What Google AI Mode is.
Google AI Mode is Google’s conversational AI search experience. It answers a question with an AI-written response, links to the sources behind it, and lets the searcher keep asking follow-up questions in the same thread. Google’s own help page calls it its “most powerful AI search experience”.
Google introduced AI Mode on 5 March 2025 as an opt-in Labs experiment, describing it as a mode that “expands what AI Overviews can do with more advanced reasoning, thinking and multimodal capabilities.” It opened to everyone in the US on 20 May 2025 and, after an October 2025 expansion, Google said it would be available in over 200 countries and territories.
Interest is large. Google search data via DataForSEO shows “google ai mode” searched around 368,000 times a month in the US and 40,500 in the UK (November 2026). If you sell to businesses, your buyers are among those searchers, using it for the long, comparative questions that used to send them to several websites.
If you want the wider vocabulary first, our explainer on AEO, GEO and SEO covers it. Our guide to generative engine optimisation (GEO) goes further into the research behind that term. For the commercial side of being visible in AI answers, see our AI search visibility service.
How it works: query fan-out.
AI Mode does not run your question as one search. Google says it uses a “query fan-out” technique, “breaking down your question into subtopics and issuing a multitude of queries simultaneously on your behalf”, then brings the results together into one response.
That matters for anyone who publishes on the web. A single buyer question can trigger searches about pricing models, integrations, compliance and reviews at the same time. Your page does not need to answer the whole question. It needs to be the best answer to one of the sub-questions.
Google also announced Deep Search inside AI Mode, which it says “can issue hundreds of searches, reason across disparate pieces of information” for questions that need a more thorough answer. And it says AI Mode, like AI Overviews, runs on “a custom version of Gemini 2.5” in the US.
Google is open that the system makes mistakes. Its help page says: “As with any early-stage AI product, AI Mode doesn’t always get it right,” and advises people to check important information in more than one place. For a B2B buyer that usually means your website, your reviews and a call.
How AI Mode arrived.
AI Mode went from an invitation-only experiment to a worldwide product in about seven months. The dates below come from Google’s own announcements and Alphabet’s earnings remarks.
Two statements in that sequence matter for planning. In May 2025 Google said it would “graduate many features and capabilities from AI Mode right into the core Search experience.” In other words, AI Mode is also where Google tests what normal search becomes. Treat it as a preview of the results page your buyers will see.
AI Mode vs AI Overviews.
AI Overviews are a summary that appears at the top of some ordinary results pages. AI Mode is a separate, conversational place where the AI answer is the page. Google’s guidance says AI Mode “is particularly helpful for queries where further exploration, reasoning, or complex comparisons are needed.”
| Question | AI Overviews | AI Mode |
|---|---|---|
| Where it appears | At the top of a normal results page, when Google decides a summary helps | Its own tab and view; the searcher chooses to go there |
| Typical question | A quick factual or how-to question | Exploration, reasoning and complex comparisons |
| Follow-ups | A single summary on the results page | Built in: each follow-up is treated as a new query |
| How it finds sources | May use query fan-out | Query fan-out, and Deep Search for more thorough answers |
| Model | Custom Gemini 2.5 in the US (May 2025) | Custom Gemini 2.5 in the US (May 2025) |
| Console position | The whole Overview occupies one position; all its links share it | Follows the same method as a normal results page |
| To appear | None beyond normal search: the page must be indexed and eligible to show a snippet | |
The last row is the important one. Google’s documentation on AI features says “There are no additional requirements to appear in AI Overviews or AI Mode”. AI Mode draws from the same pool of pages as ordinary results; what changes is which pages get picked, and how often a buyer clicks through. Our guide to Google AI Overviews optimisation sets out what Google documents for that feature.
How many people use it.
Google has given usage figures in its quarterly earnings remarks, and they show fast growth in both users and question length.
Source: Alphabet earnings remarks by Sundar Pichai, published on blog.google on 23 July 2025, 29 October 2025 and 4 February 2026. Company-reported figures.
In February 2026 Google also said “a significant portion of queries in AI Mode” now lead to a follow-up question, and that daily AI Mode queries per user in the US had doubled since launch. For comparison, Google said in July 2025 that AI Overviews had over 2 billion monthly users.
These are company-reported numbers, so read them as direction rather than precision: more people, asking longer questions, and continuing the conversation instead of clicking away.
What happens to clicks.
Independent studies of AI summaries in Google show fewer clicks to websites when an AI answer appears. Google says the clicks that remain are higher quality. Both can be true, and a B2B site should plan for both.
The cleanest independent data comes from Pew Research Center, which studied the real browsing of 900 US adults in March 2025. Users who saw an AI summary clicked a traditional result in 8% of visits, against 15% when there was no summary. Clicks on the links inside the summary happened in just 1% of visits. Pew also found sessions ended more often after a page with an AI summary: 26% against 16%.
