Google has announced a new agentic AI experience for Gemini that can go beyond answering questions to plan and carry out tasks on behalf of users, initially targeting businesses before expanding to consumers.
The announcement was made Thursday at a Google Cloud event, as technology companies increasingly develop AI agents capable of handling multi-step tasks, writing code, coordinating meetings and interacting with other software.
Google CEO Sundar Pichai said Gemini has more than one billion monthly active users, while nearly 90% of Fortune 100 companies use Gemini Enterprise at work.
Gemini agent can work towards objectives
Google Cloud CEO Thomas Kurian said the new agent can be given “objectives, not just instructions”. It can plan work, use specialised skills and tools, and connect to internal business systems to complete assigned tasks.
The agent can access services including Google Workspace, Microsoft 365, Slack, Jira, Confluence, Git, BigQuery, Databricks, Postgres and Snowflake. It can also connect securely to Model Context Protocol (MCP) servers inside or outside an organisation’s network.
Users can provide files, folders and other project materials as context for specific tasks.
Gemini will generally select the model best suited to a task, although users can choose a model themselves. Google said the initial selection will include Anthropic’s Claude models, with open-source and other private models expected to be added later.
AI agent gets its own workplace identity
The new system will have its own Google Workspace account, including an email address and separate context, allowing it to operate more like a digital colleague within an organisation.
According to Google, the agent can use workplace context such as team structures, time zones, approval requirements and calendar information to help carry out tasks.
Employees will be able to interact with it by tagging it, emailing it, sharing files or adding it to group chats. Actions taken by the agent will be recorded in an audit trail attributed to the AI rather than an individual employee.
A dedicated tasks inbox will allow users to monitor progress, see how work is delegated to subagents and track the use of specialised skills and code.
Businesses get cost controls and multiple access points
Google said the agent will be available through iOS and Android devices, Windows and Mac desktops, the command line interface, Google Workspace, Microsoft 365, ServiceNow and Slack.
Early testers included sportswear company On, e-commerce platform Shopify and payments company PayPal. Google also identified BNP Paribas, Bradesco, Merck, Orange Spain, Santee Cooper, SOMPO, Ulta Beauty and Wesfarmers among its Gemini Enterprise customers.
To help businesses manage spending, Google is introducing flexible options that include multi-model orchestration, smart routing and real-time spending limits.
Google prioritises enterprise adoption
Google is initially focusing on businesses to address the security, performance and scalability challenges associated with more autonomous AI systems, Pichai said.
The strategy comes as AI companies compete to move beyond conversational assistants towards agents that can execute tasks across applications and workplace systems.
Google plans to bring the experience to consumers at a later stage, although the company did not specify a date for the wider rollout in the announcement.
