The Pulse
OpenAI Launches Dots, Agents That Work in the Background
OpenAI launched dots, always-on agents powered by GPT-6 Astra that can work across connected apps and their own cloud computers. The rollout begins with eligible Pro and Business Premium users, while Enterprise users can try a beta enabled
AI.info Team · Written with AI ·
“This is the real deal version of AI we've always imagined.”
Sam Altman, CEO of OpenAI, quoted by Axios
OpenAI introduced dots on September 29, describing them as agents that can keep working toward a user’s goals when the user is not actively directing them. Powered by GPT-6 Astra, each dot gets its own cloud computer and can work with apps a user connects. OpenAI says its plugin ecosystem can link dots to more than 4,000 apps. The company is starting the rollout in eligible markets for Pro and Business Premium subscribers, with an Enterprise beta available when a workspace administrator enables it.
A cloud computer that keeps working
A dot can use its own browser and cloud computer to handle several projects, while users can inspect its work and, if they choose, allow it to connect to their laptop. OpenAI says dots can be reached through ChatGPT on desktop, the web and mobile, as well as Slack and Microsoft Teams. They can send progress updates or ask users to clarify a task. The company says texting is planned, but is not available yet.
OpenAI’s examples range from watching customer feedback and preparing tested software fixes to revising launch materials when a product’s scope changes. Another example has a dot updating a scientist’s analyses as new data arrives. These are use cases the company describes, not published performance results. OpenAI also recounts an early tester’s dot preparing an overdue invoice and sending it after receiving approval.
Background work has a read-only boundary
OpenAI draws a line between background research and actions that change or share information. When a user is not actively working with a dot, its “proactive research” uses connected apps with read-only tools, which the company says cannot send messages, change app content or control the user’s browser or computer. Separate permissions and custom rules let users allow particular actions, require approval or block them. Dots can be monitored in an Activity View and redirected while they work.
For actions that could affect accounts or share information, OpenAI says its auto-review system checks instructions, custom rules and safety requirements to decide whether work can proceed or needs approval. Some sensitive actions, such as changing a password, remain with the user. OpenAI cautions that dots can make mistakes and says users should review consequential work. That warning matters for a product designed to continue work outside the user’s immediate supervision.
One dot is included, with limits on deeper work
The first dot comes at no extra cost with Pro and Business Premium plans, according to OpenAI. Conversations with a dot do not count toward ChatGPT usage limits, but tasks it starts or manages in Codex or ChatGPT Work do count. Plans include an allowance for deeper work, with extended limits during the first month after launch; OpenAI says users will later be able to add dots and increase how quickly or how much work each one handles.
OpenAI is also previewing specialist dots for organizations: agents with their own identities, credentials and access to designated company systems. The company says it is beginning focused enterprise pilots, drawing on internal tests involving procurement, invoice processing, email marketing, customer support and commercial contracting. It is working with Microsoft to integrate the specialist agents with governance and security controls in Agent 365. For now, individual dots begin with one agent per user, while organizational deployments remain pilots.