Ahrefs, which sells SEO software, measured the click-through rate of the top-ranking page across 300,000 keywords. In April 2025 it found an AI Overview correlated with a 34.5% lower average click-through rate. When it re-ran the study for December 2025, published February 2026, the figure was 58%.
Google’s position is different. Liz Reid, Head of Google Search, wrote in August 2025 that “total organic click volume from Google Search to websites has been relatively stable year-over-year”, and that Google is sending “slightly more quality clicks” than a year earlier. Google’s Search Central team says clicks from AI results “are higher quality, where users are more likely to spend more time on the site.” Google has not published the figures behind either statement.
For a B2B company the honest reading is this: early, informational traffic will fall, and the visits that remain will be later in the buyer’s research.
What it means for B2B research.
B2B buying questions are exactly the kind AI Mode was designed for: long, comparative and full of conditions. Pew found that AI summaries appeared on only 8% of one- or two-word searches but on 53% of searches of ten words or more. Google says AI Mode queries are three times longer than traditional ones.
Buyer, in AI Mode: “We’re a 200-person UK logistics company on HubSpot. Which payroll and HR platforms handle UK and Irish payroll, integrate with HubSpot, and are realistic for a team without an HR systems person?”
Follow-up: “Which of those publish implementation timelines, and what do reviewers say about onboarding?”
What the engine needs to find: pages that state the markets served, the integrations, the team size suited, the implementation steps and time, and independent reviews. Each is a separate sub-search. A vendor that only says “seamless integrations for growing businesses” answers none of them.
An invented question, written to show the pattern. It is not a real query or a real answer.
Three things follow for B2B sites.
- Specifics win sub-searches. Fan-out rewards pages that answer one sub-question precisely: “who it is for”, “how it is priced”, “how long it takes”, “what it integrates with”. Vague category pages lose because they answer nothing exactly.
- Comparisons happen without you. The buyer asks the engine to compare. If your site does not say plainly how you differ, the engine will rely on review sites and competitors’ pages instead. Our guide to the off-site sources AI engines trust covers that side.
- The click is later and warmer. When a buyer does click from AI Mode, they have often already compared. The page they land on has to confirm fit and make the next step easy, not restart the pitch.
Seeing AI Mode in Search Console.
AI Mode traffic is already in your Search Console data, mixed into the normal totals. Google says AI features are “reported on in the Performance report, within the ‘Web’ search type,” and are “counted towards the overall data in Search Console.” Google’s documentation does not describe a separate AI Mode filter.
Google’s help page on how Search Console measures performance sets out the rules for AI Mode:
| Metric | AI Mode | AI Overviews |
|---|---|---|
| Click | A click on a link to an external page counts; clicks that stay within Google do not | A click on a link to an external page counts |
| Impression | Standard impression rules | Standard impression rules |
| Position | Same method as a normal results page | The Overview occupies one position; all its links share it |
| Follow-up questions | A follow-up is treated as a new query, with its own impressions, position and clicks | Not addressed |
| Search Labs experiments | Not included in Search Console | |
What you can usefully do with it
- Watch impressions and clicks diverge. Pages that gain impressions but lose clicks are often being shown inside AI answers. Note them; they are the pages whose job has changed from “attract” to “confirm”.
- Read the long queries. Look through the queries list for long, conversational questions. Because each follow-up counts as a new query, new long questions appearing in your report is a sign of AI Mode use around your topic.
- Judge by enquiries, not clicks. Google itself recommends judging search by overall visit value, such as sales and signups, rather than clicks alone. Connect landing pages to form submissions in your analytics and CRM. Our guide to website lead tracking shows how.
- Ask new leads. Add a “how did you hear about us?” field with an AI assistant option. It is the only way to capture research that ends in a typed URL or a branded search.
What to change on a B2B site.
Most of the work is making your site the precise, checkable answer to the sub-questions buyers ask, and keeping it technically eligible. Google’s own list for doing well in AI experiences starts with “unique, non-commodity content” and a page experience that works “regardless of whether visitors arrive from classic or AI results.”
The B2B change list
- ✓One page per service, use case and audience. State who it is for, the problem, the method and what happens next. Fan-out sub-searches land on specific pages.
- ✓Say how you are priced, even if you cannot publish a number: the model, what changes the cost, and typical engagement shapes.
- ✓Publish honest comparisons: when you are a good fit, when you are not, and how you differ from the usual alternatives.
- ✓Date your evidence. Case studies and numbers with a source and a date are easier for an engine and a buyer to trust.
- ✓Answer first. Open each section with a one- or two-sentence direct answer, then the detail. Our guide to page patterns AI engines quote covers this in depth.
