Most people in SEO still talk about Google AI Overviews like they are the whole story. They are not even the biggest part of the story anymore.
At Google I/O 2026, Google quietly dropped a number that should have been the headline of the entire event: AI Mode had crossed 1 billion monthly users, with query volume more than doubling every quarter. That is not a feature in testing. That is a new front door to the internet, and it is growing faster than almost anything Google has shipped in the last decade.
If you have only been optimising for AI Overviews, you have been preparing for last year’s exam. AI Mode plays by different rules, rewards different content, and based on the data so far sends almost nobody back to your website. Here is what is actually happening, and what to do about it.
What Actually Changed When AI Mode Launched
Google announced AI Mode on March 6, 2025, built on Gemini 2.5. It started small, limited to Google One AI Premium subscribers testing it inside Search Labs. By May 2025 it opened to everyone in the US, no waitlist required. From there, the growth curve looks less like a typical feature rollout and more like a land grab.
By late 2025, AI Mode had already reached 75 million daily active users and more than 100 million monthly users across the US and India, according to Digital Applied’s analysis published in March 2026. Then came the number that changed how seriously the SEO industry needed to take this: at Google I/O 2026, Google confirmed AI Mode had surpassed 1 billion monthly users, with query volume more than doubling each quarter.
That last part matters more than the headline figure. A billion users today is impressive. A billion users that is doubling every three months means by the end of 2026, AI Mode could be processing a meaningful share of all Google queries not as an experiment sitting beside Search, but as a genuine alternative to it.
How AI Mode Actually Works – Query Fan-Out Explained Simply
The easiest way to understand AI Mode is to compare what happens behind the scenes when you ask it something complicated, versus what happens with a normal Google search.
Type “best laptops under $1000 for video editing” into regular Google, and you get a results page built around that one phrase. Type the same thing into AI Mode, and the system does something different: it splits your question into several smaller questions what specs matter for video editing, which laptops under $1000 meet those specs, what reviewers actually say about each one, how prices compare across retailers — and runs all of those searches at the same time. Google calls this query fan-out. The results get pulled together into one answer that reads like a single, coherent response rather than a list of separate findings.
This is powered by a custom-tuned version of Gemini 2.5, the same family of model behind Google’s broader Gemini products, but fine-tuned specifically for search. It is not just a faster AI Overview. It supports multi-turn conversation, meaning you can ask a follow-up question and AI Mode remembers everything from before no need to repeat context or start a new search. It can pull in live data: stock prices, restaurant availability, Google’s Shopping Graph, map information. And in its more advanced form, called Deep Search, it can run hundreds of queries in sequence and produce something close to a fully cited research report in a few minutes.
None of this is really “search” in the way the term has meant for the last twenty-five years. It is closer to handing your question to a very fast, very well-read research assistant who goes and checks several sources before getting back to you and who, increasingly, can also act on your behalf. Google has already started testing AI Mode completing small real-world tasks: finding and booking event tickets, checking restaurant availability, filling in forms for local appointments.
AI Mode vs AI Overviews: They Are Not the Same Thing
People use “AI Overviews” and “AI Mode” interchangeably, and that mix-up is causing real strategic mistakes. They behave differently enough that treating them as one thing means optimising for neither one properly.
- Sits on top of the normal results page
- One-pass synthesis – built from a fixed, limited set of sources
- Static. No follow-up, no conversation
- Best suited to simple, direct factual questions
- Now appears on roughly 60% of all Google queries
- Its own dedicated tab and experience
- Multi-turn, iterative retrieval using query fan-out
- Conversational — remembers context across follow-ups
- Built for exploration, comparisons, multi-part planning
- Over 1 billion monthly users as of mid-2026
The practical difference for anyone publishing content is this: AI Overviews mostly reward content that answers one clear question well. AI Mode rewards content that can survive being broken into pieces and recombined alongside several other sources — which means depth, structure, and genuinely original information matter more here than almost anywhere else in search.
“AI Mode is about doing things. AI Overviews was about presenting information differently.”
