1300 166 138 Tell us what you need
News / AI Agents

OpenAI launches Dots, AI agents that work while you sleep

By Monika Tantau • Published 2 months ago
OpenAI launches Dots, AI agents that work while you sleep

Key takeaways

• OpenAI launched Dots on 29 September, always on agents powered by GPT 6 Astra
• Each dot has its own cloud computer and works across connected apps, Slack and Teams
• Australian Pro users get their first dot at no extra cost; EU and UK Pro users miss out
• Meta's rival agent, Muse, is still limited to the US, Canada and Mexico

Most Read

OpenAI launched Dots at its DevDay event on 29 September, a new kind of ChatGPT agent that keeps working when you are not. According to OpenAI, each dot runs on GPT 6 Astra, has its own cloud computer and browser, and learns how you like work done so it can take tasks off your plate around the clock.

It is OpenAI’s answer to Meta’s Muse, and unlike Muse it is already reaching Australia. Pro subscribers here get their first dot included in the plan, while Pro users in the European Union and the United Kingdom have been left out for now.

What a dot actually does

You give a dot a name, a goal and the apps it may use, then talk to it in ChatGPT, Slack or Teams, or on a voice call. It works through tasks, decides what should happen next and brings results back for review. OpenAI’s own example is an early tester whose dot noticed he had forgotten to invoice a publication, prepared the invoice and sent it once he approved.

When you are not actively working with it, a dot carries out what OpenAI calls proactive research, reading the apps you have connected to spot things it could help with. Dots can connect to more than 4,000 apps through OpenAI’s plugins, and you can open a dot’s computer at any time to see what it is doing.

The controls OpenAI has built in

Proactive research uses read only tools, so in the background a dot cannot send messages, change content or control your browser. Access to your own laptop is off by default and has to be granted from the desktop app. OpenAI says its monitoring can pause or stop a dot if it detects a safety concern, and it still warns that dots make mistakes and consequential work should be reviewed.

For companies, OpenAI is piloting specialist dots with their own identity and credentials for set roles such as invoice processing or customer support, and is working with Microsoft to manage them through Agent 365. Content from Business and Enterprise workspaces is not used to train OpenAI’s models by default.

Who can use it in Australia

The OpenAI help centre says Dots are rolling out to Pro users everywhere except the European Economic Area, Switzerland and the United Kingdom, and to Business Premium users in all supported regions. Enterprise, Edu and Healthcare workspaces can try a beta once an admin switches it on.

Pro starts at $154.99 a month in Australia, and the first dot comes at no extra cost, with extended limits for the first month. Access is arriving gradually over several days, and a dot can only be created from the ChatGPT desktop app or desktop web. How a dot compares with Gemini Spark and Claude Cowork on price is laid out in our AI agents buying guide for Australia.

How Dots compares with Meta Muse

How Dots compares with Meta Muse

Meta launched Muse on 8 September as a personal agent that browses, fills in forms, sends email and books travel after you close the app. The two products chase the same idea, but Muse is only available in the United States, Canada and Mexico, and Meta has given no date for Australia. Our Meta Muse guide covers what it does and what it costs.

Why Australian businesses should care

An agent that reads email, calendars and shared drives around the clock handles personal information, and under the Privacy Act 1988 the organisation stays accountable for what happens to it. Before staff connect a dot to work systems on a personal Pro plan, businesses should decide which apps are allowed and who reviews what the agent sends.

Teams that want the same always on help inside their own systems can build it with tighter controls, which is the kind of work covered in AI agent development. The difference is who sets the boundaries and where the data lives.

Agents that never clock off

Dots mark a shift from assistants you prompt to agents you supervise. With OpenAI and Meta both pushing always on agents, and the OECD working on guardrails for AI agents in government, the question for businesses is moving from what an agent can do to how closely it should be watched.

Scroll to Top