Gemini enters the era of autonomous business agents

Google is looking to go beyond just conversational experiences with this move. The new unified agent will no longer just answer users' questions, but will also independently solve tasks, write software, and coordinate meetings and travel. As the technology enters the business environment, caution will be needed, especially regarding security and reliability.

Companies will be able to give the system specific goals instead of just instructions. The agent will plan its own work, use personalized knowledge, and connect with internal systems. Users will be able to choose from a variety of models, including external solutions such as Anthropic's Claude, with support for open source and proprietary models coming soon.

The system will be able to connect to a number of tools, such as Google Workspace, Microsoft 365, Slack, Jira, Confluence, Git, BigQuery, Databricks, Postgres, and Snowflake. It will also support connecting via Model Context Protocol servers. The work will be monitored in a special task inbox, where thought processes and progress on tasks will be visible.

The system will be unique in that the AI will receive its own Workspace account and email address. It will act as an autonomous collaborator, aware of employees' schedules and time zones. It will be available on mobile devices with iOS and Android operating systems, computers with Windows and macOS operating systems, and via the command line. Among the first testers were Shopify, PayPal, and the On brand.