OpenAI Dots: an AI colleague that keeps working day and night
OpenAI launched OpenAI Dots on 29 September 2026: AI agents in ChatGPT that keep working on their own. Not yet with Pro in the Netherlands and Belgium, but with Business Premium they are. Here is what you can do with them.
Imagine giving an assignment at the end of the working day and finding the work done the next morning. That is what OpenAI promises with OpenAI Dots, new AI agents in ChatGPT that the company launched on Tuesday 29 September 2026. Important if you live in the Netherlands or Belgium: you cannot use them there yet with a Pro subscription, but you can with Business Premium. In this article you will read what Dots are, what you can do with them and what to watch out for.
What did OpenAI announce on 29 September?
On 29 September 2026, OpenAI launched Dots: always-on AI agents that keep working on your tasks by themselves. This happened at DevDay, OpenAI’s annual event for developers in San Francisco. Dots were the biggest news of the day, but not the only news.
According to BetaNews’ report, OpenAI also announced the following that day:
- ChatGPT Space: a shared workspace for colleagues, ChatGPT and your Dot, for Pro, Business and Enterprise plans.
- Pages: documents where people and AI agents write and edit together.
- Pro 500: a new plan costing 500 dollars a month, with 25 times as much usage as Plus and an extra-fast mode.
- GPT-6.1 Sol: an improved model for programming and computer use, for now in ChatGPT Work and Codex.
Dots run on GPT-6 Astra, the model OpenAI presented on 3 September. A model is the AI brain behind the chatbot. Want to understand the basics first? Then read our explanation of what AI is in plain English.
What exactly are OpenAI Dots?
OpenAI Dots are personal AI agents that keep working in the cloud on the goals you give them, even between conversations. Each Dot gets its own virtual computer and browser. You set a goal, connect the apps it needs and decide what the Dot may do by itself. Then it comes back with results for you to review.
According to TechCrunch, you start with one personal Dot that you give a name of your own. For organisations, OpenAI is working on specialist Dots, each with its own task, its own login details and its own tools. For now, these can only be seen in trials with large companies. You can open your Dot’s computer at any time to see what it is working on.
The difference from a normal ChatGPT conversation is easiest to see side by side:
| Question | Normal ChatGPT chat | OpenAI Dots |
|---|---|---|
| When does it work? | Only when you ask for something | Continuously, even when you are not there |
| What does it do? | Gives answers in the chat window | Uses websites, apps and documents itself on its own cloud computer |
| Where do you talk to it? | ChatGPT app and website | ChatGPT on desktop, web and mobile (typing or calling), plus Slack and Microsoft Teams. Text messaging will follow later. |
| Does it learn from you? | Limited, through memory | Yes, it adapts to your feedback over time |
A Dot can connect to more than 4,000 apps through ChatGPT’s plug-ins. You create a new Dot in the ChatGPT desktop app or in a browser on your computer, so not on your phone.
What can you do with OpenAI Dots in your work?
Dots are mainly useful for work that keeps coming back, has many steps or takes longer than a few minutes. Think of research, keeping an eye on something and first drafts of texts. OpenAI itself gives examples ranging from booking a table to large projects.
A few real-world examples:
- Marketer: your Dot tracks what competitors post on their website each week and makes a short overview on Monday.
- Freelancer: your Dot collects enquiries from different channels and prepares draft replies that you still check. You can find more ideas in our article on AI customer service for freelancers and small businesses.
- Team leader: your Dot turns meeting notes into a shared Page in ChatGPT Space. That way everyone sees the same action points.
- Business owner: your Dot reads customer feedback and lists recurring complaints. According to BetaNews, an early tester’s Dot noticed an invoice that had not yet been sent, prepared it and only sent it after approval.
A tip few people know: do not just give an agent a goal, but also rules for when to stop. Dots have their own setting for this, Custom Rules. There you set out what the Dot may do by itself, when it must ask first and what it must never do. For example, state that it may not send, buy or delete anything without your approval. That way you stay in control, even when the agent keeps working overnight.
