Google pushes post-search AI agents

Google used its May 19 I/O conference to present AI agents as a core consumer interface, not a side feature. The company tied that push to Search, Gemini, shopping, video calls and Android XR glasses, with executives describing a product set built to browse, compare, summarize and take actions for users. Google’s official I/O roundup said Gemini 3.5 Flash combines “frontier intelligence with action,” while a separate company post said Gemini-powered eyewear is due this fall. Outside observers described the move as an attempt to move users further from classic search pages and closer to AI-mediated workflows. The Verge said I/O marked the start of a “post-search AI era,” and TechCrunch wrote that Google was pitching consumers an “AI agent ecosystem” for using the web. Those descriptions came as Google tried to show the same AI logic across multiple surfaces rather than in a single chatbot window. (blog.google) ### Where did Google put the agent pitch? Google’s May 20 I/O recap listed new models, agents and tools spanning search, shopping, creation and developer products. The company said Gemini Omni can work from multiple inputs, starting with video, and said Gemini 3.5 Flash is designed for action as well as reasoning. In the same post, Google said advances to its “agent-first” development platform, Google Antigravity, were meant to move beyond tools that help users write toward agents that help them act. (techcrunch.com) TechCrunch reported on May 19 that Google also introduced agentic capabilities in Search that let users create, customize and manage multiple AI agents around topics of interest. A separate TechCrunch analysis on May 21 said the consumer pitch was promising but confusing, because Google was presenting not just one assistant but a broader system for navigating the web through agents. (blog.google) ### What did Google show beyond Search? CNET’s live recap said Google highlighted new Gemini capabilities, AI agents and an experimental video AI agent inside Google Beam. That demo turned Beam from a telepresence showcase into a setting where users could interact with an AI avatar on video, according to CNET’s account. Google’s broader I/O summary also pointed to “Universal Cart” and other shopping-related tools, reinforcing that the company was demonstrating action flows, not only answers. (techcrunch.com) TechCrunch said that approach could let consumers ask software to handle browsing and comparison work that previously happened across websites and apps. ### Why were glasses and device partners part of the same message? (cnet.com) Shahram Izadi, Google’s vice president and general manager for XR, said in a May 19 company post that Gemini is being built into glasses, headsets and other Android XR devices. Google said the platform is being developed with Samsung and Qualcomm, and that intelligent eyewear with Gemini is coming this fall. The company named Gentle Monster and Warby Parker among eyewear partners. (blog.google) Sameer Samat, president of the Android ecosystem, said in Google’s Android Show recap last week that developers would get a preview of glasses launching later this year. That tied the glasses announcement to the same I/O product cycle as Gemini updates and broader Android changes. ### Why does this matter for startups that live on the web? TechCrunch’s May 21 analysis said Google was pitching a new way for consumers to use the web through AI agents rather than direct site visits. (blog.google) The Verge’s I/O coverage framed that as a post-search shift, where discovery increasingly happens inside AI layers instead of traditional result pages. (blog.google) That matters because Google’s demos centered on upstream interfaces that can browse, compare and summarize before a user reaches a standalone app or service. Google’s next public milestones include the fall launch window for Gemini-powered eyewear and the wider rollout of the I/O features listed in its May 20 announcements roundup. (blog.google) (techcrunch.com)