Follow ZDNET: Add us as a preferred source on Google.
ZDNET’s key takeaways
- Google Search is getting a new AI Search box.
- Search information agents monitor topics in the background.
- Agentic coding lets you build small apps in Search.
Remember when Google Search felt simple? I’d type a few words into a box, scan a list of blue links, and hope for the best. That version of Search is long gone, buried deep under AI. At I/O 2026, Google announced a bunch of Search updates that made it clear the product is becoming something more conversational, more personal, and more like an assistant that can do things for me.
Also: Sick of AI in Search? These 7 Google alternatives still put links first
The company said it is bringing “advanced model capabilities to Search with new AI features,” including a new AI Search box, information agents, agentic coding, and a personalization feature that pulls from my Google app data, to name a few things.
“The goal of Search has always been simple: to help you ask anything on your mind,” said Liz Reid, vice president and head of Search. The difference now, at least as I see it, is that Search is designed not just to answer, but to research, shop, book, monitor, and create on your behalf.
Here’s what’s new, who gets it, and when I think you’ll actually be able to use it.
1. AI Mode runs on Gemini 3.5 Flash
The AI Mode tab in Google Search on desktop and mobile is powered globally by a new model called Gemini 3.5 Flash.
At I/O, Google announced the Gemini 3.5 family and described it as a “major leap forward in building more capable, intelligent agents.” It said 3.5 Flash can provide “frontier performance for agents and coding, excelling at complex long-horizon tasks.”
It’s basically a faster, agent-ready model that can reason across sources, handle longer prompts, understand images and video, and complete multistep workflows. Google said it is also making it easier to “continue the conversation” in Search. I can now ask follow-up questions from an AI Overview, move into a “conversational back and forth with AI Mode,” and Search keeps my context.
Also: How to remove AI Overviews from Google Search: 4 easy ways
According to Google, AI Mode has passed one billion monthly users. Now, it’s rolling out 3.5 Flash as the default model behind it.
- Who can access? AI Mode with Gemini 3.5 Flash is available globally across devices.
- When will it be available? Now, according to Google’s I/O 2026 announcement.
2. A new AI Search box
Google is overhauling the Search box and decoupling it from keywords.
Instead of making me compress a messy thought into a few search terms, the new box is built for conversational, multimodal questions. I can just go ahead and enter whatever I want now. As Google put it, “Because your curiosity doesn’t always fit into keywords,” it is introducing “the biggest upgrade to our Search box in over 25 years… now completely reimagined with AI.”
For me, that means Search should better understand my specific, multipart ramblings. Instead of typing “best portable Bluetooth speaker waterproof Alexa,” I could ask something closer to how I would ask a person: “I want a portable Bluetooth speaker to take out by the pool. It’d be nice if it were waterproof and supported Alexa. Which ones are worth buying?”
Also: OpenAI’s new image watermarks make it easier to spot AI fakes – here’s how
I can ask with text, images, files, videos, or Chrome tabs. Google said the new AI Search box would put “powerful AI tools right at your fingertips” with AI suggestions that “go beyond autocomplete.”
- Who can access? The new AI Search box is available to all users.
- When will it be available? Google said it is “starting to roll out today” in areas where AI Mode is available.
3. Search agents can research things for you
One of the more interesting new Search features, to me, may just be agents.
Google said it’s entering “the era of Search agents,” where you can create, customize, and manage multiple AI agents inside Search. The idea is that these Search agents can keep working in the background after I ask a question.
The first version is information agents. I can tell it what I want to monitor, and it will keep checking the web, blogs, news, social posts, and other recent sources. Google used apartment hunting as an example: “You can brain dump all of the exact requirements you’re looking for, and your agent will continuously scan for you, notifying you when listings meet your needs.”
Also: How Google just revamped Gemini Enterprise for the agentic era – here’s what’s new
- Who can access? Google said information agents will come to Google AI Pro and Ultra subscribers first.
- When will they be available? Planned for this summer.
