Imagine hiring a helper who never sleeps, never forgets what you told them, and works on your to-do list while you do something else. That is the idea behind Dots, a new product from OpenAI, the company that makes ChatGPT.
So, what is a Dot?
A Dot is an always-on assistant that lives inside ChatGPT. You can think of it as a digital worker. You give it a job, and it keeps working on it, even when you close the app.
OpenAI announced Dots at its DevDay 2026 event. The company describes them as proactive systems that can chip away at tasks every hour of the day.
How is this different from regular ChatGPT?
With normal ChatGPT, you ask a question and get an answer. Then the conversation stops until you speak again.
A Dot is built to keep going. It can continue working toward a goal between conversations. It can also learn your preferences and working habits over time. The more you use it, the better it should fit the way you like things done.
Where does it do its work?
Each Dot gets its own space online. OpenAI says a Dot has its own cloud computer, its own browser, and the apps you've connected. You can open your Dot's "computer" at any time to see what it is doing. If you want, you can also connect a Dot to your own PC so it can work alongside you.
Dots can also work with a huge number of other tools. They can interact with more than 4,000 apps.
How do you talk to a Dot?
The way you would talk to a coworker. You can message or call Dots in ChatGPT on desktop, web, and mobile. They can also reach you through Slack and Microsoft Teams. A Dot can message you with progress, questions, or decisions that need you.
What could a Dot actually do?
Here are some everyday examples of the kind of work a Dot is made for:
Keep an eye on customer feedback and suggest fixes
Work through your inbox and calendar
Update reports when new information comes in
Handle tasks that repeat every week
OpenAI also says Dots can turn repeated feedback into tested changes for people to review. Think of it as a helper that does the first round of work, so you only need to check and approve.
Is it safe?
This is the big question. A tool that can click, type, and use your apps needs strong guardrails.
OpenAI says Dots start with built-in rules for when to act on their own and when to ask for approval. The company also says it tested Dots against fake emails designed to trick them, many of which were written as attacks.
Still, OpenAI is honest about the limits. The company admits Dots can still make mistakes and tells users to review important work. So the smart approach is simple: let a Dot do the heavy lifting, but check anything that really matters.
Who can use Dots right now?
Not everyone yet. Dots are rolling out in select markets to Pro and Business Premium subscribers. The first Dot is included at no extra charge for those plans. Free, Go, and Plus plans do not have Dots at launch, and nobody under 18 can use them.
Many details are still unknown. OpenAI has not said how much extra Dots will cost, or how much work each plan allows. There is also no announcement yet of a way for outside developers to build with Dots.
What comes next?
OpenAI has bigger plans. The company imagines a future where people control whole teams of Dots. It is also testing "specialist" Dots for organizations, meaning helpers trained for specific jobs inside a company.
The bottom line
Dots are OpenAI's attempt to move AI from "something you chat with" to "something that works for you." If it delivers, it could save people a lot of time on boring, repeated tasks. But it is brand new, access is limited, and mistakes can still happen.
If you get access, start small. Give your Dot one simple job, watch how it handles it, and build trust from there.