Google's Gemini 3 Now Powers AI Overviews for 1 Billion Users, and Your Website Needs to Adapt

Google announced on 27 January 2026 that Gemini 3 is now the default AI model powering AI Overviews globally, reaching over one billion users. In the same update, users can now flow directly from an AI Overview into a full AI Mode conversation, asking follow-up questions without leaving Google. For UK businesses that rely on search traffic, this changes how customers find and evaluate them.

"We're rolling out Gemini 3 as the default model for AI Overviews globally, so even more people will be able to access best-in-class AI responses, directly in the results page," said Robby Stein, VP of Product for Google Search, in the official announcement.

The upgrade has immediate implications. Model changes affect which pages get cited and how responses are structured. As Search Engine Journal reported, Gemini 3 brings deeper reasoning capabilities that could produce different citation patterns from its predecessor, meaning the websites Google's AI chooses to reference may shift.

What Changed on 27 January

Two things happened simultaneously. First, Gemini 3 replaced the older Gemini 2.5 family as the engine behind AI Overviews worldwide. Previously, Google had been routing only complex queries to Gemini 3 through a selective system. Now it handles all queries by default, with lighter models like Gemini 3 Flash for simpler searches and Gemini 3 Pro for complex ones.

Second, Google introduced a seamless path from AI Overviews into AI Mode. When a user reads an AI Overview and wants to know more, they can ask a follow-up question directly. The conversation carries forward. No restarting, no context loss. Google tested this in December and found users strongly preferred the continuous experience.

The combination matters. AI Overviews already appear for a significant proportion of Google searches. Adding conversational depth means users can now get their second, third, and fourth questions answered without ever clicking a website link.

The Second Click Is Dying

Most coverage of this update focuses on the technology. The business impact cuts deeper. Research shows AI Overviews have already reduced organic click-through rates by 20-40% on average, and Ahrefs data puts the CTR drop at 58% for queries with AI Overviews. Some sites ranking number one organically have lost up to 79% of their traffic for queries where an AI Overview appears above them.

Before the AI Mode integration, there was at least a chance the user would click through for more detail. Now Google is explicitly designed to handle those follow-up questions itself. The "second click" (the moment a user needs more information and visits your website) is being absorbed into Google's conversational interface.

"This will result in less traffic to your site and more traffic to Google's AI Mode."

- Barry Schwartz, Founder, Search Engine Roundtable

For businesses, this means winning the initial citation is no longer enough. Your content needs to feed the conversation, providing the specific, structured answers that Gemini 3 can confidently attribute to you across multiple follow-up questions. This emerging discipline is called generative engine optimisation (GEO), and it requires different thinking from traditional SEO.

What This Means for UK Businesses

AI Overviews are live in the UK. The Gemini 3 upgrade is rolling out globally, and UK searches are included. For businesses that depend on Google for customer acquisition (which is most of them), this represents a structural change, not a temporary fluctuation.

The scale of the visibility gap is already measurable. When we tested 100 UK small businesses for AI visibility, 97 had no AI discovery files whatsoever. Their average visibility score was 31 out of 100. These businesses are largely invisible to the AI systems that are now mediating how customers find services.

Meanwhile, 47% of B2B buyers already use AI for vendor research. AI-referred visitors convert at 23 times higher rates than those from organic search. The businesses that AI systems cite and recommend are capturing disproportionate value. The ones that AI systems don't know about are losing ground they may not recover.

This is not limited to large companies. A plumber in Kettering, a solicitor in Northampton, a hairdresser in Corby: when a potential customer asks Google "who's the best plumber near me?" and Google answers with an AI Overview that flows into a conversation, the businesses that get cited are the ones whose websites provide clear, structured, verifiable information about what they do, where they operate, and why they're qualified.

How to Adapt Your Website

The shift from ranking to citation requires specific changes to how websites are built and structured.

Structure content for AI extraction. AI models cite content they can confidently attribute. This means clear headings that match real questions, concise factual statements, and logically grouped sections. A page titled "Our Services" with vague marketing copy is harder for AI to cite than a page with specific headings like "Emergency Plumbing Prices in Kettering" followed by direct, factual answers.

Answer the follow-up questions. With AI Mode, Google now handles multi-turn conversations. Think about the three questions a customer asks after their initial search. If your website answers "How much does a boiler replacement cost?" but not "How long does installation take?" or "Do you offer finance?", you're losing the conversation to a competitor whose site covers the full thread.

Implement AI discovery files. Files like llms.txt, ai.json, and identity.json help AI crawlers understand your business identity, services, and expertise. Without them, AI systems have to guess, and they often guess wrong or simply don't mention you. Google is already testing llms.txt for Discover and AI Mode, making early adoption a strategic advantage. Our AI visibility service covers implementing these files alongside broader AI optimisation. You can also check your current status using the AI Visibility Checker.

Demonstrate real expertise. Gemini 3's deeper reasoning capabilities mean it's better at distinguishing real authority from surface-level content. First-hand experience, original data, specific case studies, and verifiable credentials all increase the likelihood of citation. As one industry analysis noted, "You can't SEO-optimise your way into having experiences you never lived or expertise you were never accredited for."

Maintain technical performance. AI crawlers and traditional Googlebot both benefit from fast, well-structured sites. Pages that load quickly, use semantic HTML, and pass Core Web Vitals thresholds are more reliably crawled and indexed, and therefore more available for AI citation.

What to Watch

Google stated that AI Mode availability varies by market, with the full AI Mode conversational experience already available to Search Labs users and expanding. The Gemini 3 upgrade to AI Overviews is rolling out globally as of this week.

The EU's Digital Markets Act proceedings against Google, which we covered in our analysis of EU forcing Google to share search data with AI rivals, could reshape how AI Overviews source their information. The UK's CMA has also acted, proposing conduct requirements that would let publishers opt out of AI Overviews without losing traditional search visibility. If Google is required to share ranking data with competing AI platforms, the businesses that structure their sites for AI citation will benefit across multiple systems, not just Google's.

Google previously revealed that AI Mode can access Gmail and Google Photos to personalise search results. Combined with Gemini 3's deeper reasoning, this means AI Overviews will increasingly tailor answers based on individual user context, making broad, generic content even less effective and specific, authoritative content even more valuable.

Gemini 3 is also the engine behind Chrome's new auto-browse feature, which lets an AI agent navigate websites, fill forms, and complete purchases on behalf of users. And Gemini's reach extends beyond Google's own products: Apple confirmed that Gemini now powers Siri across 2.5 billion devices, meaning the same AI model shaping your search visibility is also processing voice queries from every iPhone. Websites need to work for AI visitors, not just human ones.

For UK businesses, the message is clear: the shift from traditional search to AI-mediated discovery is accelerating. Websites built to rank in a list of ten blue links need to be rebuilt to be cited in a conversation. The businesses that adapt their website structure now will capture the customers that AI systems direct their way. Those that wait may find the conversation has moved on without them.

Frequently Asked Questions