Type a question into AI Mode and the results page never appears. Gemini splits the query into a fan of sub-searches, reads across the sources it retrieves, and returns one synthesised answer in a thread the user can keep questioning. That is the difference that matters commercially. An AI Overview sits above links a reader might still click; AI Mode removes the links page and replaces it with a conversation. Your article is read, summarised and sometimes cited, but it is no longer the place the journey ends, and so it is no longer a place you can reliably monetise.
What is Google AI Mode?
Google AI Mode is a conversational search experience, built on the Gemini model, that treats a query as a dialogue rather than a list of results. The user asks a full question, receives a synthesised answer, and follows up in the same thread. Behind the scenes Google uses "query fan-out": it breaks the question into related sub-queries, runs them in parallel, and assembles a single response with a handful of supporting links. AI Mode is multimodal, and Google's CEO Sundar Pichai said on the company's third-quarter 2026 earnings call that one in five AI Mode searches globally use voice, images or documents. He also said AI Mode queries are on average three times longer than traditional searches. Google made AI Mode available to everyone in the United States in May 2025 and has since extended it to many more markets.
How is AI Mode different from AI Overviews?
AI Overviews and AI Mode are separate surfaces, and the gap between them decides how much traffic a publisher loses. An AI Overview is the generated summary at the top of an otherwise normal results page, so the blue links remain underneath it. AI Mode is a full conversational experience in which there is no list of links to fall back on. The two do not draw consistently on the same sources: one 2026 analysis found that they cite the same URLs only about 13.7 per cent of the time, even though they reach broadly similar conclusions. Being visible in one surface therefore does not guarantee visibility in the other. Google says it is "building one seamless experience across AI Overviews, AI Mode, and traditional Search", so publishers should expect the boundary between them to keep blurring.
How many people use AI Mode?
Google said in May 2026 that AI Mode had passed one billion monthly users within about a year of its public launch. Treat that as Google's own figure: the October earnings call gave query-length and input-type data but no updated user count. AI Overviews is larger, reaching around 2.5 billion users a month on Google's earlier figures, but AI Mode is the faster-growing and the more self-contained of the two. The scale is the point. AI Mode is not a beta to watch from a distance; it is already part of how a very large number of people search.
What does AI Mode do to publisher traffic?
AI Mode lowers referral traffic by resolving the query before the user needs to click. The strongest evidence now comes from randomised experiments rather than correlations.
A University of Pennsylvania and Northeastern University preregistered field experiment, posted to arXiv on 18 August 2026, randomly assigned 1,100 US adults to a suppressed-AI, default or forced-AI-Mode search experience in March 2026. Forcing AI Mode reduced external click-through by 18.8 percentage points, and the drop was concentrated in clicks to news sites, Reddit and Wikipedia. Participants also became 11.2 percentage points more likely to search on a rival engine such as Bing, DuckDuckGo or Yahoo. The authors note a limitation: Google changed the HTML of AI Overviews during the study, so the extension hid them in only 51.1 per cent of the suppressed-AI cases. These details come from TechWyse's 22 September 2026 write-up; please read the paper before quoting it.
A separate field experiment by Saharsh Agarwal of the Indian School of Business and Ananya Sen of Carnegie Mellon, covering 68,089 searches, found that outbound organic clicks fell 39.8 per cent when an AI Overview appeared, with zero-click searches up 34.5 per cent. Its AI Mode arm was smaller and the authors call it exploratory because of higher attrition, but it points the same way: 0.36 external clicks per search in AI Mode against 0.53 in the control group, alongside lower satisfaction scores.
Observational data agrees. Ahrefs found AI Overviews correlated with a 58 per cent reduction in click-through for top-ranking pages in February 2026, and Pew Research Center found users clicked a traditional result 8 per cent of the time when an AI summary appeared against 15 per cent when it did not. Correlation is not causation, but experiments and observational studies now converge.
Is Google putting ads in AI Mode?
