If Google can’t make AI agents useful, maybe no one can

For years, tech companies have promised AI will give everyone a capable personal assistant but delivered something more like a clueless intern. Over the past six months, that has started to change, thanks largely to the viral open-source AI agent platform OpenClaw. And among the top AI labs now chasing similar success, one seems particularly well-poised to make agents succeed at a large scale: Google.

At I/O 2026, Google announced new AI agents for gathering information, planning events, summarizing your inbox and calendar, and more. The agents can run continuously in the background, and the company claims they’ll seamlessly integrate into Google’s own tools and external ones. It’s also expanding its developer tools and revamping Search with additional generative AI capabilities. Some are rolling out this week, and some will be available in the coming months, but the company’s strategy seems clear: adopt some of the features that have helped fuel OpenClaw’s success and amplify them with Google’s deep knowledge of our digital presence.

“Before this, I think AI agents were more of an idea in research,” Koray Kavukcuoglu, CTO of Google DeepMind and Google’s chief AI architect, told The Verge in an interview. This year, he hopes, they’ll be “really in our lives.”

AI agents have been a buzzword since just after ChatGPT’s launch in late 2022, but they remained mostly a science-fiction concept until the rise of OpenClaw, which has gained millions of users since its launch last November. OpenClaw let people chat with their agents via everyday apps like WhatsApp and Telegram, and (as long as a laptop was open) the agents could run around the clock. They performed well enough to handle basic tasks reliably, albeit with some clear flaws.

It made all the AI labs immediately sit up and take notice, but OpenAI was one of the first players to take action, acquiring OpenClaw (though it remains open-source) in February and hiring its creator Peter Steinberger. But Google’s existing empire of services gives it a major leg up. Where OpenClaw drove adoption by integrating with tools people already used, Google can do this too via MCP, but also build deeper links into its in-house suite of products, including Gmail, Drive, Docs, Photos, and Search. If anything, it’s surprising it took so long.

One of Google’s big bets this year is Gemini Spark, its new AI agent for consumers. Google promises Gemini Spark can perform tasks across Google’s own services and more than 30 external partners coming soon, including Dropbox, Uber, and Spotify. Gemini Spark is cloud-based; it can run 24/7 without keeping a laptop open and can sync across the web, Android, and iOS. The agent rolls out to trusted testers this week, and a beta will be available in the US next week on Google’s Ultra plan.

Google touts the typical uses for Gemini Spark, like shopping, researching, and coordinating with other people’s schedules and plans. Google also hopes people will find their own uses. Josh Woodward, Google’s Gemini app lead, says he’s been using Gemini Spark to plan a neighborhood block party, deploying agents to track RSVPs and what attendees are bringing, send reminders, and figure out when his homeowners’ association allows placing a giant inflatable. Outside Spark, Google is also introducing the Daily Brief, a morning update similar to OpenAI’s ChatGPT Pulse.

Gemini Spark isn’t available yet, but if it works the way Google says it does, it could be a big step forward for traditional tech companies’ AI agents. Google’s earliest agentic experiments completed tasks at a snail’s pace while hijacking your browser. By last year’s Gemini 3 release, its agents worked well for some jobs — like cleaning out an inbox — but still failed at others. Now, Google is taking a promising step by mirroring some key elements of OpenClaw: long-running agents that operate around the clock in the background, giving them the ability to have a lot more context about their tasks — and giving users the ability to text and email their agents directly.

Starting this summer, Google’s AI search, too, is getting agents — and promising to finally do more than eat up screen real estate and recommend pizza with glue. Its “information agents” are supposed to perform continuous background research — like tracking stock market shifts or weather for the best picnic day.

Google also announced an expansion to Antigravity, the agentic development platform it introduced about six months ago. A new standalone Antigravity desktop app will serve as a central hub for agent interaction, and the whole system is now designed as a platform to build and manage autonomous agents, Google says. The expansion follows on the heels of similar tools from OpenAI and Anthropic, which have tried to broaden their successful coding services to more approachable tools for non-programmers.

All this will be underpinned by a new model series: Gemini 3.5, whose initial entry Gemini 3.5 Flash should be available next month. The model is supposed to have significantly better coding capabilities than Gemini 3, which was released to great fanfare last November. It’s clearly intended to leapfrog updates from Anthropic, which is known for its coding prowess, and OpenAI. Gemini 3.5 Flash is especially good “when deploying multiple agents simultaneously and completing long-running tasks,” Kavukcuoglu told reporters Monday. It’s also supposed to be four times faster than other frontier models and less than half (or in some cases, one third of) the price — a fact that’s relevant to 24/7 AI agents, with token costs that quickly add up.

In the world of AI agents, Google will still be playing catch-up with the one-man team behind OpenClaw. But it’s a long-standing frontrunner in the AI race, and its app has the benefit of scale: It now serves more than 900 million users per month, executives told reporters on Monday, in more than 230 countries and more than 70 languages. Compared to dedicated AI companies under increasing financial pressure, it’s able to at least temporarily subsidize costs to attract users. And while its agents haven’t yet had to weather the real world, they’re headed in a promising direction. If any AI company can make agents truly useful, it’s Google. If it can’t, it won’t have many excuses to fall back on — and the whole idea might need a rethink.