- ✓Keep company facts consistent across your site, Google Business Profile, directories and review sites. Google lists keeping Business Profile information up to date among its recommendations.
- ✓Stay technically eligible. Pages must be crawlable, indexable and allowed to show a snippet. A
nosnippetormax-snippetsetting limits what can be shown from the page in Search, including AI features. - ✓Make structured data match the page. Google says structured data should match the visible content. It helps; it is not a shortcut.
- ✓Design the landing page for a warm visitor. Clear fit, proof and a low-friction next step above the fold, because the comparison has often happened already.
An order of work that fits a small team
- 01 · LISTList 20 real buyer questions and their sub-questions
- 02 · TESTAsk them in AI Mode; note who is named and cited
- 03 · GAPSMap each sub-question to a page, or a missing page
- 04 · FIXWrite or rewrite the missing and vague pages first
- 05 · MEASURETrack impressions, clicks and enquiries per page
Repeat the test monthly; AI answers change from run to run.
For the way to run that test properly and keep a baseline, our guide on checking your AI search visibility gives the method. If the manual test becomes a routine job, see our comparison of the main AI visibility tools.
What you can ignore.
A lot of advice sold as “AI Mode optimisation” is unnecessary for Google. Google’s guide to optimising for generative AI features in Search lists things site owners “can ignore”. We cover llms.txt, schema and AI crawlers in a separate guide.
| Tactic | What Google says | Do this instead |
|---|---|---|
| llms.txt and “AI text files” | Google Search ignores them; they “will neither harm nor help” | Make the HTML pages clear and crawlable |
| “Chunking” content into tiny pieces | No requirement; Google can understand several topics on a page | Use clear headings and answer-first sections for readers |
| Rewriting content just for AI | “You don’t need to write in a specific way just for generative AI search” | Write for the buyer’s real question |
| Inauthentic “mentions” | “Isn’t as helpful as it might seem” | Earn genuine reviews and coverage |
| Special AI schema | “No special schema.org markup you need to add” | Keep normal structured data accurate |
If a proposal puts any of these at the centre of an AI Mode plan, ask what else is in it. Our guide on what AI search optimisation should cost lists the other red flags.
The AI Mode readiness check.
Tick each statement that is true for your website today. You get a score out of 100 and a list of fixes in priority order. Nothing you tick is sent anywhere.
Readiness score: 0 / 100
Tick what is true today. If you are unsure, leave it unticked.
The weights put eligibility and a tested question list first. The score is a planning aid, not a prediction of how often you will be named. If the list of fixes is more than your team can take on, our GEO and AEO services cover this work.
Questions owners ask.
What is Google AI Mode?
Google AI Mode is a conversational mode of Google Search. It uses a custom Gemini model to break a question into subtopics, run many searches at once, and write one answer with links to its sources. Searchers can ask follow-up questions in the same thread.
How is AI Mode different from AI Overviews?
AI Overviews are summaries at the top of some normal results pages. AI Mode is a separate view where the AI answer is the page, designed for exploration, reasoning and complex comparisons, with follow-up questions built in. Both use the same pool of indexed pages.
Do I need to do anything special to appear in AI Mode?
No special markup or files. Google says there are no additional requirements beyond normal search: the page must be indexed and eligible to show a snippet. What helps is clear, specific, well-evidenced pages that answer the sub-questions your buyers ask.
Can I see AI Mode traffic in Search Console?
Yes, but not separately. Google reports AI Mode inside the Performance report’s Web search type. Clicks to your site count, position follows the normal method, and each follow-up question counts as a new query.
Will AI Mode reduce traffic to our website?
Probably for early, informational questions. Independent studies show fewer clicks when AI summaries appear, while Google says remaining clicks are higher quality. Plan for fewer, later-stage visits and make sure your pages turn them into enquiries.
Is AI Mode available in the UK?
Yes. Google said in October 2025 that AI Mode would be available in over 200 countries and territories, and its help page lists the supported countries and languages.
Sources: Google, AI Mode Labs launch, 5 Mar 2025; Google, AI Mode US rollout, 20 May 2025; Google, AI Mode expansion, 7 Oct 2025; Alphabet Q2 2025 remarks; Alphabet Q3 2025 remarks; Alphabet Q4 2025 remarks; Google Search Help, AI Mode; Google Search Central, AI features and your website; Search Console Help, how performance is measured; Google Search Central Blog, 21 May 2025; Google Search Central, generative AI features guide; Google, Liz Reid, 6 Aug 2025; Pew Research Center, 22 Jul 2025; Ahrefs, 17 Apr 2025; Ahrefs, 4 Feb 2026; Google search data via DataForSEO, US/UK, November 2026. All accessed 10 November 2026.