Google Unveiled a Lot of AI Agents at I/O. It Just Forgot to Explain Why Anyone Should Care.
Google's I/O developer conference this week was full of AI announcements — more, perhaps, than at any prior event. But the most revealing thing about Tuesday's keynote wasn't what Google showed. It was who it forgot to show it for.
The centerpiece of Google's consumer AI push is Gemini Spark, a personal AI agent that lives across Gmail, Google Docs, and Google Workspace, surfacing insights from newsletters, managing home inventory, and coordinating group logistics. Alongside it, Google unveiled information agents — a souped-up reinvention of Google Alerts that runs 24/7 in the background, tracking market trends, price changes, and weather events in real time. A third feature, Daily Brief, compiles a personalized digest from a user's inbox, calendar, and task list each morning.
New Google AI Agent Products Announced at I/O
Gemini Spark — personal AI agent across Google's productivity suite
Information Agents — always-on AI reinvention of Google Alerts
Android Halo — notification hub for Spark activity on Android
Daily Brief — AI morning digest from Gmail, Calendar & Tasks
Agentic Chrome — conversational AI layer for web browsing
On paper, these are genuinely interesting products. The problem is accessibility. Most of these features are being rolled out exclusively — or first — to subscribers of Google Ultra, Google's new $100-per-month tier. Information agents and Spark go to Pro and Ultra subscribers this summer. Android Halo arrives "later this year." Daily Brief is the most broadly available, reaching Plus, Pro, and Ultra subscribers in the US. Free users, the vast majority of Google's audience, are largely left out — at least for now.
Google's early days were built on giving everyone free access to transformative tools. In 2026, its most transformative tools cost $100 a month.
The irony runs deep. Google's historic strength was democratizing access: free search, free email, free maps. The products that built Google's cultural footprint were given away. Now the company is asking consumers to subscribe to a premium tier to access features it acknowledges are still being iterated on. The stated rationale — that heavy users will push the limits and improve the product — is a reasonable engineering argument. It's a poor consumer pitch.
The messaging at I/O didn't help either. Between presentations, Google flashed AI-generated imagery and aired an animated segment featuring talking AI chips. In the Android glasses demo, a presenter's photo of the audience was modified in real time to add a floating blimp and sent to their smartwatch. The audience reaction was polite. The implicit question — is any of this worth the societal cost of new data centers and infrastructure being built to support it? — went unaddressed.
There's a version of this story that lands very differently. AI agents that take over monitoring, research, and scheduling tasks could give people back something genuinely scarce: time offline. Framed correctly, Spark could be positioned as the tool that reduces screen time, not increases it — especially appealing to younger users who are already pushing back against always-on digital culture. Google nodded at free-user access being on the roadmap but gave no timeline.
Meanwhile, messaging-first AI startups like Poke, Poppy, and Wingman are building simpler interfaces — AI agents accessible through everyday text messaging — that sidestep the complexity entirely. Google has the scale and the infrastructure. What it's missing, at least based on Tuesday's showing, is the clarity of purpose that makes someone outside the tech industry want to use what it's building.