OpenAI has introduced workspace agents in ChatGPT as shared, Codex-powered agents for teams. The product is designed to handle complex tasks and long-running workflows while operating within an organization’s permissions and controls. Workspace agents extend GPTs into collaborative business use cases, allowing teams to build one agent, use it across ChatGPT or Slack, and improve it over time. The system is built for work that depends on shared context, approvals, handoffs, and coordination across tools and teams.
Teams can create an agent from the Agents tab in the ChatGPT sidebar by describing a recurring workflow or uploading a file, then refining the agent through guided setup, tool connections, skills, and testing. OpenAI highlights several example use cases, including software request review, product feedback routing, weekly metrics reporting, lead outreach, and third-party risk management. The agents can gather context from connected systems, write or run code, use apps, retain memory, and continue work across multiple steps. They can also run on a schedule or operate in Slack, where they can respond to requests, answer questions, link documentation, and file tickets when needed.
OpenAI positions the product as a way to turn team knowledge and internal processes into reusable workflows. Shared agents can be made available across an organization and refined through ongoing use, with memory and conversational guidance helping them improve over time. An accounting example shows an agent preparing key parts of month-end close, generating workpapers with inputs and control totals for review, and following internal policies. For sensitive actions such as editing spreadsheets, sending emails, or adding calendar events, teams can require approval before the agent proceeds. Usage analytics show how many runs an agent has completed and how many people are using it.
Enterprise governance is a central part of the rollout. ChatGPT Enterprise and Edu admins can control which connected tools and actions user groups can access, manage who can use, build, and share agents, and monitor configurations, updates, and runs through the Compliance API. Built-in safeguards are intended to keep agents aligned with instructions when encountering misleading external content, including prompt injection attacks. OpenAI says admins will soon be able to view every agent built across their organization in the admin console, including usage patterns and connected data sources.
Workspace agents are available in research preview for ChatGPT Business, Enterprise, Edu, and Teachers plans. For Enterprise and Edu plans, admins can enable agents using role-based controls. Workspace agents will be free until May 6, 2026, with credit-based pricing starting on that date. OpenAI says upcoming additions will include new triggers that can start work automatically, better dashboards, more actions across business tools, and support for workspace agents in the Codex app.
