OpenAI is taking its AI assistant strategy beyond the traditional chatbot with Dots, a new generation of persistent AI agents designed to keep working even when the user is away. Announced at DevDay 2026, Dots can monitor projects, use software, respond to changing information and return completed work for human approval. They are powered by GPT-6 Astra and operate through their own cloud computer and browser.
Alongside Dots, OpenAI has introduced ChatGPT Space, a shared environment where employees, ChatGPT, Codex and Dots can work with the same files, Pages and project information. Together, the two products show how OpenAI wants AI to move from answering individual questions to taking responsibility for ongoing workflows.
What Are OpenAI Dots?
Introducing dots, powered by GPT-6 Astra.
Remarkably capable, always-on agents built to handle everything. pic.twitter.com/inY5BZj3Kz
— OpenAI (@OpenAI) September 29, 2026Dots are designed to act more like persistent digital coworkers than conventional chat assistants. A user can give a Dot an objective and the standards it needs to follow. The agent can then continue working in the background and return when there is something useful to review.
Each Dot gets its own cloud computer and browser. It can use connected applications and communicate through ChatGPT, Slack and Microsoft Teams. OpenAI says Dots can also work with recurring tasks and follow up across multiple communication channels.
The idea is simple. Instead of opening ChatGPT every time something needs to be done, users can delegate an ongoing responsibility to an agent.
Powered By GPT-6 Astra
GPT-6 Astra is rolling out today to a limited set of organizations and over the coming days will become available to all ChatGPT Plus, Pro, Business, and Enterprise users, as well as through the OpenAI API and AWS.
Be ready to experience Astra at its best. Get the ChatGPT… pic.twitter.com/6u0irS8iF0
— OpenAI (@OpenAI) September 3, 2026Dots are built on GPT-6 Astra, OpenAI’s flagship model for complex end-to-end work, including reasoning, coding, computer use, research and document creation. Astra supports a context window of up to 1.05 million tokens and up to 128,000 output tokens through the API.
OpenAI says Dots can connect through its plugin ecosystem to more than 4,000 applications. This gives the agents access to the software and information that businesses already use rather than forcing companies to rebuild their workflows from scratch.
From AI Copilots To Delegated Work
The biggest difference is continuity. A developer, for example, could ask a Dot to monitor customer feedback, identify recurring bugs, work on smaller fixes and return completed pull requests with demonstrations of the changes. The developer can then concentrate on a larger feature.
The same approach can be used by product and marketing teams. A Dot can learn an organisation’s audience, positioning and creative standards. It can then update launch materials when the underlying product changes.
OpenAI also highlights research, sales and content workflows. A sales agent can compare a prospect’s requirements with product documentation and account history, identify missing technical tests and keep a proposal updated. A content-focused Dot can work with an interview transcript, identify potential clips, prepare show notes and draft social posts.
The common idea is not simply generating content. It is allowing the AI to continue managing a workflow after the initial instruction.
ChatGPT Space Creates A Shared AI Workspace
ChatGPT Space extends the concept from individual agents to entire teams. Space is designed as a shared home for files and Pages, with Pages acting as editable documents that people and their ChatGPT agents can work on together. It replaces the existing Library, while Projects remain separate.
A team could maintain a project Page and ask ChatGPT or a Dot to keep it updated. Information from connected tools can be used to update action items, deadlines and other project details as circumstances change.
The same Dot can also be accessed through supported communication channels. This means an employee can interact with the agent through ChatGPT, Slack or Microsoft Teams while the resulting work continues to live inside the team’s shared environment.
Space also supports collaboration around files, presentations, spreadsheets and other work materials. Access can be controlled with view or edit permissions, and those permissions can later be changed or revoked.
Background Work Comes With Controls
An always-on AI agent needs stronger controls than a normal chatbot. OpenAI says Dots can perform proactive research using read-only tools. These background tools cannot send messages, modify content or control a user’s computer or browser.
For actions that can change information or affect external systems, users can configure app permissions and Custom Rules. OpenAI also provides an Activity View to inspect background work. An auto-review system can check potentially consequential actions against instructions, rules and safety requirements before allowing the agent to continue or asking for human approval.
Dots normally work inside their own cloud computers. Local computer access can be separately permitted in supported workspaces, giving the agent access to local files and commands. OpenAI’s current Enterprise controls show that local computer access and Custom Rules are configurable workspace permissions, with Dots access itself off by default for Enterprise.
Privacy And Shared Work
ChatGPT Space also introduces an important distinction between private and shared information. Sharing a Page does not expose a user’s private chats or personal ChatGPT Memory. However, if information from personal context is written into a shared Page, people with access to that Page can see it.
OpenAI says Business and Enterprise data is not used to train its models by default. For personal accounts, training depends on the user’s individual settings. That makes review and permissions particularly important when Dots are handling company information or making changes to shared work.
Conclusion
With Dots and ChatGPT Space, OpenAI is pushing ChatGPT toward a more continuous model of AI-assisted work. The user does not simply ask a question and wait for an answer. Instead, they can assign an objective, provide rules and let an agent continue working as information changes.
The bigger shift is from AI assistance to AI delegation. Dots handle ongoing tasks, while Space gives humans and agents a common place to build, update and review the resulting work. OpenAI’s latest move suggests that the next stage of workplace AI may be less about having a smarter chatbot and more about deciding which responsibilities can safely be handed over to an agent.
Follow Us: Facebook | X | Instagram | YouTube | Pinterest
