Until now, we have used generative AIs as simple single-use assistants. You write a command, the platform returns a text and that's where the interaction ends. However, a recent leak discovered by TestingCatalog reveals that Google is testing a new agent tab within its environment Gemini Enterprisemarking a clear leap towards autonomy.
The underlying idea is to move from isolated responses to complex workflows. The new interface allows you to create, manage and deploy multiple agents focused on different tasks. You will no longer ask for a simple draft, but you will assign a broad objective. The system will divide the mission into logical steps, accessing your folders and connected applications to complete the job with as little intervention as possible.
Comprehensive automation, but with human review
This revamped workspace includes a task pane, inbox, and goal-setting sections, programming reusable skills and define recurring schedules. With this aggressive movement, they seek to compete directly against other corporate productivity proposals, such as: Claude Cowork from the firm Anthropic.
Granting full access to a computer program over our corporate files generates obvious risks. Aware of the danger, Google developers have implemented a kill switch mandatory human review. This feature ensures that the system stops and waits for the user's green light before performing sensitive actions, an essential safeguard against AI errors in a strict work environment.
It remains to be seen whether, in practice, this tool will be a true revolution or whether it will end up hindering the workers' routine. Honestly, we had not seen such an interesting novelty since Google integrated the possibility of create songs in Geminia leap that now goes from the creative field to pure productivity.
Would you let an artificial intelligence take control of your work tools without your constant supervision?
