OpenAI has introduced workspace agents in ChatGPT, a new research-preview feature aimed at teams that want AI systems to handle shared, multi-step work instead of one-off prompts.

According to OpenAI, workspace agents are an evolution of GPTs built on Codex and designed to run in the cloud. That means they can keep working after a user leaves the session, pull context from connected tools, follow team processes, and ask for approval before sensitive actions such as editing files, sending emails, or updating calendars.

The launch is framed as a way to move beyond solo productivity use cases. OpenAI says teams can create one agent and share it across ChatGPT or Slack for recurring workflows like software request reviews, routing product feedback, weekly metric reports, lead qualification, and vendor risk checks. The company also says admins can control which tools agents may access, who can build or share them, and how runs are monitored through analytics and the Compliance API.

Workspace agents are available in research preview for ChatGPT Business, Enterprise, Edu, and Teachers plans. OpenAI says the feature will be free through May 6, 2026, after which it will move to credit-based pricing. The release suggests OpenAI is pushing ChatGPT further into collaborative back-office automation, where the value is less about answers in a chat window and more about keeping cross-team workflows moving without constant manual follow-up.