OpenAI launches dots, an always-on agentic assistant

OpenAI used its Dev Day event on Tuesday to unveil Dots, a new personal agentic assistant powered by GPT-6 Astra. The company describes the agents as capable, always-on systems built to handle tasks continuously with minimal supervision.

Unlike Codex or standard ChatGPT sessions, Dots are designed to operate independently of any single device or interface, running in the background to pursue goals users define upfront rather than waiting for step-by-step prompts.

OpenAI framed the launch as an evolving system, describing plans for users to eventually deploy multiple Dots working together on their behalf, starting with a single primary Dot that can be named and customised.

OpenAI Dots feature availability

Dots became available Tuesday within ChatGPT for Pro and Business Premium subscribers in eligible markets, launchable directly from either Codex or ChatGPT. OpenAI also confirmed users will be able to message their Dots through workplace platforms like Slack and Microsoft Teams, with text message support planned for a future update.

A vision of a software engineer allocating a Dot to listen to customers and fix bugs or implement new features by itself is possible.

An illustration of this is a scientist utilising a dot to repeatedly perform analysis when new data arrives and highlighting any unusual outcomes.

Specialist Dots, where each dot is assigned an identity, credentials, and certain tools according to its function, is another thing that OpenAI brings forward.

It was stated that OpenAI is currently collaborating with Microsoft to incorporate Dots within Agent 365, which is Microsoft's framework for enterprise AI agents.