Are OpenAI Dots available in the Netherlands and Belgium?
Only if you use ChatGPT through a Business Premium seat or a business Enterprise or Edu account. With a Pro subscription, you cannot use Dots in the Netherlands and Belgium for now. According to BetaNews, which bases this on OpenAI’s announcement, the rollout for Pro subscribers does not cover the European Economic Area (EEA), Switzerland and the United Kingdom. The Netherlands and Belgium are part of the EEA.
This is what it looks like per plan:
- Free and Plus: no Dots.
- Pro: not yet in the Netherlands and Belgium. It is not yet known when European Pro users will be next.
- Business Premium: available, in every region where ChatGPT is supported.
- Enterprise, Edu and Healthcare: a test version (beta) that an administrator must switch on first. It is off by default.
So check for yourself:
- Log in to ChatGPT on your computer, in the app or in the browser.
- See whether you have an option to start a Dot.
- Do you work for a company with ChatGPT Enterprise or Edu? Then ask your IT administrator whether the beta is switched on.
- Check the current plans and prices on OpenAI’s website before you sign up for anything.
What are the risks of an AI agent that is always on?
An agent that works independently can also make mistakes independently. The more access you give a Dot, the bigger the consequences if something goes wrong. OpenAI itself says that Dots can still make mistakes.
OpenAI has built in a few safety measures. Actions that affect your accounts or share information are first checked automatically against your rules. When you are not actively working with your Dot, it only searches using tools that cannot send or change anything. Changing a password is always left to you. And you can connect your own computer, but this is off by default.
- Privacy: do you connect a Dot to your email, Teams or customer data? Then OpenAI processes that information. Under the GDPR (the European privacy law), your organisation remains responsible for customers’ personal data. If in doubt, talk to your employer or a privacy expert.
- Mistakes: AI can make things up or misunderstand instructions. Always check important work yourself before it goes to a customer.
- Cost: your first Dot is included at no extra cost in Pro and Business Premium, but those are the more expensive plans. Work out whether the time saved outweighs the price.
- Access: give a Dot only the permissions it really needs. According to TechCrunch, OpenAI is working with Microsoft on a link to the security settings of Agent 365, so that companies can manage this better.
There is also a broader debate about the safety of AI agents. According to Al Jazeera, OpenAI is under close scrutiny after incidents with its AI agents. The day before DevDay, the company decided not to release a new model, GPT-6.1 Astra, for now because of safety problems. Reason enough to start slowly.
What can you do right now?
You do not have to wait for Dots to work smarter with AI. Start by describing one recurring task clearly. That is exactly what an agent will need from you later.
Write down what your goal is, which sources the AI may use and what it must never do without you. The better your instruction, the better the result. Also read what a prompt is and how to get more out of it. Try the prompt below in ChatGPT, or soon in your own Dot.
Frequently asked questions
What are OpenAI Dots?
OpenAI Dots are AI agents in ChatGPT with their own virtual computer and browser. They work continuously on your goals, even when you are not there, and learn from your feedback.
Can I use OpenAI Dots with a free or Plus subscription?
No. At launch, Dots were only available to Pro and Business Premium users, with a test version for Enterprise, Edu and Healthcare. Check the current terms with OpenAI.
Are OpenAI Dots already available in the Netherlands?
For Business Premium users, yes. For Pro subscribers in the EEA, including the Netherlands and Belgium, not yet. It is not yet known when that will change.
Does a Dot cost extra?
Your first Dot is included at no extra cost in your Pro or Business Premium subscription. Conversations with your Dot do not count towards your normal usage limits in ChatGPT. OpenAI wants to add the option to run more Dots later.
Is it safe to give a Dot access to my email and customer data?
That comes with risks. Only give access to what is truly needed, set out in Custom Rules what must always come to you first, and take the GDPR into account if you share customers’ personal data.
Do you not want to wait for Dots? Then build your own helper today with our guide to making your own AI assistant without programming. That way you will be ready as soon as agents become available for your subscription too.
Was this article helpful?
