This article was AI-generated as part of an experimental historical-content project. The date reflects the period being analyzed rather than the date the article was originally written.
On Tuesday Google made two changes to Search that read like product housekeeping. According to Robby Stein’s post, Gemini 3 is now “the new default model for AI Overviews globally.” And people can ask a follow-up question directly from an AI Overview and continue into a conversation in AI Mode, keeping the Overview’s context. The follow-up flow is available on mobile, globally.
Neither change involves a new tab, a new app or a sign-up. That is the part worth noticing.
A model change nobody opts into
When Gemini 3 arrived in November, it came to Search through AI Mode first, for paying U.S. subscribers who selected “Thinking” from a menu. Google said it would also start routing the hardest questions in AI Mode and AI Overviews to the new model. Those were changes people had to go looking for, or that touched a subset of queries.
An AI Overview is different. It appears on the ordinary results page, above the links, for questions where Google decides it helps. Changing the default model behind it means the first generated paragraph many people read about a company can change for the same query, on the same page, without the user doing anything differently. The query is the same. The writer is new.
For reputation teams the consequence is plain. Any baseline captured before January 27 describes a different system. If you saved screenshots of how AI Overviews summarized your company, your executives or last year’s crisis, they are now historical documents. Useful for comparison. Not a picture of what people see today.
The follow-up is the new second page
The second change may matter more over time. Until now an AI Overview was a short answer with links underneath. Someone who wanted more clicked a link, refined the query, or opened AI Mode and started over.
Now the Overview is the opening move of a conversation. Google says that in its testing people preferred “an experience that flows naturally into a conversation,” and that keeping the context from the Overview makes follow-ups more helpful. For a company, that means the Overview frames the questions that come next. If the first paragraph mentions a lawsuit, the obvious next question is about the lawsuit, and AI Mode answers it already primed by the paragraph above.
In 2023 I suggested that Gemini might change who summarizes you first. The new design adds a wrinkle. The first summary now also sets the agenda for the second.
What to check this month
Re-run the questions that matter most, on mobile, and treat the results as a new baseline. Note which sources the Overview cites now and whether they differ from what you recorded before.
Then use the follow-up. Ask what a skeptical reader would ask after reading the Overview, and see what AI Mode says when it inherits that context. A neutral first paragraph followed by a pointed second answer is a different problem from a pointed first paragraph, and it usually needs different evidence to fix.
None of this suggests the new model is worse. It may well be better. But better on average is not the same as unchanged for you, and an upgrade nobody opted into is exactly the kind of change that monitoring tends to miss.