OpenAI launched ChatGPT Work, a cloud-based AI agent that connects directly to a company's workplace tools -- Slack, email, calendars, Google Drive, CRMs and project-management software -- and can complete multi-step tasks across them rather than answering isolated prompts. It's OpenAI's clearest move yet to reposition ChatGPT as a workplace operating layer rather than a conversational chatbot.
The product is powered by GPT-5.6, OpenAI's newest model family, and is built to gather context spread across apps, files and existing workflows to autonomously produce finished output -- documents, spreadsheets, presentations, reports and even simple websites -- instead of requiring a user to assemble the final product from a series of chat responses. Agents run in the cloud, meaning they can continue working on a task after a user closes the app, and they're explicitly designed to be shared across a team: one person builds and refines an agent, and the rest of the organization can use and improve it inside ChatGPT or directly in Slack.
The rollout is staged. ChatGPT Work is going first to Pro, Enterprise and Edu subscribers on web and mobile, with Plus and Business users gaining access in the days that follow; the updated ChatGPT desktop app is available globally on both Mac and Windows. That sequencing prioritizes OpenAI's highest-paying and most enterprise-oriented customers first, consistent with how the company has rolled out other business-facing features.
The launch puts OpenAI in more direct, day-to-day competition with Microsoft's Copilot agents and Google Workspace's Gemini-based tools, both of which have spent the past year building similar cross-application task automation into the productivity suites enterprises already pay for. OpenAI's advantage is ChatGPT's existing enterprise seat base and model quality; its disadvantage is that it doesn't own the underlying productivity suite the way Microsoft and Google do, meaning ChatGPT Work has to integrate into workflows rather than natively hosting them.
For enterprise buyers, the real test isn't the demo -- it's reliability and auditability when an agent is autonomously touching email, calendars and shared documents inside a live organization. What to watch next: how OpenAI prices ChatGPT Work relative to its existing Business and Enterprise tiers, and whether shared, cross-team agents introduce data-governance questions large enterprise IT departments will want resolved before broad rollout.