OpenAI’s Dots Are AI Agents Designed to Keep Working

Image: OpenAI

Chatbots wait for instructions. OpenAI's new Dots are designed to keep working.

Introduced at OpenAI's DevDay on September 29, Dots are described as always-on AI agents that can take on ongoing tasks, connect to applications and continue making progress between conversations.

The idea represents a shift from AI that responds to individual requests toward AI that can work toward a broader objective.

From answering to doing

A conventional chatbot might help write an email after being asked. An agent can potentially take the next steps itself.

OpenAI says users can give a Dot a goal, connect the applications it needs and define what it is allowed to do. The agent has its own cloud computer and can work on tasks before returning results for the user to review.

That could make AI useful for workflows that normally require repeated instructions, such as organizing information, monitoring projects or preparing work across multiple applications.

Control becomes more important

The more freedom an AI agent receives, the more important its boundaries become.

OpenAI says Dots allow users to establish controls over what an agent can do. That distinction matters because an AI that can interact with applications has a different risk profile from one that simply generates text.

An agent that can read information is one thing. An agent that can modify documents, communicate with other people or perform actions on a user's behalf requires considerably more trust.

OpenAI's approach therefore puts permissions and user approval alongside capability.

A different kind of AI interface

Dots also reflect a broader change taking place across the AI industry.

Instead of opening an AI service every time something needs to be done, users could increasingly delegate ongoing work to an agent and check the results afterward.

That does not mean agents can reliably handle every task without supervision. The technology is still developing, and real-world workflows introduce problems that demonstrations cannot always reveal.

But the direction is clear: AI companies are increasingly building systems designed not just to answer questions, but to pursue goals.

Dots are one of OpenAI's clearest attempts yet to turn that idea into a practical product.