Google brings agentic AI to Gemini, starting with businesses

Google revealed a unified, agentic version of its Gemini AI at a Google Cloud event, positioning the system to not only answer queries but to carry out tasks on a user’s behalf from a single interface. The company said it will prioritize rolling the agent out to businesses first to tackle enterprise challenges around security, scale and performance before expanding to consumers. According to Google executives, the new agent accepts objectives rather than step-by-step instructions: it can plan workstreams, call on custom skills and tools, access company data, and connect to internal systems to complete tasks. By default Gemini selects what it considers the best model for a job, though users can override that choice and pick from other models, including Anthropic’s Claude models; Google plans to add open-source and private models over time. The agent can take attachments and structured project material as input and integrate with a wide set of business systems, such as Google Workspace, Microsoft 365, Slack, Jira, Confluence, Git, BigQuery, Databricks, Postgres, and Snowflake. It can also work securely with any Model Context Protocol (MCP) server located inside or outside a company network, the company said. Users will monitor activity through a “tasks inbox” that shows the agent’s reasoning, task delegation to subagents, skill loading, code and progress. Uniquely, each agent receives its own Google Workspace account and email address so it functions like another co-worker—aware of teams, time zones, approvers and calendars—and produces an audit trail attributed to the agent rather than to individual employees. Google said the agent will be accessible across mobile and desktop platforms, command line, Google Workspace, Microsoft 365, ServiceNow and Slack. Early testers include sportswear brand On, Shopify and PayPal, and Google listed a number of Gemini Enterprise customers such as BNP Paribas, Bradesco, Merck, Orange Spain, Santee Cooper, SOMPO, Ulta Beauty and Wesfarmers. Google also said it will offer flexible spending features—multi-model orchestration, smart routing and real-time spend caps—to help businesses manage enterprise AI costs.