Yes, and the rollout accelerated this week. On 28 October 2026 Google said it is scaling ads in AI Mode across Australia, Canada, India, Kenya, Malaysia, New Zealand, Nigeria, Pakistan, the Philippines and Singapore, with Brazil, Mexico and the United Kingdom planned for the fourth quarter. It is also launching multiturn ads in the United States so that ads can appear in follow-up questions. Ads are labelled "Sponsored", and Google says click-through rates are already similar between sponsored and organic product listings in AI Mode, though it gave no percentages.
Monetisation is not yet at parity. On the same earnings call, Chief Business Officer Philipp Schindler said AI Mode query growth had an impact "because of the lower monetization we were seeing in AI Mode", as reported by 24/7 Wall St on 29 October, while expecting it to improve. For publishers the structure matters more than the current yield. Google is building a way to monetise the answer assembled from publishers' content, and the publishers who supplied that content receive neither the click nor a share of the ad beside it.
What has Google changed for publishers in AI Mode?
Google has made some concessions on links. In May 2026 it announced updates across AI Overviews and AI Mode that include labels on links from a user's active news subscriptions, which it said early testing showed users were significantly more likely to click, and forum previews that credit the creator or community. Publishers are also pressing the issue legally and politically. Penske Media has filed an antitrust lawsuit against Google, and the European Publishers Council has lodged a complaint with the European Commission arguing that publishers face a forced choice between unlicensed use of their content and losing search traffic, according to The Next Web's 6 May 2026 report. These measures change the presentation of links; none yet changes the underlying economics of an answer that does not require a visit.
How can publishers respond to AI Mode?
The response runs along three lines. First, measure server-side. AI surfaces often strip or suppress the referrer, so the reliable record of who is reading you sits in server logs and at the CDN edge, not in client-side analytics that file this traffic as direct. Second, structure content for retrieval as well as ranking: lead each section with a direct answer, use question-format headings, add Article and FAQ structured data, and publish original data an AI system can attribute to you. This improves the odds of being the named source even when no click follows. Third, treat the read itself as the monetisable event. If the user resolves the question inside Google's surface, monetisation has to move to where the content is actually consumed. That is the layer blankspace works in: detecting AI and agent traffic at the CDN edge and placing contextual brand mentions into the response, so value is captured at retrieval rather than lost when the visit never happens.
Frequently asked questions
What is the difference between Google AI Mode and Gemini?
Gemini is Google's underlying model and its standalone assistant app; AI Mode is the search experience inside Google Search that uses Gemini to answer queries conversationally. Gemini is the engine and AI Mode is one place the engine is applied. Publishers care about AI Mode specifically because it sits inside Search, where the audience that once arrived as referral traffic now begins, and often ends, its query.
Is AI Mode replacing the normal Google results page?
Not entirely, but it is being positioned as a primary way to search. The classic results page still exists and AI Overviews still appear on it, yet Google describes a single experience spanning AI Overviews, AI Mode and traditional Search. As adoption grows, a rising share of searches resolves without a list of links, which is why publishers see referral declines even where their content is being read.
Does AI Mode cite and link to publisher sites?
Yes, AI Mode shows supporting links and citations, but they are surfaced selectively and clicked rarely. Pew Research found that links inside AI summaries were clicked about 1 per cent of the time, and experiments show a large fall in external clicks overall. Being cited is valuable for brand visibility, but it seldom replaces the click volume of the old results page.
Does Google pay publishers when it shows ads in AI Mode?
Google has not announced any revenue share with publishers for ads shown in AI Mode answers. Its 28 October announcement describes how advertisers reach users in AI Mode, not how the sources behind the answer are compensated. Publishers wanting payment for use of their content currently rely on licensing agreements, which are negotiated individually, or on legal and regulatory action.
How can publishers prepare for AI Mode now?
Start by reading server and edge logs rather than relying on client-side analytics, so you can see AI traffic that conventional tools miss. Structure your content for extraction with direct-answer openings, question-format headings and structured data. Then plan for monetisation that does not depend on a click, because a growing share of your audience gets what it needs inside the answer. Edge-level approaches that monetise the AI response itself, such as blankspace, are designed for that shift.