Google announced a significant expansion of Gemini's agentic capabilities at its Android Show event, moving the assistant from single-app interactions into genuine multi-step, cross-app task execution. The headline feature lets users press the power button, describe a task, and have Gemini handle the entire process — copying a grocery list from a notes app and adding items to a shopping cart, for example — using whatever is currently on screen as context and waiting for final confirmation before completing any transaction. A web-browsing feature that allows Gemini to autonomously book appointments and complete online tasks is also moving from experimental to Android-wide availability, alongside a form-filling capability that draws on Google's Personal Intelligence system to populate fields using information the user has previously shared. Gemini is also coming to Gboard through a feature called Rambler, which transcribes speech in the user's own tone while removing filler words and formatting the output — a dictation tool that competes directly with standalone apps in that category.For freelancers specifically, the cross-app task execution and form-filling features represent a meaningful reduction in the administrative friction that consumes hours of every working week.

Client onboarding forms, invoice submission portals, project management platform updates, and repetitive data entry across disconnected tools are exactly the category of multi-step, low-creativity work that Gemini's new capabilities are designed to handle. A freelancer who can describe a task in natural language and have their phone execute it across multiple apps while they focus on billable work gains a real productivity advantage over one still doing those tasks manually. The features are rolling out to Samsung Galaxy and Google Pixel devices first this summer before expanding to other Android hardware later in the year.