Google Cloud introduces a business agent that works with its own unique email address
The tool grants the software a separate identity to complete work and log actions.
9 Oct 2026
Google Cloud unveiled a new AI tool this week at a public event on Thursday. There are slight differences in the reported dates of the announcement among news sources. The company calls it the Gemini agent for enterprise users. It marks a move away from simple chatbots toward tools that finish jobs.
Unlike normal software, this agent obtains its own Workspace account and email address. It acts as a coworker with a unique digital identity inside the company directory. This design helps companies know exactly who performed an action at any time.
Regulators have asked about this kind of accountability for months regarding new tools. Google says it will allow administrators to approve what roles each agent can fill. Every step the AI takes is written to a record linked to the agent itself.
The system links directly to business data and tools like Slack or Jira. It reads from company databases and files stored in Microsoft 365 or Google Workspace. This allows the AI to understand specific work environments without manual setup for each new task.
Users can choose which artificial intelligence model handles each specific request within the platform. The system might pick a fast option for small tasks or Anthropic Claude for complex work. Teams can also set hard limits on spending to keep costs predictable.
Work continues running even after a laptop is turned off for the night. The AI remembers context and goals across different devices and screens during the process. It maintains this memory as long as the task remains unfinished or pending approval.
Accessible on phones, desktops, or through a command line interface anywhere in the office. Workers can tag it in chats or add it to group conversations for quick access.
Reports say early testers include PayPal and Shopify among others using the technology now. Google plans to support more open source models later this year for flexibility. The focus now remains on solving security and scale problems for large firms globally.