— MindStudio AI, “Google AI Search Mode Explained: What It Means for Your Workflows and Agents,” 2026
Why 93% of AI Mode Searches End With Zero Clicks
Here is the number that should genuinely concern every publisher, content marketer, and business owner reading this: Seer Interactive studied 25.1 million AI Mode impressions and found that 93% of those queries ended without a single click to any external website.
For context on how dramatic that is, the broader zero-click trend across all of Google sits somewhere between 60% and 68% depending on which 2026 study you read – SparkToro and Datos research puts it around 64.82%, climbing steadily from roughly 50% back in 2019. AI Mode’s 93% is not a slightly higher version of that trend. It is a different category of behaviour entirely.
The reason is structural, not accidental. AI Mode is specifically designed to answer the question completely inside the interface — comparing options, citing sources inline, letting the user refine and go deeper without ever leaving the page. The entire point of the product is to remove the need to click away. Where a normal search result might give you one perspective and expect you to click through for the rest, AI Mode tries to hand you the synthesised version of all the perspectives at once.
What Still Drives Traffic Even Inside AI Mode
The data is not uniformly bleak. A few categories continue to send real clicks, and understanding which ones is more useful than treating every query type the same way.
SEO consultant Rand Fishkin, cited in Search Engine Land’s 2026 zero-click study, pointed out that branded searches, local business queries, and high-intent transactional searches continue to perform well even as informational queries get swallowed by AI surfaces. Someone searching for your exact business name, someone looking for “plumber near me,” someone ready to compare prices before buying — these searchers still want a destination, not just an answer.
There is also a citation-position effect worth knowing about. According to Digital Applied’s 2026 research, sidebar citations inside AI search responses achieve a 6 to 10% click-through rate — comparable to ranking in organic positions four through ten on a traditional results page. That is a meaningful number. It means being cited prominently inside an AI Mode response, while it will never replace a top organic ranking in volume, is not worthless either. It is closer to a mid-page organic placement than to total invisibility.
And the clicks that do happen convert better. Ahrefs data shows that while AI Overviews can cut click-through rates by up to 58% on the affected queries, the traffic that survives that filtering converts at meaningfully higher rates — the people who click through after reading an AI summary have already done their initial research and are further along, not just browsing.
How to Actually Show Up in AI Mode Answers
Because AI Mode pulls from multiple sources per sub-question rather than picking one winner, getting cited here is less about ranking number one and more about being clearly, unambiguously useful for one specific slice of a bigger question.
Write for the sub-question, not just the headline keyword
Remember the laptop example — “best laptops under $1000 for video editing” gets broken into smaller searches about specs, pricing, and reviews. A page that thoroughly answers just one of those pieces — say, exactly which processor and RAM specs matter for video editing on a budget — has a real shot at being pulled into the fan-out, even if it never would have ranked for the broad head term on its own. Structuring an article around clearly labelled sub-topics, each answerable on its own, gives AI Mode more entry points to cite you from.
Lead every section with a direct, self-contained answer
Because AI Mode retrieves and recombines passages rather than entire pages, a section buried three paragraphs deep with no standalone answer at the top is far less likely to survive that process. Open each section with a clear, complete answer in the first sentence or two, then explain and support it afterward.
Back every claim with a named, dated source
Generic claims give an AI system nothing to verify against. A specific, attributed statistic “according to Seer Interactive’s study of 25.1 million impressions” rather than “research shows” gives the retrieval system an actual entity to cross-reference, which research from Princeton, Georgia Tech, and IIT Delhi found increases the odds of being cited by roughly 37%.
Make sure Google can actually crawl you
AI Mode draws from Google’s existing index. If your technical SEO is broken — pages not indexed properly, slow load times, blocked crawl paths — none of the structural work above matters, because the content was never eligible for retrieval in the first place. This is the same foundation covered in our guide on technical SEO for WordPress.
What to Do This Month — A Realistic Plan
You do not need a separate AI Mode strategy bolted onto everything else you are doing. You need to adjust the content you are already producing so it survives being broken apart and recombined.
