What are OpenAI Dots?
- Written by
- Tom Houpt
- Published
A Dot is OpenAI’s version of an always-on AI agent. You give it a job, and it keeps working after you close the chat.
The basic idea
A normal chat assistant waits for you to type. A Dot runs on its own cloud computer with its own browser. It can work toward a goal for hours, send you progress updates and ask a question when it’s stuck. OpenAI announced Dots at its developer conference on 29 September 2026. For how Dots compare with similar AI agents from Meta and xAI, read Dots, Muse and Grok Bot.
What a Dot can do
Each Dot can use more than 4,000 connected apps through OpenAI’s plugins. You can watch what it does on its computer while it works. It can also do what OpenAI calls proactive research, looking for ways to help in the background with read-only access to your connected apps. Press coverage says it learns your preferences from feedback.
Typical jobs are the kind of ongoing admin that suits an AI agent: drafting invoices, chasing late payments, tracking a project or summarizing a pile of email.
Who gets one
At launch, Pro and Business Premium subscribers get one Dot included. Enterprise, Edu and Healthcare workspaces have a beta, with admin approval. Reports say Pro users in the European Economic Area, Switzerland and the UK were left out at first. You create your Dot in ChatGPT on the desktop or in a browser, then reach it from the mobile apps, Slack or Microsoft Teams. Check OpenAI’s page for current plans and regions, because this is changing quickly.
What you control
You can set rules for when a Dot acts on its own and when it has to ask you first. OpenAI says sensitive actions, such as changing a password, always need your approval. You can watch its activity and pause it. Start with “ask first” for everything, then loosen the rules once you trust the results.
What to watch out for
OpenAI says Dots can still make mistakes and tells people to review consequential work. A Dot that can send messages or touch money needs the same checks as a new employee. Connect one app at a time, keep passwords out of the chat and read what it did. If a Dot writes to a customer for you, the message still carries your name.
Who it’s for
People with recurring work that follows a pattern, who are already on an eligible plan and are willing to supervise at first. If you only need one-off answers, a regular chat assistant like ChatGPT is simpler.
Where to start
A Dot is a plain example of handing work to AI instead of chatting with it. It’s new, so pick one low-stakes job, leave approvals on, and read the result before you give it more.
Common questions
What is a Dot?
An AI agent from OpenAI that runs in the cloud with its own computer and browser. You give it a job, it works on its own, and it sends updates or asks questions along the way.
Who can get a Dot?
At launch, Pro and Business Premium subscribers get one Dot, and Enterprise, Edu and Healthcare workspaces have a beta. Reports say Pro users in the EEA, Switzerland and the UK were excluded at launch, so check current availability.
Is a Dot the same as ChatGPT?
No. ChatGPT waits for you to ask. A Dot keeps working in the background on a standing job, and you reach it through ChatGPT, Slack or Microsoft Teams.
Can a Dot make mistakes?
Yes. OpenAI says Dots can still make mistakes and tells people to review consequential work. Set it to ask before it sends, spends or deletes anything.
Get our new guides first on Google: add Less Admin HQ as a preferred source (opens in a new tab).