Let me tell you about the moment I realized how much trouble we're in. I was running a search in Google AI Mode last week — something about comparing CRM platforms for small businesses — and I got a complete, well-sourced answer with feature comparisons, pricing tiers, and implementation timelines. It took me about 90 seconds to read. I never clicked a single link. And I'm someone who writes about SEO for a living. I should know better. But the answer was just... good enough.
That's the problem. Not that AI Mode gives bad answers. The problem is that it gives answers that are good enough that most people never leave Google.
Google AI Mode launched in the US in March 2025, powered initially by Gemini, and it's been steadily expanding since. By Q3 2025, it had reached 75 million daily active users globally. As of early 2026, it's the default experience for a growing percentage of Google searches. And the data on what it's doing to web traffic is, frankly, alarming.
What AI Mode actually is (and isn't)
There's still some confusion about the difference between AI Overviews and AI Mode, so let me clear that up.
AI Overviews are the AI-generated summaries that appear at the top of regular search results. You've seen them — they show up above the traditional blue links, provide a quick answer, and include citation links. The blue links are still there below. AI Overviews are an enhancement to existing search. They change the experience, but they don't replace it.
AI Mode is something else entirely. It's a full-screen, conversational AI search experience. When you're in AI Mode, there are no blue links. No sidebar. No "People Also Ask." Just a Gemini-powered AI giving you a direct, detailed answer to your question, sometimes citing sources in inline links, sometimes not. It uses a fan-out technique that issues up to 16 simultaneous queries to assemble its response. The result feels more like talking to a very knowledgeable assistant than searching the web.
The distinction matters because the two features have radically different impacts on website traffic. AI Overviews coexist with organic results. AI Mode replaces them.
The zero-click numbers are staggering
Here are the numbers that should be keeping every SEO professional up at night.
In traditional Google search — no AI features at all — about 34 percent of searches end without a click to any external website. That's been the baseline for years, and it's always been a concern. But it was manageable.
When AI Overviews appear on a standard search results page, the zero-click rate jumps to 43 percent. That's a meaningful increase, but organic links are still visible. People still scroll down. The CTR drops from the organic average, but clicks still happen.
In AI Mode, 93 percent of searches end without a click. Ninety-three percent. That's not a drop in traffic. That's traffic functionally ceasing to exist for the vast majority of queries handled by AI Mode.
The overall picture is getting worse too. According to Click-Vision's 2026 zero-click study, 58.5 percent of all US Google searches now end without a single click to any external website. For queries that trigger any AI feature, that number climbs to 83 percent. The organic CTR for AI-enhanced results has dropped from 1.76 percent to just 0.61 percent — a 61 percent decline.
I want to sit with those numbers for a second because they represent something genuinely new. For twenty-five years, the implicit bargain of the web was: you create content, search engines send you traffic in exchange for being able to index and serve that content. AI Mode breaks that bargain. Google gets your content, feeds it to Gemini, generates an answer, and the user never visits your site.
Google cites itself more than anyone else
A study of 1.3 million AI Mode citations revealed something remarkable: Google.com accounts for 17.42 percent of all citations inside AI Mode responses. That's more than YouTube, Facebook, Reddit, Amazon, Indeed, and Zillow combined.
Let that sink in. When AI Mode does cite a source, it cites Google's own properties more than any external website. And 59 percent of those self-citations point to traditional organic search result pages — meaning AI Mode is essentially saying "for more info, go back to regular Google search."
This creates a closed loop. You search in AI Mode. AI Mode answers your question using content from various websites. If you want more detail, AI Mode sends you to another Google property. The original content creator — the website that actually produced the information — gets cut out of the loop entirely.
I've been trying not to be conspiratorial about this, but it's hard to see it as anything other than what it is: Google using other people's content to keep users within the Google ecosystem. Whether that's an intended strategy or an emergent behavior of the system, the effect on publishers is the same.
What this means for different types of sites
The impact of AI Mode varies dramatically depending on what kind of site you run.
Informational content sites are getting hit the hardest. If your business model depends on ranking for "how to" queries, definition queries, or comparison queries, AI Mode can answer most of those without ever sending a user to your page. Recipe sites, tech explainer sites, health information sites — these are the categories seeing the most dramatic traffic declines from AI-powered search features.
E-commerce sites are in a slightly better position, but not as safe as they might think. AI Mode can compare products, summarize reviews, and recommend purchases without users visiting individual product pages. The transaction itself still requires clicking through to a store, but AI Mode is absorbing the research phase that used to generate multiple site visits before a purchase.
