OpenAI Launches Dots: Always-On AI Agents Powered by GPT-6 Astra That Work While You Sleep

Picture an assistant that doesn’t wait to be asked. It books the flight, schedules the meeting, and gets on with the research — while you’re making a cup of tea. That’s the idea behind Dots, OpenAI’s new class of AI agents announced on 29 September 2026.

Unlike the ChatGPT most people are familiar with — where you type a question and get an answer — Dots are built to take on ongoing work and keep making progress between conversations. OpenAI describes them as always-on agents, each equipped with its own cloud computer and browser, capable of connecting to more than 4,000 applications.

What Dots Actually Do

The core idea is delegation. You set a goal — say, coordinating a project across a team, tracking a research topic, or managing a recurring scheduling task — and your Dot keeps working towards it, using the permissions and connected tools you’ve given it.

OpenAI says Dots can book flights, assign tasks, and schedule meetings, subject to the permissions granted by connected applications and the instructions users set. They retain context over time, meaning a Dot can build on previous conversations rather than starting from scratch each time. The company describes this as 24/7 operation, though it’s worth being clear that independently verified, uninterrupted performance across all use cases hasn’t been established.

Dots are powered by GPT-6 Astra, the model OpenAI introduced in September 2026 for coding, research, computer use, and complex multi-step tasks. It’s a step up from the conversational models most users know, built with autonomous action in mind rather than simple back-and-forth dialogue.

You create a Dot through ChatGPT on desktop web or the desktop application. For some users, they can also work through workplace tools including Slack and Microsoft Teams.

The Access Picture — and Why UK Users Should Pay Attention

Here’s the catch for anyone reading this in the UK. OpenAI’s support information states that Pro access to Dots excludes the European Economic Area, Switzerland, and the UK at launch. So if you’re a ChatGPT Pro subscriber hoping to try this today, you’ll need to wait.

Business Premium access is a different story. OpenAI says that tier is available across supported ChatGPT regions, which may include the UK, depending on an organisation’s specific plan and settings. Enterprise access is described as a beta and is off by default.

OpenAI has not confirmed that Dots are generally available to all ChatGPT users anywhere. The rollout is gradual, limited to eligible users in supported markets.

What This Technology Represents

This isn’t just a new feature. It marks a genuine shift in what AI assistants are expected to do. The previous model — prompt in, response out — is giving way to something closer to a delegated colleague: one that acts on your behalf, across systems, over time.

But that shift brings real questions. Always-on agents with access to cloud computers, browsers, and external applications can make mistakes. They can take actions that are hard to reverse. Questions around unauthorised access, data privacy, cybersecurity, and accountability become considerably more pressing when an AI isn’t just answering questions but actually doing things.

Sam Altman, OpenAI’s chief executive, has previously described the company’s direction as building AI that can take on “cognitive labour” on behalf of users. Dots appear to be a concrete step in that direction.

From a business perspective, the appeal is obvious. Repetitive administrative work, research tasks, scheduling coordination — these eat up hours that people would rather spend elsewhere. But the value depends heavily on reliability, the quality of integrations, and whether the subscription cost makes sense for the organisation.

And on employment: there’s a reasonable question about what happens to roles that consist largely of the tasks Dots are designed to handle. The scale and timing of any effect on UK employment isn’t something anyone can state with confidence right now, but it’s a conversation that’s starting to happen in offices and policy circles alike.

What Competitors Are Doing

OpenAI isn’t alone in this space. Persistent AI agents — ones that work autonomously across applications over time — are an area where several major technology companies are competing. Independent reporting has described Dots as entering a market alongside other agent-style products, though whether GPT-6 Astra gives OpenAI a meaningful edge over rivals hasn’t been independently established.

What This Means for Kent Residents

For most ChatGPT Pro users in Kent, Dots aren’t available yet — OpenAI has explicitly excluded the UK from Pro access at launch, and no confirmed UK release date has been given. Kent businesses on Business Premium or Enterprise ChatGPT plans may be closer to access, depending on their regional settings, but any organisation thinking about connecting Dots to workplace systems would need to think carefully about data protection, human oversight, and procurement rules before doing so. The broader question — about what always-on AI agents mean for knowledge-work jobs across the county — is one that’s worth watching as availability expands.