OpenAI launches Workspace Agents to replace GPTs, enabling intelligent agents to autonomously run continuously in the cloud

According to Beating Monitoring, OpenAI has introduced Workspace Agents in ChatGPT, positioned as an upgraded version of GPTs. Workspace Agents are powered by Codex, run in the cloud, and can continue executing tasks even after the user goes offline. They support scheduled triggers and Slack deployments, allowing teams to co-build and share.

On the ChatGPT sidebar, users click “Agents” and describe a workflow in natural language. The system guides them through step-by-step agent setup, including defining steps, connecting tools, adding skills, and testing. OpenAI has listed 5 example agents: software procurement approval, product feedback routing, weekly report auto-generation, lead follow-up, and third-party supplier risk assessment. The official also provides templates for scenarios such as finance, sales, and marketing, with built-in preset skills and tool configurations.

Each agent has its own file, code, tools, and memory workspace. You can write code, call integrated applications, and maintain context across steps. For sensitive operations (such as editing spreadsheets, sending emails, or adding calendar events), you can set up that manual approval is required before execution. Workspace Agents also support the Compliance API, allowing administrators to view each agent’s configuration, update and run records, and pause the agent when necessary.

This feature will be rolled out in research preview for the ChatGPT Business, Enterprise, Edu, and Teachers plans, free until May 6, after which it will switch to quota-based billing. Existing GPTs will remain temporarily, and later will support one-click migration to Workspace Agents.