Google Cloud has introduced Gemini Agent, a universal AI assistant designed to carry out workplace tasks across multiple applications, as technology companies expand their focus on agentic artificial intelligence for enterprise operations.
Announced at the Gemini at Work 2026 event on October 8, the new agent allows employees to assign tasks through a single interface, including preparing documents, analysing information, creating content and writing or executing code.
Unlike conventional AI chatbots that primarily respond to queries, Gemini Agent is designed to complete multistep assignments by accessing relevant business systems, selecting tools and coordinating workflows.
According to Google Cloud, the agent can operate across Google Workspace applications, including Gmail, Drive, Docs, Sheets, Slides, Chat and Calendar. It also supports connections with external enterprise platforms such as Microsoft 365, Slack, Salesforce, ServiceNow and Jira.
The system maintains context across devices and applications, allowing users to continue assignments without repeatedly providing the same information. Tasks can also continue running in the cloud after a user closes their device.
Google Cloud has introduced multi-agent orchestration capabilities, enabling Gemini Agent to create and coordinate specialised AI agents for different stages of an assignment. These agents can work sequentially or simultaneously, depending on the requirements of the task.
The company is also introducing coworker agents that can function as members of enterprise teams. These agents can receive dedicated identities, email accounts and access permissions, allowing employees to assign responsibilities through familiar collaboration tools.
For marketing teams, the capabilities could support activities such as preparing campaign documents, coordinating launch schedules, analysing performance data and managing information across departments.
Google Cloud said Gemini Agent can select different underlying AI models depending on the task. Its model orchestration capabilities currently include Google's Gemini models and Anthropic's Claude models, with additional options planned.
The company has also introduced cost-management features, including automated model routing and spending limits, alongside enterprise security controls covering identity management, permissions and governance.
Google Cloud CEO Thomas Kurian said the company is positioning Gemini as a single agent capable of handling knowledge work, content creation and coding through one interface.
Google reported that nearly 80% of its Cloud customers currently use its AI products, while almost 90% of Fortune 100 companies use Gemini Enterprise. The figures were disclosed as part of the company's announcement.
The new Gemini Agent is initially available to enterprise customers through a private preview, with broader availability details yet to be announced.
The launch comes as enterprise technology providers increasingly develop AI systems capable of executing business processes rather than only generating responses. Google Cloud's latest offering brings these capabilities into existing workplace applications, with an emphasis on connected workflows, enterprise controls and task completion.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.