Google is updating how AI Overviews work in Search by making Gemini 3 the default model powering them. At the same time, the company is making it easier for users to move from a short AI-generated summary into a longer, conversational exchange through AI Mode.
The changes were announced in a Google blog post and are rolling out globally in markets where AI Overviews are available. Google says the feature now reaches more than one billion users.
Until now, Gemini 3 — Google’s latest large language model — was used mainly inside AI Mode, the company’s conversational search experience built for complex queries. With this update, the same model is now handling AI Overviews as well, bringing stronger reasoning and follow-up capability directly to the main search results page.
Table of Contents
Gemini 3 moves into AI Overviews
Gemini 3 was introduced in November and deployed to AI Mode on launch day. Google is now expanding its role by using it as the default engine behind AI Overviews.
Robby Stein, Google’s Vice President of Product for Search, said the change is meant to bring higher-quality AI responses into the parts of Search where users typically begin.
“We’re rolling out Gemini 3 as the default model for AI Overviews globally,” Stein wrote, adding that the goal is to surface more capable AI responses directly on the results page when they are useful.
For most users, the shift will not be obvious. But under the hood, it means AI Overviews are now generated by the same model that powers Google’s more advanced conversational search. That could influence how answers are structured, how queries are expanded, and which sources are cited.
Moving from summary to conversation
The more visible update is how users can continue beyond an AI Overview.
Google now lets users ask follow-up questions directly from an AI Overview and carry that interaction into AI Mode. The original context is preserved, so users don’t have to restate or reframe their query to continue the conversation.
According to Google, this reflects how people already use Search. Straightforward queries — such as checking the weather or a sports score — often need nothing more than a quick answer. More complex questions, however, tend to evolve into comparisons, follow-ups, and deeper exploration.
“People come to Search for an incredibly wide range of questions,” Stein said, noting that some require only a snapshot, while others benefit from a more conversational approach.
Google describes the result as a single, fluid experience, allowing users to move from a brief overview to a longer exchange when needed. The company says it tested this flow globally on mobile in December and found that users preferred being able to continue naturally into conversation without losing context.
Why this matters
Since their introduction, AI Overviews have increasingly drawn attention away from traditional search results and toward AI-generated responses. Each update has made it easier for users to stay within Google’s AI layer rather than clicking through to individual websites.
The shift to Gemini 3 is notable because changes in the underlying model can affect how answers are framed, how many sources are cited, and which pages receive visibility. When Gemini 3 was first introduced in AI Mode, it enabled broader query expansion and more dynamic response layouts, already altering how results were generated.
Using the same model for AI Overviews could further narrow the gap between a short summary and a full conversational response — especially now that users can move directly into AI Mode from the overview itself.
What comes next
Google says the rollout begins today, though availability may differ by region.
The company has previously said it plans to introduce automatic model selection, routing simpler questions to faster models while sending more complex queries to Gemini 3. It is not yet clear how that approach will affect AI Overviews beyond this default model change.
For now, the broader direction remains consistent. Google Search is continuing its shift away from static results pages and toward an experience built around ongoing, AI-driven interaction — with Gemini playing an increasingly central role.
