Google’s New Gemini Agent Can Handle Work Across Multiple Apps

Google has unveiled Gemini Agent for businesses, an AI assistant capable of handling tasks across multiple apps and collaborating in team discussions. At the Gemini at Work 2026 event, Google showcased how the enterprise AI agent can complete assignments without requiring employees to manually switch between workplace applications.

Employees can delegate multi-step work to the agent, which can participate in team conversations, draft reports, and coordinate schedules. The agent can also create temporary AI agents to tackle individual parts of complex assignments, preserving context for later continuation. Businesses can invite the agent as a coworker in Google Chat, assigning it a Workspace account, email address, and calendar.

The agent can identify relevant participants from past team conversations and coordinate schedules, with suggestions appearing under its name in Google Docs. It can also prepare project-update presentations based on email requests. The agent connects with Microsoft 365, Slack, and other business systems for seamless automation. For complex tasks, Gemini's cloud-based execution allows work to continue even after the user closes their laptop.

However, the official product FAQ indicates that features like multi-step background delegation, mobile and desktop access, and third-party model selection are currently available to a limited number of businesses in early access. Google has not set a firm timeline for wider availability. Despite these capabilities, businesses must carefully consider how much work the agent should perform independently, as mistakes or errors could have significant consequences.

Employees, managers, and IT administrators all play a role in ensuring the agent's actions align with company protocols and do not introduce additional risks or workloads. Google offers a free 30-day trial of Gemini Enterprise Plus, but access to all announced features is not guaranteed.