Imagine giving ChatGPT a task today, closing your laptop, and coming back later to find that the work has continued without you.
Not another answer. Not another conversation. Actual progress on a project you asked it to handle.
That is the idea behind ChatGPT Dots, OpenAI's always-on AI agents designed to work toward your goals, use connected tools, and keep projects moving between conversations.
Announced on September 29, 2026, Dots introduce a different way of working with AI. Instead of treating every request as a separate question, you can give an agent ongoing responsibilities and return to review what it has done.
The interesting part isn't that AI can answer questions. We already know that.
The interesting part is what happens when AI can keep working after you stop asking questions.
What Exactly Are ChatGPT Dots?
ChatGPT Dots are personalized AI agents that can research information, analyze data, prepare documents, work with software, and make progress toward longer-term goals.
According to OpenAI, each dot has its own cloud computer and browser. It can use connected apps, draw on relevant context from ChatGPT memory, and learn from your feedback over time. Dots are powered by GPT-6 Astra. Source: OpenAI
Think of the difference this way.
A regular chatbot helps you write a marketing plan. A dot could help research the market, organize findings, prepare content, track relevant updates, and return to you when a decision needs your input.
You still provide the direction. The difference is that the work doesn't have to end with the first response.
How Are Dots Different From Regular ChatGPT?
The biggest difference is continuity.
In a typical conversation, you ask something, receive an answer, and decide what to do next. With a dot, you can establish an ongoing goal, provide the necessary context, and let it work between conversations.
Dots can track progress, use connected tools, and ask for your judgment when a decision requires it. You can also give feedback or change priorities without starting the entire process again. Source: ChatGPT Learn
This does not mean every task runs perfectly or that every action happens without approval. Dots have defined permissions, safety checks, and limitations.
They are better understood as agents that can carry work forward, rather than chatbots that simply produce responses.
What Can You Actually Do With ChatGPT Dots?
OpenAI describes several practical uses, including preparing presentations, turning interviews into content, and improving applications.
For example, a founder could ask a dot to collect customer feedback, investigate changes in a market, and update a presentation as new information becomes available.
A developer could ask it to investigate a bug, propose a code change, and test the result for review.
A content creator could use it to turn interview transcripts into show notes and social posts, then apply feedback across future drafts.
These are examples of intended uses, not guarantees that every task will succeed without supervision.
The larger idea is simple: instead of repeatedly telling AI what to do next, you give it a goal and let it work out the next steps within the permissions you have granted.
Can ChatGPT Dots Work While Your Computer Is Off?
Yes. A dot has its own cloud computer, so it can continue supported work even when your personal computer is switched off.
You can connect your own computer separately if you want the dot to work with local files, applications, or tasks that depend on your device. That connection has additional requirements, including keeping your computer online and the ChatGPT app open while it uses your machine. Source: OpenAI Help Center
This distinction matters. The cloud computer and your personal computer are separate environments, and connecting one does not automatically grant access to everything on the other.
Are ChatGPT Dots Free?
At launch, OpenAI includes a first dot with eligible ChatGPT Pro and Business Premium plans at no additional cost. Enterprise users can access the beta when their workspace administrator enables it.
Availability is rolling out gradually, with regional and plan restrictions. OpenAI's published guidance excludes the European Economic Area, Switzerland, and the United Kingdom from the initial Pro rollout. Check the official Dots guide for current eligibility.
The important point is that Dots are not a universally available free feature for every ChatGPT account.
Are ChatGPT Dots Safe to Use?
Giving an AI agent more responsibility also means thinking more carefully about access and control.
Dots use permissions and approval rules to determine which actions they can perform independently, which require approval, and which are blocked. You can manage connected apps, establish custom rules, review activity, and pause a dot when needed.
OpenAI also explains that proactive research uses restricted, read-only tools. That research cannot directly send messages, change content through plugins, or control a browser or computer. Source: OpenAI's Dots Safety FAQ
These safeguards reduce risks, but they do not eliminate them. Dots can still make mistakes, and malicious instructions embedded in websites or documents remain a security concern.
Review consequential work before relying on it, especially when it involves sensitive information, financial decisions, or actions affecting other people.
Why ChatGPT Dots Matter
For years, using AI has largely meant opening a conversation whenever you need help.
Dots point toward a different model: an AI agent that remembers relevant context, follows a project, uses permitted tools, and keeps making progress between conversations.
That could change how people approach research, software development, content creation, and everyday work. But the value will depend on reliability, permissions, and whether the agent actually saves more time than it takes to supervise.
Dots are not proof that AI can do everything. They are a practical experiment in giving AI more continuity and responsibility.
And that may be the more interesting shift.
The future of AI might not be about getting better answers. It might be about giving better instructions once, then having the work continue.