An AI assistant that can reach into your calendar and inbox is useful only when it finishes the job without turning you into its full-time supervisor. Google's new free access makes that question easier to test for people in the initial markets.
What actually happened
On 17 November, Google expanded Tasks in the Gemini app to people in the United States, Canada, Japan, South Korea, Australia and India, at no additional charge. It runs on Google's Gemini 4 Hydrogen model. Users switch from Chat to Task and assign multistep work instead of completing each browser or app interaction themselves.
Examples include researching options, coordinating a calendar and examining bills. Additional capabilities start in the US: calling businesses, using supported account sign-ins through Google Password Manager and checking out with Google Pay. Google says each task runs in a hardware-isolated virtual machine and sensitive steps such as purchases require confirmation or advance approval.
Why a small team cares
Picture a solo consultant coordinating several customer appointments every week. They need to compare incoming emails, find free calendar slots and confirm times with third parties. Delegating the discovery and scheduling steps could remove a small but recurring administrative burden.
The subscription change matters because a free trial no longer competes with the other paid tools on a small business's software bill. But the business result still depends on how many tasks finish correctly. A five-minute automated task that takes ten minutes to check is not a productivity gain.
Hype vs useful
Google's examples are demonstrations, not evidence that every website and third-party service works reliably. The company itself acknowledges that some tasks fail or hand control back to the user for the final interaction. Calling, purchasing and account access also have separate rollout and approval constraints.
Europe is excluded from the six-country launch. For a business based in Spain, this is currently a watch item rather than a feature to deploy. For an eligible solo operator, a narrow test is justified, with a normal manual workflow kept as fallback.




