OpenAI unveils Dots, autonomous AI agents that work around the clock

OpenAI has introduced Dots, a fresh category of autonomous AI agents designed to operate continuously in the background while pursuing goals on behalf of users. Unveiled at Dev Day, these agents run on GPT-6 Astra and differ fundamentally from traditional chatbots in their ability to work without constant human input.

What Dots Does#

Unlike previous tools such as Codex or ChatGPT, Dots function as independent workers that can be deployed for specific tasks. Users can create and customize their own primary Dot, and eventually assemble teams of multiple agents working in concert. The system accepts instructions through messaging platforms including Slack and Teams, with SMS support forthcoming.

OpenAI has already identified several practical applications: a developer might assign one Dot to track customer issues and implement fixes autonomously, while a researcher could deploy another to re-examine experimental findings as new data arrives. The platform supports "specialist Dots" configured with particular roles, credentials and integration with existing enterprise tools—Microsoft is collaborating on integrations with its Agent 365 security infrastructure.

Rolling Out Soon#

Dots becomes available starting Tuesday through ChatGPT for Pro and Business Premium members in supported markets. Users can launch them directly from Codex or the ChatGPT interface.

The Humanization Angle#

The agents carry a playful visual identity—small, floating cartoon characters—reflecting a broader industry trend of making AI feel more approachable and less abstract. This approach mirrors recent efforts from competitors like Meta's Muse agent to give AI systems a more engaging persona.