Local businesses might actually benefit from AI Mode in some cases. When someone asks "best plumber near me" in AI Mode, the response typically includes business names, ratings, and contact information drawn from Google Business Profiles. If your local SEO is strong, you might get mentioned in AI Mode answers more reliably than you appeared in traditional results. But you'll get fewer website visits — the user might just call the number AI Mode provides directly.
SaaS and service businesses occupy an interesting middle ground. AI Mode struggles with nuanced, comparison-heavy queries that require understanding specific use cases. If your content addresses specific scenarios — "best project management tool for agencies with 10-50 employees" versus the generic "best project management tools" — you're more likely to get cited because the AI needs specialized sources to answer specialized questions.
How to get cited in AI Mode
Getting cited in AI Mode isn't the same as ranking in traditional search. The mechanics are different, and some of the strategies that work for organic ranking are irrelevant or even counterproductive for AI citation.
AI Mode pulls from multiple sources simultaneously and synthesizes them into a single answer. To be included as a citation, your content needs to be the kind of source that a synthesizing AI would naturally reference. That means being a primary source rather than a secondary one. If you're summarizing information available elsewhere, AI Mode will go to the original source instead.
Structured data matters more for AI Mode than it ever did for traditional search. Schema markup helps the AI understand what your content actually represents — whether a page contains a product review, a how-to guide, a scientific study, or a business listing. Pages with rich structured data are cited at higher rates than equivalent pages without it.
The format of your content affects citation rates significantly. AI Mode tends to cite content that presents information in clearly delineated, factual statements rather than content that buries facts within narrative prose. Tables, data points with clear labels, and well-defined comparison criteria all increase your chances of being cited. Ironically, the kind of content that reads worst to humans — dry, factual, highly structured — reads best to the AI.
Author authority signals appear to influence citation selection too. Content attributed to named authors with verifiable credentials gets cited more often than anonymous content. This makes intuitive sense — if AI Mode is trying to provide trustworthy answers, it's going to prefer sources that demonstrate expertise.
The uncomfortable strategic question
Here's what nobody in SEO really wants to talk about: should you even optimize for AI Mode citations?
Think about it. Even if you get cited in an AI Mode response, 93 percent of users aren't clicking through. The citation is a footnote, not a traffic source. You're spending resources to optimize for a system that, even in the best case, sends you very little traffic.
There are two arguments for doing it anyway. The first is brand visibility. Even without a click, having your brand name appear in AI Mode responses keeps you in front of users. Over time, that recognition compounds. Users who see your brand cited by Google's AI begin to associate your brand with authority on that topic.
The second argument is that AI Mode adoption is still growing. At 75 million daily users globally, it's significant but not yet dominant. Optimizing for AI Mode now means establishing your position before the competition catches on. The sites that figured out featured snippets in 2016 are the ones still dominating position zero a decade later.
But I'll be direct: I think the smart long-term play isn't optimizing for AI Mode specifically. It's building the kind of content — original, data-rich, expertly authored — that performs well across every format Google throws at us. Featured snippets. AI Overviews. AI Mode. Whatever comes next. Content that is genuinely the best source on a topic tends to surface regardless of the specific mechanism.
What I'm telling clients right now
I'm telling clients to track three things that most people aren't tracking yet.
First, citation share. How often your brand appears in AI Mode responses for your target queries, regardless of clicks. This is a new metric and the tools for tracking it are still immature, but it's going to matter more than traditional rankings within a year or two.
Second, branded search volume. If AI Mode is reducing direct traffic from informational queries, your branded search volume becomes a crucial indicator of whether your content marketing is still building brand awareness even without generating direct visits. If people are still searching for "[your brand] + [topic]," your visibility is working even without clicks.
Third, conversion rate from remaining organic traffic. If total organic traffic declines but you're still getting the most valuable, highest-intent visitors, your revenue might not decline proportionally. Focus on the quality of remaining traffic, not the vanishing quantity.
I won't pretend this isn't a difficult transition. The entire SEO industry was built on the assumption that creating useful content would be rewarded with traffic. AI Mode breaks that assumption. But the underlying principle — create content so good that Google's systems can't ignore it — still holds. The reward just looks different now. Instead of traffic, you get citations. Instead of clicks, you get brand mentions. It's less satisfying, sure. But adapting to it is better than pretending it isn't happening.
The 93 percent zero-click rate isn't going to improve. If anything, as AI Mode gets better and more widely adopted, it'll get worse. The question isn't whether to adapt. It's how fast you can.