OpenAI’s new Dots are designed to move ChatGPT from answering questions to actually working on your behalf.
ChatGPT changed how people search for information.
Now OpenAI wants to change how people get things done.
On September 29, 2026, OpenAI unveiled Dots, a new generation of always-on AI agents designed to work autonomously on multi-step projects rather than simply responding to individual prompts.
The launch puts OpenAI directly into competition with Meta’s rapidly expanding Muse personal AI agent.
And it could mark one of the biggest changes in consumer AI yet.
Why It Matters
Traditional chatbots wait for you.
You ask a question.
They answer.
You ask another question.
They respond again.
AI agents are different.
You can give an agent a goal and allow it to work through multiple steps, use connected applications, research information and continue working after you leave the conversation.
That’s what OpenAI is trying to achieve with Dots.
Instead of asking:
“Can you write this report?”
The future AI interface could look more like:
“Take care of this project and update me when something important happens.”
What Are OpenAI Dots?
Dots are OpenAI’s new always-on AI agents.
They are designed to work across applications and perform multi-step tasks on behalf of users.
OpenAI says Dots are powered by GPT-6 Astra.
The agents can interact with tools and services and can work with platforms such as Slack and Microsoft Teams.
They are designed to maintain context and continue working rather than treating every request as an isolated conversation.
Dots vs ChatGPT
The biggest difference is autonomy.
Traditional ChatGPT
You:
- Ask a question
- Receive an answer
- Review it
- Give another instruction
- Repeat
Dots
You:
- Give the agent a goal
- Provide permissions
- Let it work
- Review its progress
- Intervene when necessary
That sounds like a small interface change.
It isn’t.
It changes the role of AI from software you operate into software that operates for you.
The 4,000-App Opportunity
According to reporting around the launch, Dots can work across thousands of supported applications.
That matters because the modern worker rarely uses one application.
A typical workflow might involve:
Slack → Google Calendar → email → documents → spreadsheets → project management → browser research.
If an AI agent can move between these systems, it can potentially automate entire workflows rather than individual tasks.
That is much more valuable than simply generating text.
OpenAI Is Starting With Businesses
Dots are initially available to users on selected premium business and enterprise plans.
That tells us something about OpenAI’s strategy.
The company isn’t treating agents simply as another consumer chatbot feature.
It is targeting workflows where saving human time has an obvious economic value.
Consider a sales manager.
Instead of asking AI to summarize CRM data, the manager could eventually ask an agent to:
- Review this week’s pipeline.
- Identify stalled deals.
- Check recent customer communication.
- Prepare follow-up recommendations.
- Draft messages.
- Update the team.
- Flag anything requiring human approval.
That is an entirely different category of software.
But Always-On AI Creates a New Risk
The more autonomous an AI becomes, the more dangerous mistakes can become.
A chatbot producing the wrong sentence is usually easy to fix.
An agent that sends the wrong email, changes a document, modifies data or makes an unauthorized purchase is different.
OpenAI has therefore added controls that allow users to regulate what Dots can do.
The system also includes review and permission mechanisms.
This is becoming a central issue in agentic AI:
How much autonomy should you give software?
Dots vs Meta Muse
OpenAI’s timing is significant.
Meta introduced Muse earlier in September as a personal AI agent designed to work across a user’s apps and online services.
Muse can perform tasks such as browsing the web, making purchases, managing reminders and connecting to services.
OpenAI’s Dots therefore enter a market that Meta has already started defining.
The competition is no longer simply:
ChatGPT vs Meta AI.
It is becoming:
OpenAI agents vs Meta agents.
The Bigger AI Shift
The biggest change isn’t the name Dots.
It isn’t the interface.
It isn’t even GPT-6 Astra.
The important change is that AI is becoming persistent.
Instead of opening an AI application when you need help, your agent could remain available in the background.
It could monitor projects.
Track information.
Prepare work.
Look for changes.
And come back to you when human attention is actually needed.
That is much closer to having a digital employee than a digital search engine.
What’s Next
OpenAI’s Dots are still at the beginning of this transition.
The biggest questions will be:
- How reliable are they?
- How much autonomy will users actually grant them?
- How much will businesses pay?
- How much computing power will always-on agents require?
- Can OpenAI make them safe enough for sensitive workflows?
- Will consumers eventually get the same capabilities?
The answer could determine whether AI agents become another software feature or an entirely new computing platform.
FAQ
What are OpenAI Dots?
Dots are OpenAI’s new always-on AI agents designed to autonomously work on multi-step tasks.
What model powers Dots?
OpenAI’s Dots are powered by GPT-6 Astra.
How are Dots different from ChatGPT?
ChatGPT generally responds to individual prompts, while Dots are designed to pursue longer-running tasks with greater autonomy.
Can Dots use other applications?
Yes. Dots are designed to work with connected applications and services, including workplace platforms.
Are Dots available to everyone?
The initial rollout is focused on selected premium business and enterprise users, with broader availability expected to evolve over time.
Related Reading
- Meta Muse Explained: The AI Agent That Wants to Run Your Digital Life
- OpenAI vs Meta: The New AI Agent War Has Begun
- AI Agents Explained: Why Chatbots Are Becoming Digital Workers