OpenAI launches workspace agents in ChatGPT

OpenAI is pushing ChatGPT from a personal assistant into a shared work surface. In a new product launch, the company introduced workspace agents: shared agents that can handle long-running workflows across files, code, apps, and Slack, while operating inside the permissions and controls set by each organization.

What is new

OpenAI positions workspace agents as the next step beyond GPTs. They run in the cloud, can keep working when users are away, and can be shared across teams in ChatGPT and Slack. OpenAI says teams can build agents for software review, weekly metrics reporting, lead outreach, and third-party risk work. The product includes approvals for sensitive actions such as email, calendar, and spreadsheet edits, plus analytics and compliance visibility for admins. Workspace agents are launching in research preview for ChatGPT Business, Enterprise, Edu, and Teachers plans, and will be free until May 6 before credit-based pricing starts.

Why this matters

This matters because the market is shifting from personal copilots to team-level operational agents. OpenAI is not just trying to help individual users write faster. It is trying to become the place where workflows are built, shared, governed, and measured. For CIOs, that means the competition for the enterprise agent layer is happening now, and access controls, logging, approvals, and Slack integration matter almost as much as the model itself.

Source and date validation

The original source is OpenAI’s own product post, “Introducing workspace agents in ChatGPT,” published on April 22, 2026. That is an official publication date and well within the 48-hour freshness window, so the story qualifies as current.

Source: https://openai.com/index/introducing-workspace-agents-in-chatgpt/