AI agents are everywhere. Every briefing I’ve attended for software companies over the past year has involved some mention of agents—using generative AI tools to automate digital tasks. Despite breakout moments at the start of 2026, like the plucky OpenClaw agent that early adopters used to manage their online life, most people are not yet embracing this style of automation day-to-day. That won’t last for long if Google gets its way.

At Google I/O, the tech behemoth shared its vision to make its popular search engine a core way to expose billions of existing users to the company’s agentic prowess.

“You will be able to create, customize, and manage multiple AI agents for your many tasks, right in Search,” says Liz Reid, who leads Search at Google. She gives the example of setting up an agent to track stock market trends and send you alerts using real-time data—when specific conditions are met.

Alongside these agentic additions coming to Search, Google also announced a new underlying model, Gemini 3.5 Flash, as the global default model for AI Mode answers, as well as improvements to the Search box that make it more responsive to user inputs.

Always-On Answers

One of the key agentic experiences rolling out is a new form of dynamic data gathering through “information agents” that can be more proactive than previous search experiences and automate alerts through your conversational requests.

“Ask Google to just keep you updated on anything, and now our agents can do work for you even if you're not using Google,” says Robby Stein, a vice president of product for Search. “So, you could be asleep, and it's still helping you.” This feature will arrive first for subscribers to Google’s AI Pro and Ultra plans this summer.

In Google’s example of how these can work, a user asks AI Mode through the Google app to keep them “updated when any of my favorite athletes announce sneaker collabs or signature drops.” Then, AI Mode generates an info agent unique to that user, designed to monitor the request. When a new shoe drops that fits these criteria, like A'ja Wilson’s pink Nikes, the user gets a notification alert with critical context and ways to buy the sneaks.

Booking agents are another automation style in Search that Google will be expanding this summer. At previous I/O conferences, Google debuted similar AI features, like the now defunct Duplex, which called companies on your behalf to set up restaurant reservations or salon appointments. Google continues to iterate on that core idea with agents that can search for relevant context about local companies, even calling that barber down the street for a price quote on a beard trim if it doesn’t show up on the website, to help users gather info with less active participation.

Much of this keeps users trapped within Google’s ecosystem without actually browsing the web themselves, a practice we’ve seen before when features like AI Overviews launched in 2024. Even so, Google has explicitly stated that it’s not attempting to replace web page links.

Vibe-Coded Results

Google first launched Antigravity, its version of an agentic coding tool like OpenAI’s Codex or Anthropic’s Claude Code, late last year. Now, an updated version of Antigravity will be used to create bespoke outputs in Search, like an adjustable visualization of black holes to help you better understand deep space.

“Search can build you custom experiences,” Stein says. Google’s vibe-coding tool, Antigravity, will craft unique answer layouts, when helpful, and incorporate interactive visual elements, such as infographics and movement simulations. The update is expected to arrive later this summer for all users.

An expanded version of that tool, which Google says is a way to create “super widgets” and “mini apps,” will be available first for Google AI plan subscribers in the US this summer. (I vote to call them super widgets, which sounds more unhinged.) These can be used for tracking longer-term plans, like a healthy diet, or managing complex tasks, like a cross-country move, through Search.

Search Doesn’t Need You

Google of yore, with its 10 blue web links, asked users to be fairly active participants in the info-gathering process: Here are a bunch of webpages that appear relevant to your search. How about you click around and see which site has the best info? The process was clunky at times and could be time-consuming, but I was personally hopping site to site as Google nudged me into different online rabbit holes.

The launch of AI Overviews, which generates summaries of web pages and online information, two years ago, and the rollout of AI Mode, the chatbot-style search, last year, have automated aspects of this process. Google enabled these changes to generate more personalized answers for users—with some high-profile flubs as part of its growing pains.

AI Overviews and AI Mode further abstracted my daily experience of using a search engine. Instead of clicking on a bunch of random website links, I was reading an AI summary positioned at the top of my search results and sometimes clicking through to double-check the accuracy of the output.

The next evolution of Search that Google is building asks for even less active participation from users. You’re really the most involved at the start of the journey, and that’s it. You tell the agents what you want to know, and they do the clicking and even calling on your behalf. Rather than you going off on some online adventure, it’s the agent that’s hoovering up anything it can find and bouncing between different sites.