Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124

Google Searching doesn’t look like Google anymore, and in I/O By 2026, the company has gone full-throttle. The annual developers conference held this week in Mountain View, California, was the clearest statement yet about where Google will take its flagship product: away from the blue link model it has perfected over 25 years and toward something closer to conversation. Amnesty International factor.
The announcements ranged from Gemini 3.5 Flash becoming the new default engine behind global AI mode, to a complete reimagining of the search box itself, which Google described as the biggest upgrade to that interface in more than two decades.
So far, AI has appeared in Google search in the form of so-called AI Overviews and in a separate AI mode that looks more like talking to a Gemini chatbot. The new interface will instead be modified to match the style and results of your search query – including a ‘smart search box’ that lets you ask longer, more complex questions. Here’s what to research from Google I/O.
Robbie Stein, Google’s vice president of search product, portrayed this year’s I/O updates as a major step in bringing together Google Search and advanced AI, tracking the progress from AI overview to AI mode, and now a unified AI search experience. He said that a billion people use Google’s AI mode every month, and they ask him more questions. These tools allow people to ask virtually anything and get rich, real-time answers from Google’s extensive knowledge systems, he said.
“This is a very exciting time for research,” Stein told reporters ahead of I/O. “People can really ask anything that’s on their mind, and people’s curiosity is fairly endless.” The company is doubling down on its efforts to integrate leading AI models with live Google data (web pages, business listings, products, photos, finance) to deliver deeper, conversational search results.
The changes come as Google also announced the introduction of Gemini 3.5 Flash, a more capable model that focuses on thinking, coding, and complex tasks. Building research tools around the new model increases the quality of the overall answers to the research, Stein said.
Along with the form upgrade, Google is introducing a “smart search box” that accommodates longer queries, accepts uploads (images and PDFs), automatically completes granular prompts and can access contextual sources like open Chrome tabs to support multi-step searching.
AI Overviews now seamlessly transitions into AI mode to continue. So, instead of just getting an AI-generated answer in your search, you can have a conversation with the AI that provides your search results to get the answers you’re looking for.
New “gadgets” can simulate physics, visualize concepts, build calculators, or become persistent applets for tasks like movement, health tracking, or trip planning.
Stein also introduced dynamic, interactive “widgets” and larger system-generated “super-widgets” (enabled by Gemini and developer tools). They can simulate physics, visualize concepts, build calculators, or become persistent applet apps for tasks like navigation, health tracking or trip planning – optionally using connected personal data (Gmail, photos, calendar) to personalize results across 200 markets and 98 languages.
Stein described search moving into the “agent” era where AI agents can help you with a range of tasks, such as monitoring topics, sending alerts (such as when your favorite artist announces a tour) or booking services. Although an agent can’t book a reservation on your behalf, you can share your details — such as preferred dates and times and the number of people joining your party — to receive a list of matches with availability and updated prices, and links to officially finalize your reservation. These capabilities will be available this summer.