OpenAI has added Team Tasks to ChatGPT, letting a team automate recurring work instead of one person maintaining it alone. In a post from the OpenAI Developers account at 23:24 Beijing time on October 20, the company gave the example of pulling together what shipped and what's blocked every Friday, with teammates able to review the results and change what the next update covers.
According to OpenAI's ChatGPT Learn documentation, a Team Task can run on a schedule or when a specific event occurs. The examples include summarizing a week of customer updates each Monday and summarizing launch progress whenever a message arrives in a launch Slack channel.
The tasks run in the cloud with the team's service account and configured app connections, so a connection's account, not the creator's personal access, determines what data a task can reach and what actions it can take. The documentation also says tasks do not inherit the creator's personal memories, custom instructions or chat history.
Access is shared across the team. New members can see all earlier runs and generated files, and individual tasks and runs cannot have separate access restrictions, so OpenAI advises reviewing that access before inviting colleagues.
Enterprise admins decide who can create teams and manage Team Tasks through workspace permissions. The documentation names the task permission two ways, as "Create and manage team automations" in its overview and FAQ and as "Create and manage team tasks" in its setup checklist, alongside a separate "Create teams" permission.
On billing, the documentation says only this: "Team Tasks use workspace credits. Team spending limits are separate from user limits." It does not list a per-run cost or say which plans include the feature.