OpenAI's New 'Dots' Are AI Helpers That Work While You Sleep

At its DevDay event on Tuesday, OpenAI introduced Dots, a new kind of AI helper powered by its newest model, GPT-6 Astra. Unlike ChatGPT, which mostly waits for you to type something, a Dot is built to keep working on its own — you give it a name, tell it what you want, and it runs in the background with almost no supervision. OpenAI says you can start with one 'primary dot' today and eventually manage whole teams of them.

The company's own examples are telling. One Dot watches customer complaints for a software team, fixing bugs and adding requested features as they come up. Another reruns a scientist's analysis whenever new data arrives, flagging surprises. You can message your Dots through Slack, Teams and other work apps, with regular text messages coming soon. 'Specialist' Dots can be given their own logins, credentials and tools so they can actually get things done inside systems you already use. OpenAI is working with Microsoft on security controls called Agent 365.

Why this matters: this is the shift from AI that answers questions to AI that does jobs. Much of this was already possible with OpenAI's Codex tool, but Dots packages it into something friendlier and always on — less like a chatbot, more like a junior colleague who never sleeps. For busy professionals, that could mean hours of admin, research and follow-up handled without you. It also means trusting software with access to your accounts, files and inboxes.

The catch: Dots are only available now inside ChatGPT for Pro and Business Premium subscribers in eligible markets, so most casual users are locked out. And an assistant that acts on its own can also make mistakes while you're not looking — or go further than you intended. The bubbly cartoon branding, similar to Meta's recent Muse agent, is designed to make you comfortable with exactly that.