Google on Thursday unveiled an artificial intelligence agent built for office work, taking its entry in an increasingly crowded race among technology giants to sell virtual assistants that can carry out tasks on their own.
The agent, called Gemini, can write and run computer code, create images and other media, and plug into Google’s own products, including Gmail, Docs, Drive and Sheets, as well as outside tools like Slack and Microsoft 365, the company said in a statement, as CBS News reported.
Thomas Kurian, the chief executive of Google Cloud, introduced the product in a keynote address at Gemini at Work 2026, the company’s annual event for AI products. He described Gemini as a “universal agent” that can act like a co-worker: employees delegate a task to it and get back finished work.
“Gemini can function as your personal assistant or as a team member, where it works on behalf of a group of people, like a project manager within a team, or on behalf of a specific role in an organization, like an analyst in your finance department,” Mr. Kurian wrote in a company blog post.
For jobs too large for one agent, Gemini can generate what Google calls “sub-agents,” described in the company statement as “temporary, job-specific agents, each with their own identity,” that can work together on a goal over several days. The agent runs in the cloud and builds memory and skills from the tasks it is assigned and the questions users ask it. “You never have to re-brief it or wonder which machine you told it to do something on,” Mr. Kurian wrote.
The market for such agents has filled in fast. Meta brought out Muse, a personal AI assistant, in September and says it has been downloaded millions of times. OpenAI followed late last month with Dots, an “AI helper” available to ChatGPT users on Pro and Business Premium plans. At Google, the agent arrives a week after the company announced Gemini 4 Argon, its first frontier model in months, which it is initially offering only to vetted cybersecurity partners.