Google I/O shifts agentic AI

Google used its May 19 I/O keynote to put “agentic” AI at the center of its product lineup, tying new features to Search, the Gemini app and developer tools rather than introducing a stand-alone moonshot. Sundar Pichai called it the “agentic Gemini era” in Google’s official keynote transcript, while product posts described a shift from prompts to actions across consumer and enterprise software. Google also made Gemini 3.5 Flash the default model in AI Mode in Search globally and said the model is available across the Gemini app and developer platforms. ### Which announcements showed Google’s AI strategy most clearly? Google’s clearest product signal was Search. In a May 19 post, the company said AI Mode is now powered by Gemini 3.5 Flash by default for users globally, and described the update as the biggest overhaul of the Search box in more than 25 years. The same post said new Search features are built around AI agents that can help complete tasks, not just return links. (blog.google) Gemini 3.5 Flash was the other anchor. Google said the model combines “frontier intelligence with action,” runs faster than earlier systems and is intended for “real-world agentic workflows.” In a separate product note, Google said 3.5 Flash is now the default model for the Gemini app and AI Mode in Search, with 3.5 Pro due next month. (blog.google) ### Why did Google keep tying AI to Search instead of splitting it off? Axios reported on May 21 that Google is trying to disrupt its own products with AI while still protecting the businesses that generate most of its profit. That balancing act is most visible in Search, where Google is inserting AI features into its core product instead of replacing the business model outright. (blog.google) Google’s own rollout supports that reading. The company did not present agentic AI as a separate experimental lab project; it embedded the features inside Search, Gemini, Android Studio, Google AI Studio and enterprise products. That distribution model leans on Google’s existing scale rather than asking users to adopt an entirely new destination. (axios.com) ### What does “agentic” mean in the products Google actually shipped? Google’s language at I/O focused on software that can take actions on a user’s behalf. In its keynote materials, the company said it had moved “beyond AI tools that just help us write, to agents that help us act,” and described Google Antigravity as an “agent-first” development platform. (blog.google) Search executive Liz Reid, as quoted by Ars Technica’s coverage of the event, framed the changes as part of Google’s next step in AI search. Google’s product posts similarly described AI Mode and Gemini 3.5 Flash as infrastructure for task completion, coding and workflow execution, not only text generation. (blog.google) ### How did investors respond after the keynote? Alphabet shares slipped after the event as traders weighed the commercial payoff from the new AI push against execution and regulatory risks. TS2 reported that Alphabet’s Class A shares fell 0.4% to $387.66 after the regular Nasdaq session on May 21, while Class C shares were also down 0.4%. (arstechnica.com) A separate pressure point came from Europe. TS2 said investors were also digesting a new complaint over scam ads, adding a regulatory overhang to the post-I/O reaction. CNBC had framed the conference before it began as a test of whether Google could turn its AI narrative into something convincing for Wall Street. (ts2.tech) ### What comes next after I/O? Google said on May 19 that Gemini 3.5 Pro is already being used internally and will roll out next month. The company also said Gemini 3.5 Flash is already live across Search AI Mode, the Gemini app, Google AI Studio, Android Studio and enterprise offerings, making the next phase less about unveiling a new concept than expanding the products already announced. (ts2.tech) (blog.google)