- Pick your three highest-traffic existing articles and check whether each major section can stand alone as a complete answer if pulled out of context — most cannot yet, and that is the single highest-leverage fix available
- Identify the natural sub-questions hiding inside your broader topics and make sure each one has its own clearly headed, fully answered section
- Replace vague claims with named, dated sources throughout your content — this single change compounds across AEO, GEO, and now AI Mode citation eligibility
- Stop measuring every page purely by click volume — start tracking branded search trend in Search Console as a secondary signal of AI-driven brand exposure
- Keep investing in branded, local, and transactional keyword content specifically, since these remain the query types most likely to still send a click even as informational search gets absorbed into AI surfaces
- Re-check your robots.txt file to confirm Googlebot is not blocked anywhere — AI Mode cannot cite what it cannot crawl
Frequently Asked Questions
Q. What is Google AI Mode?
Google AI Mode is a separate, conversational search experience inside Google Search, built on a custom-tuned version of Gemini 2.5. Rather than returning a list of links, it breaks a question into multiple smaller sub-questions using a technique called query fan-out, searches all of them simultaneously, and combines the results into one cited, conversational answer that supports follow-up questions without starting a new search. It launched in testing in March 2025, opened to all US users by May 2025, and reached over 1 billion monthly users by Google I/O 2026.
Q. What is the difference between Google AI Mode and AI Overviews?
AI Overviews are static summaries that appear above the normal results on a regular Google search — built from a single-pass synthesis of a limited set of sources, with no follow-up or conversation possible. AI Mode is a completely separate, dedicated experience that uses multi-turn, iterative retrieval: it can hold a conversation, remember context across follow-up questions, and use query fan-out to research multiple sub-topics at once before combining everything into a single response. AI Overviews work well for simple factual questions. AI Mode is built for exploration, comparisons, and multi-part research tasks.
Q. Does Google AI Mode send any traffic to websites?
Very little, on average. A study by Seer Interactive covering 25.1 million AI Mode impressions found that 93% of queries ended without any click to an outside website — meaningfully higher than the roughly 60 to 68% zero-click rate seen across Google search overall in 2026. However, certain query types still perform better: branded searches, local business queries, and high-intent transactional searches continue to drive real clicks. Citations placed prominently inside an AI Mode response also achieve a 6 to 10% click-through rate, which is comparable to ranking in organic positions four through ten on a traditional results page.
Q. How do I optimise content for Google AI Mode?
Structure content around the natural sub-questions hiding inside a broader topic, since AI Mode’s query fan-out process searches for those pieces individually before recombining them. Open every section with a direct, self-contained answer that makes sense even if pulled out of its surrounding context, since AI Mode retrieves passages rather than full pages. Back claims with specific, named, dated sources rather than vague statements, which increases the likelihood of being cited as a trustworthy source. And make sure your technical SEO foundation is solid, since AI Mode draws from Google’s existing index — content that cannot be crawled and indexed properly is never eligible for retrieval in the first place.
Q. Is Google AI Mode the same as Gemini?
No, though they are closely related. Gemini is Google’s broader AI model, used across many products for tasks ranging from coding to image generation. AI Mode is a specific application built on a custom-tuned version of Gemini 2.5, fine-tuned specifically for search queries within Google Search. Gemini is the underlying engine; AI Mode is the search-specific product built on top of it.
Q. Should I stop investing in traditional SEO because of AI Mode?
No. AI Mode draws its answers from Google’s existing search index, which means strong traditional SEO — proper crawlability, indexing, content quality, and backlink authority — remains the prerequisite for being eligible for citation inside AI Mode at all. The smarter response is not to abandon SEO but to restructure content so it survives being broken into sub-questions and recombined: clear, self-contained answers within each section, specific named sources, and continued investment in branded and transactional keywords where clicks still flow normally even as informational search increasingly gets absorbed into AI-generated answers.
Kia has worked in SEO and digital marketing for over a decade, building and optimising websites across different industries. He founded Technexies to share what actually works in modern search written from direct professional experience rather than theory. All content on Technexies is researched, written, and reviewed by Kia personally.
