Summary
OpenAI has begun rolling out Dots, AI agents that can carry out ongoing tasks using a separate cloud computer and connected apps. Users can set permissions and review work, while specialist versions for organisations are in focused pilots.
OpenAI announced Dots on September 29, 2026, and has begun rolling them out as AI agents in ChatGPT. They are designed to handle continuing, multi-step tasks through their own cloud computer and connected apps, rather than waiting for a prompt at each stage. OpenAI says Dots are powered by GPT-6 Astra and can connect to more than 4,000 apps through its plugin ecosystem.
How Dots are designed to work
A Dot can work on several projects, continue tasks in the background and send progress updates or questions. Users can interact with it in ChatGPT, Slack or Microsoft Teams; voice calls are also supported through ChatGPT. OpenAI says context carries across these channels. Text messaging is planned but is not yet available.
Each Dot operates on a separate cloud computer with its own browser. The user's computer and files remain separate unless the user gives the Dot permission to connect to that device. OpenAI also describes a background feature called “proactive research”: when a person is not actively working with a Dot, it can look for useful information in connected apps using read-only tools. Those tools cannot send messages or change app content.
For actions that could affect accounts or share information, an automated review checks the proposed action against the user's instructions, custom rules and safety requirements. Depending on those checks, a task may proceed, require approval or need the user to complete it. Users can inspect the Dot's cloud computer, track its activity and redirect it. Some sensitive tasks, such as changing a password, remain with the user.
OpenAI presents examples ranging from preparing software fixes for review to updating research analyses when new data arrives. It also describes an early tester's Dot preparing and sending an invoice after the tester approved it. These examples illustrate the intended workflow: the agent carries out preparatory or delegated work, while people retain oversight of consequential decisions.
Rollout and organisational use
Dots are rolling out to Pro and Business Premium users in eligible markets. The first Dot is included in those plans at no extra cost, with an allowance for deeper work. Enterprise users, including those in Education and Healthcare workspaces, can try a beta when their workspace administrator enables it. Initial setup takes place in the ChatGPT desktop app or a desktop browser; after setup, users can message their Dot in the mobile app.
OpenAI is also beginning focused enterprise pilots of specialist Dots. Unlike a personal Dot, each specialist version is set up by an organisation with its own identity, credentials and access to selected systems for a defined responsibility. OpenAI is working with Microsoft to integrate these specialist agents with governance and security controls in Microsoft Agent 365.
The design gives agents more room to carry work forward between interactions, while permissions, activity monitoring and approvals provide ways for people and organisations to supervise that work. OpenAI cautions that Dots can make mistakes and advises reviewing consequential work.