase. Now that research is happening in a black box, and you might only show up as a citation—if you're lucky.

The study also revealed something about transactional behavior: 77.6% of sessions had zero external visits, and 18% of exits were caused by users exiting AI Mode and going directly to another site. That last part is particularly tricky for attribution—people are researching in AI Mode, then going directly to your site (or a competitor's) without any trackable referrer. Your attribution models are about to get a lot messier.

So What Are You Supposed To Do Right Now?

Alright, enough diagnosis. Here's what I'm actually telling clients:

1. Become The Authoritative Source In Your Category

If AI Mode is citing sources instead of sending traffic, your job is to be the source it cites. This isn't about gaming the system—it's about genuinely being the most authoritative, comprehensive resource in your space.

For B2B, this means:

• Publishing original research and data that other sites reference

• Building comprehensive, definitive guides that become the category standard

• Getting coverage in the review sites and publishers that AI Mode actually surfaces for comparison queries

The study found that for comparison queries, review sites take the majority of exits. So yes, you need to actually invest in your G2, Capterra, and industry analyst relationships—not just as lead gen, but as visibility in AI Mode.

2. Rethink Your Site Classification

Google is clearly categorizing sites into types and matching those types to intent. You need to understand where you fit.

Are you a brand? A marketplace? A publisher? A review site?

Your site architecture, content strategy, and even your domain strategy might need to evolve based on which queries you want to be visible for. Some companies might need multiple properties—a brand site, a content/publisher site, and a community site—each optimized for different intent types.

3. Build For The Zero-Click Reality

Since most AI Mode interactions don't result in clicks, visibility itself becomes the conversion metric. Brand awareness is no longer a soft metric—it's the primary outcome of being cited in AI Mode.

This means:

• Your brand needs to be memorable enough that when someone sees it cited in AI Mode, they'll type it directly into their browser later

• You need direct response channels (email, community, events) that don't rely on search traffic

• Your brand building efforts just became your most important SEO work

4. Optimize For The Clicks You Do Get

The study found that when users do click out, it's almost always transactional—they're ready to buy or take a specific action. The study also revealed that product previews show up in about 25% of AI Mode sessions and get about 9 seconds of attention.

That means your product pages, pricing pages, and demo request pages need to convert immediately. No educational preamble. No slow trust-building. If someone lands on your page from AI Mode, they've already done their research—give them what they need to convert.

One participant quote from the study captures this perfectly: "It looks like it has a lot of positive reviews. That's one thing I would look at if I was going to buy this bag. So this would be the one I would choose." Social proof, availability, clear next steps—all above the fold.

5. Don't Abandon What's Working

Here's where people get stupid. They see a study like this and think "SEO is dead, time to pivot entirely to AI optimization."

Let me be clear: AI Mode is at 1% adoption. Regular Google search still drives the majority of web traffic. Even Google's AI Overviews, which are much more widely deployed, only appear on a fraction of searches.

Keep investing in traditional SEO, paid search, and content marketing. They still work. They'll continue to work for years. Just add AI-focused strategies to your mix—don't replace everything wholesale.

My Contrarian Take: The Market Will Push Back

Here's what I think most people are missing in the "AI Mode will kill everything" narrative: the market has a say in this too.

Publishers are already fighting back against AI summaries. Regulators are paying attention to AI's impact on the information ecosystem. And most importantly, Google's shareholders care about that $200+ billion in ad revenue.

The study compared AI Mode to AI Overviews and found that on desktop, AI Overviews had a 93% zero-click rate within the AIO panel, but users still clicked on regular organic results below it. AI Mode produces roughly double the in-panel clickouts compared to AIO, but there are no organic results below to click on.

See the problem? AI Overviews can have high zero-click rates because there are still blue links below them. AI Mode can't scale at zero clicks because there's no monetization path.

Something has to give. Either Google figures out how to inject commerce and ads into AI Mode (which will degrade the experience), or they throttle the rollout, or they find some hybrid model we haven't seen yet.

My bet? It's the hybrid model, and it takes 3-5 years to figure out, not 3-5 months.

The Real Question For Leaders

Look, I'm not here to tell you AI Mode doesn't matter. The behavioral data is clear: when people use AI Mode, they stay inside it and make decisions without clicking out. That's real, and it's significant.

But the question for you as a leader isn't "should we panic?" It's "how do we hedge our bets intelligently?"

The companies that win over the next five years will be the ones that:

• Continue investing in what works today (traditional search, paid media, direct channels)

• Simultaneously build for an AI-first future (becoming the authoritative source, building brand equity, optimizing for citations)

• Stay flexible enough to adapt as the market evolves

As the study's authors note: "Strategy shifts from 'get the click' to 'earn the citation'". That's the fundamental change. But that shift is happening gradually, not overnight.

So yes, pay attention to studies like this. Use them to inform your strategy. Start building for a world where AI Mode has meaningful adoption.

Just don't bet the company on AI Mode taking over search in the next 12 months. Because Google, publishers, and the market have something to say about that too.

And in my experience, when that many powerful interests are involved, things move a lot slower than the hype suggests.

Source: Indig, Kevin and Van Buskirk, Eric. "AI Mode Is Sticky: New Usability Study Reveals How Google's AI Mode Changes Search." Search Engine Journal, 2025.