4. Search will help book local services
Search is adding agentic booking capabilities for local services and appointments.
Google’s example: Need a private karaoke room for six people on a Friday night, with food? Google Search with Gemini 3.5 Flash will show the latest pricing and availability with “direct links to finish booking through the provider of your choice.” But for select categories, like home repair, beauty, or pet care, you will be able to ask “Google to call the business on your behalf.”
Also: Google’s new Omni AI tool will let you video clone yourself – I’m intrigued (and concerned)
- Who can access? Google said, “These capabilities will roll out to everyone in the US.”
- When will it be available? Summer 2026.
5. Shopping is getting more agentic, too
Google also announced a new AI-powered shopping cart feature connected across Search, Gemini, Google Pay, Gmail, and YouTube. Called Universal Cart, it can follow my shopping research across Google services. It remembers products I’m considering, watches for price drops, finds alternatives, and helps build a cart using my payment, membership, loyalty, and shipping details.
Also: I let Chrome’s AI agent shop, research, and email for me – here’s how it went
Google said it was built on Google Wallet and “lets you quickly find opportunities for hidden savings or points.” It gave the example of building a custom PC, which needs a few parts from several retailers. “Your cart will proactively flag any product incompatibilities and suggest alternatives,” Google said. “It understands your payment method perks, loyalty information, and merchant offers.”
- Who can access? Universal Cart is coming first to Search and the Gemini app in the US, with YouTube and Gmail to follow.
- When will it be available? Universal Cart starts this summer.
6. Agentic coding comes to Search
Google is putting one of AI’s most popular use cases front and center in Search. Or, as Google put it, it’s bringing “the power of Google Antigravity and the agentic coding capabilities of Gemini 3.5 Flash right into Search.” That means I can ask Search to code small tools or apps, complete with a custom generative UI, layout, and real-time components such as interactive graphs.
Also: I stopped using ChatGPT for everything: These AI models beat it at research, coding, and more
Google’s examples include an astrophysics visualization, a wedding-planning dashboard, a moving tracker, and a fitness app that pulls in new data from reviews, live maps, local sources, and weather.
- Who can access and when? Google said “generative UI capabilities” will be available to everyone in Search this summer, “free of charge.” It also said users will be able to “build custom experiences with Antigravity, like mini apps, right in Search in the coming months,” starting first with Google AI Pro and Ultra subscribers in the US.
7. Personal Intelligence in AI Mode
Google is bringing its opt-in personalization features to Search through what it calls Personal Intelligence in AI Mode.
Also: Google’s new Omni AI tool will let you video clone yourself – I’m intrigued (and concerned)
If I choose to connect apps such as Gmail, Google Calendar, or Google Photos, Search will use my information from those apps to provide more personalized answers. That means it could be better at more personal questions, like finding a receipt buried in Gmail or surfacing relevant Google Photos while I’m researching something. I can connect or disconnect any app at any time.
- Who can access? Google said Personal Intelligence in AI Mode is “expanding” to people in “nearly 200 countries and territories across 98 languages — no subscription required.”
- When will it be available? It appears to be rolling out now.
My takeaway
Google Search has come a long way from the list of blue links.
What Google is building now looks more like a control panel for AI agents. I see this new version as one built for people who already use AI to search. It can handle long, messy, oddly specific questions, monitor topics for in the background, shop for products, book local services, build small tools, and personalize answers with information pulled directly from connected Google apps.
For power users, that’s the fun part: Google is putting agentic tools directly into Search.
For everyone else, I suspect the most noticeable changes will be the new Personal Intelligence features (if you opt in), the Gemini 3.5 Flash capabilities in AI Mode, and the new AI-powered Search box. Google Search is no longer just a place where I type a few keywords and click around. It is smarter, more personal, and, in some cases, willing to do more of the tedious work for me.
The trade-off is that the more Search becomes an agent that can personalize, monitor, and act for me, the more I have to trust Google with my data and the context behind what I’m asking.