Google launches Gemini agent for businesses at Cloud event

Google Cloud introduced the Gemini agent for work on 8 October 2026 at the Gemini at Work 2026 event, describing it as a single, universal agent that plans and completes tasks rather than only answering questions. The company said the agent can be given an objective, plan the steps needed, and return finished work inside the documents, inbox, and developer environments employees already use, connecting to a company's business systems from one prompt box.blog.google+1

The launch matters because Google is targeting businesses before consumers, following the rise of consumer-facing agents like Meta's Muse and OpenAI's Dots. Google CEO Sundar Pichai said focusing on businesses first lets the company solve harder problems around security, scale, and performance before rolling the agent out more widely. Gemini today has over 1 billion monthly active users, giving Google a large base to build on.theregister.com

What the agent does

Thomas Kurian, CEO of Google Cloud, said users give the agent objectives, not just instructions. It plans the work, uses custom skills and tools, connects to internal systems, and brings back something finished. The agent can connect to Google Workspace, Microsoft 365, Slack, Jira, Confluence, Git, BigQuery, Databricks, Postgres, and Snowflake, and can work with any Model Context Protocol server inside or outside the company's network.

Users track the agent's work through a tasks inbox showing its thinking process, delegation to subagents, loaded skills, code, and progress. The agent can create temporary, job-specific sub-agents with their own identities to tackle multi-step tasks in parallel or sequentially. It runs in the cloud, so it keeps context and can continue tasks for hours or even days after an employee leaves their computer.

By default the agent picks the best model for the task, but users can choose a model themselves, starting with Anthropic's Claude models. Kurian said Google orchestrates across its Gemini family and Claude models today, with other leading private and open models to come, to deliver quality and lower costs. Google said flexible spending options including smart routing and real-time spend caps will help control enterprise AI costs.americanbazaaronline.com

How model choice works

The agent gets its own Workspace account, email address, and context, functioning like a digital co-worker. It knows who is on which team, their time zones, who approves items, and what is on calendars. As it acts, it writes an audit trail attributed to the agent rather than a person. Kurian said every agent has its own identity, separate from the user's, and that policies and budgets set once apply across agents.

Pichai told the event that nearly 80% of Google Cloud customers use its AI products, and nearly 90% of Fortune 100 businesses use Gemini Enterprise. Nearly 500 enterprise customers each processed over 1 trillion tokens on the platform in the past year. Pichai also said Gemini 4 Argon scores 77.9% on the DeepSWE v1.1 coding benchmark, outperforming Claude Opus 5.5, Claude Fable 5.1, and GPT-6 Astra.

Kurian cited customer gains, saying Japanese telecom NTT Docomo saved 450,000 hours annually using data agents. Early testers of the agent included sportswear brand On, Shopify, and PayPal; Gemini Enterprise customers include BNP Paribas, Bradesco, Merck, Orange Spain, Santee Cooper, SOMPO, Ulta Beauty, and Wesfarmers. The agent is accessible on iOS, Android, Windows, Mac, the command line, Google Workspace, Microsoft 365, ServiceNow, and Slack.

Adoption figures and rivals

Specialized versions for financial and legal work are already in preview, with versions for government, health care, and retail planned. The launch comes as OpenAI's Dots and Meta's Muse compete in the agent race.