Google Introduces Gemini Agent for Enterprise Workflows

Google has launched a unified Gemini agent for businesses that can answer questions and carry out work from a single interface. The company says Gemini has more than 1 billion monthly active users, while nearly 90% of Fortune 100 businesses use Gemini Enterprise at work.

Google will initially make the agent available to business customers before a later consumer rollout. Sundar Pichai said that beginning with enterprise use should help the company address the harder problems of security, scale and performance associated with powerful agents.

From instructions to objectives

Thomas Kurian, chief executive of Google Cloud, said the new system can receive objectives, not merely instructions. It can plan work, use custom skills and tools, and connect with an organisation’s internal systems in pursuit of a goal.

Gemini selects a model by default, but users can choose one themselves. The initial third-party option is Anthropic’s Claude models; Google said the picker will later expand to open-source and other private models. Requests can also include attachments, including files, folders and workstream projects that combine files with skills.

Connections, oversight and an agent identity

The agent can connect with Google Workspace, Microsoft 365, Slack, Jira, Confluence, Git, BigQuery, Databricks, Postgres and Snowflake. It can also work securely with any Model Context Protocol server inside or outside a company network, extending the range of systems that can participate in an assigned task.

The product includes a tasks inbox where users can follow Gemini’s thinking process, delegation to subagents, loading of specialised skills, code and progress. Google says the agent has its own Workspace account, email address and context, allowing it to identify teams, time zones, approvers and calendar commitments. That direction follows Google’s expanding AI agent competition as Google intensifies the competition around AI agents.

Employees can invoke the agent by tagging it, emailing it, sharing material with it or adding it to a group chat. Actions are recorded in an audit trail attributed to the agent rather than to an individual. Access is planned through iOS and Android devices, Windows and Mac desktops, the command-line interface, Google Workspace, Microsoft 365, ServiceNow and Slack.

Early enterprise use and cost controls

Google named On, Shopify and PayPal as early testers. It also listed BNP Paribas, Bradesco, Merck, Orange Spain, Santee Cooper, SOMPO, Ulta Beauty and Wesfarmers among Gemini Enterprise customers.

Flexible spending options include multi-model orchestration, smart routing and real-time spend caps. For businesses, the practical implication is to define data access, approval paths and budget controls before delegating multi-step work to an agent connected to internal systems.