Incident sheet · /enterprise-ai
- System
- Enterprise AI
- Filed
- Read
- 3 min · 518 words
- Severity
- Routine
Google's Gemini Can Now Operate a Mac Untended
Google's Gemini desktop app for macOS can now control the machine itself, handling multi-step tasks across local files, apps and the web — in U.S. English, behind a paid Google AI subscription.

Event log
- Google announced computer use in Gemini's macOS desktop app on Thursday
- The feature requires a Google AI Plus, Pro or Ultra subscription and is U.S. English only
- OpenAI, Anthropic and Meta already offer comparable agentic desktop features
- Google's 'Tasks' agent feature began its broader rollout only this week
- The revamped 'CC' family agent, announced in September, just started admitting early testers
Google announced on Thursday that Gemini's desktop application for macOS can now operate the computer itself, navigating native Mac apps and browsing websites in the background to complete tasks on the user's behalf.
The launch puts Google in catch-up position. OpenAI, Anthropic and Meta have shipped comparable agentic features in their desktop applications for some time. Google only this week announced the broader rollout of Gemini's "Tasks," the in-app feature that lets the AI act as an agent — signing into websites, making online purchases, booking appointments. It also only recently began admitting early testers to its revamped "CC" agent, announced in September, which is built to keep families organized across email, calendars, chats and tasks.
What does computer use actually do?
Computer use is a distinct class of agentic capability: the AI takes actions directly on the machine rather than merely answering questions. In Gemini's case, the machine is a Mac. The agent executes multi-step tasks that span local files, installed applications and the web.
The mechanics are straightforward. The user states what they want done. A live preview of the agent's work then runs inside the chat window, while the screen and cursor stay free for other work. Users can pause, stop, and approve Gemini's actions before they happen — the same control pattern every vendor in this category has settled on.
Google's suggested use cases are the tedious ones:
- Filling out forms
- Repeating the same task across a batch of files
- Organizing files into folders based on specific criteria
- Creating slideshows from files or from information in online dashboards
That list reads like a worklog, not a wishlist. Form-filling and batch file organization are exactly the chores that cost organisations real hours, which is presumably the point.
What does this change for IT departments?
Availability is narrow. The feature ships in English in the U.S. only, requires the Gemini app for Mac, and sits behind a Google AI Plus, Pro or Ultra subscription. Google did not say when it would reach other markets or platforms. No enterprise tier was mentioned.
So the immediate answer is: nothing yet, formally. The longer answer is that an agent granted cursor-level control of a local machine is now a consumer-grade Google product, not a developer preview. IT teams that spent the last year writing policy around OpenAI and Anthropic agents now have a third vendor to account for, arriving through a personal subscription rather than a procurement channel.
The lag is the story. Google commands the models and the distribution, yet its agentic rollout — Tasks this week, CC testers only recently, computer use now — trails rivals who moved months earlier. Thursday's announcement closes a gap rather than opening a lead. For shops evaluating desktop agents, that means one more logo on the comparison chart, and one more approval dialog to model in policy.
Worth watching: whether Google extends the feature beyond English, beyond the U.S., and beyond the Mac. Until it does, this is a US-only convenience feature wearing an enterprise problem's clothes.
via techcrunch.com (Source)
More from Tom Whitfield
Show full bio
Market editor covering people, launches and day-to-day practice in enterprise information technology and the leaders who run at CIO Monitor.
188 articles
The newsroom digest
One email a day with the stories that matter. No noise.
Cross-references
- Cognizant expands Google Cloud deal around Gemini Enterprise and Workspace
- Cognizant Expands Google Cloud Partnership Around Gemini Enterprise
- Anthropic launches Claude Meetings as workplace race intensifies
- AG-UI standardizes the last-mile between agents and users
- Otter Ships Agentic Work Mode, Days After $320M Raise