Gemini’s Task Button: Who Gets It and What It Does

Google put a Task option right below Gemini’s chat box on October 13. That sounds small because it is. Gemini could already take on tasks. What changed is that eligible subscribers can now make the jump from chatting to assigning an errand without hunting for a separate place to do it.

That’s more interesting than another announcement about what an AI model can supposedly do on a test nobody outside the industry has heard of.

At gemini.google.com, select Task below the text box and give it a specific job. You could ask it to sort through newsletters or research options and put the results in a file. The task has its own thread where you can see progress. Google says it can keep working while you’re away.

Notice the wording: keep working, not finish anything you throw at it without bothering you. If a website needs a sign-in or another step Gemini can’t handle on its own, it may ask you to take over. Shopping and reservations are among Google’s examples, but that doesn’t mean every connected service can complete every purchase or booking for you.

Access is the part worth checking before you start looking for the button. Tasks require a personal Google Account, an age of at least 18, and a Google AI Pro or Ultra subscription. You also need Keep Activity turned on and must opt in to Personal Intelligence. Free accounts, AI Plus, and work or school accounts don’t qualify. Google lists the web app, Gemini mobile apps, and Mac app as supported, but tasks aren’t available in the European Economic Area, Nigeria, Switzerland, or the United Kingdom for now.

In the U.S., Google lists AI Pro at $19.99 a month. That’s a lot to pay just to see whether a button can manage your newsletters. Usage limits vary by plan, too.

And please don’t treat “assign a task” as “stop paying attention.” Google warns that tasks can make mistakes and says not to put passwords, payment details, or sensitive information into a task thread. Some work needs your permission or confirmation. If a task uses your local Chrome browser, your computer and browser need to stay awake, though a remote browser may handle some work when your device isn’t available.

My read: this is a sensible improvement to an existing feature, not an AI assistant that can quietly run your life. If you already pay for Pro or Ultra, try it on a boring job whose results you can check. If you don’t, the easier-to-find button alone isn’t much of a reason to subscribe.

Google described the change in its October 13 announcement; SiliconANGLE also reported the chat-to-task switch and paid-plan rollout.

Information checked: October 14, 2026.