Google is continuing its all-out push toward an era in which the search engine no longer merely returns links to websites, but becomes a personal AI agent.
At its annual developers conference, Google I/O 2026, held Tuesday evening Israel time at the company’s headquarters in Mountain View, California, Google unveiled a long list of announcements — from an upgraded AI-powered search engine to smart agents and new Gemini-based AI models.
Google’s message is clear: The company wants to transform itself from a search engine into an artificial intelligence company that manages users’ digital lives. Along the way, it is also refreshing the design of the Gemini app and making online shopping easier than ever.
At the center of the event were four especially significant announcements: the new and powerful Gemini Omni model, a redesign and upgrade of Google’s intelligent search engine, smart glasses expected to arrive this fall and a unified shopping cart that will allow AI agents to make purchases for users with almost no need to touch a smartphone.
Google Search undergoes a deep transformation
A year after introducing “AI Mode,” Google announced a dramatic change to its search engine. Instead of the classic search bar, Google is moving to what it calls the AI Search Box, a new search box based entirely on artificial intelligence.
According to the company, the new box can understand context far more deeply, complete thoughts more intelligently than familiar sentence completion tools and enable searches using almost any format: text, images, videos, files and even open tabs in Chrome.
Google's new AI-powered search engine
(Video: Google)
But the truly interesting part is the shift into the world of AI agents. Google announced Information Agents — smart agents that can track topics for users and update them automatically.
For example, a user looking for a four-room apartment for rent with a budget of up to 6,000 shekels in the Ra’anana area could define the requirements for the agent, which would regularly scan the web and send alerts when relevant listings are published.
Gemini Omni: The new model that wants to create everything
One of the conference’s biggest announcements was Gemini Omni, Google’s new AI model, which can generate almost any type of content, from text to full video. Google describes it as a major leap in the company’s multimodal capabilities, with an emphasis on a deeper understanding of the world and especially advanced editing abilities.
Google's new Gemini Omni
(Video: Google)
According to the company, Gemini Omni allows users to combine text, images and videos in the same creative process, including uploading files directly from the gallery and editing video through a free-flowing conversation with the model. Google says the system can maintain consistency of characters, voice and identity throughout the entire editing process.
Video created by Gemini Omni
(Video: Google)
Google also announced Gemini 3.5 Flash, a new model designed to combine especially high speed with performance on par with the company’s flagship models. Google says it is its strongest AI agent and coding model to date.
Smart glasses are back, this time with Gemini and Samsung
Years after Google Glass became an industry punchline, Google is once again trying to enter the smart glasses market. This time, it is doing so with Gemini and with a much clearer focus on everyday use.
The new glasses, developed in cooperation with Samsung, Qualcomm and the fashion brands Gentle Monster and Warby Parker, will come in two models: one with a voice interface only, and another with an internal display that will show information in front of the user’s eyes. The first model is set to launch this fall.
According to Google, users will be able to receive navigation directions, send messages, take photos, translate text in real time and even perform complex tasks using Gemini, without taking their smartphone out of their pocket.
The glasses will also allow users to ask questions about the world around them. For example, they could receive reviews of a restaurant whose facade they are looking at, understand what is written on a sign in a foreign language or even identify phenomena in the sky. Google has also integrated AI-based photography and editing capabilities, including commands such as adding funny hats to people in a photo using the popular “Nano Banana” tool.
One shopping cart for everything
Perhaps the most ambitious announcement at the conference was in the field of shopping. Google unveiled a new “unified shopping cart,” aimed at turning Gemini into a personal shopping agent that works for the user in the background.
The new cart will be connected to nearly all Google services: Search, Gemini, Gmail and even YouTube. Users will be able to add products to the cart from any of these places, and the system will begin working on its own: looking for price drops, identifying deals, checking price histories and sending updates when a product comes back in stock.
Google's smart shopping cart
(Video: Google)
Google stresses that the system can also understand complex contexts. For example, if a user is building a gaming computer from different components, the cart will check that the parts are compatible and suggest alternatives if needed. It will also be connected to Google Wallet and know how to recommend the payment method or loyalty club that will provide the best discount.
At the end of the day, behind all these announcements lies a deeper shift. After years in which Google was the place where people searched for information, it is now trying to become a system that understands what users need before they even search for it, and carries out actions for them in the real world.
The big question is whether users are truly ready for this new world.