Google Cloud introduces Gemini agent for work as AI race heats up

Alphabet's Google Cloud unveiled the Gemini agent on Thursday, a single AI tool designed to perform work-related tasks. The agent can answer questions, manage tasks, generate content and write code, reflecting the intense competition among tech giants to offer autonomous AI solutions. This announcement follows OpenAI's September launch of always-on agents called dots, which autonomously pursue user goals across multiple applications.

Earlier this month, Meta released Muse, a personal AI agent capable of handling shopping, booking travel, sending emails and making payments for users.

The Gemini agent plans work autonomously, utilizes tools, integrates with a company's business systems, and delivers finished work in various formats such as documents, emails, and within developer environments. The agent operates directly within Google Workspace applications, including Gmail, Docs, Sheets, and Calendar, as well as through Microsoft 365 and Slack.

It automatically selects the most suitable model for each task, currently leveraging Google's Gemini models and Anthropic's Claude models, with additional models to be integrated in the future.

Users can create "coworker agents" that function as team members, possessing their own email addresses and access to information as specified. Previews of versions tailored for financial services and legal work are now available, with versions for government, healthcare, and retail sectors expected to follow soon.