OpenAI Launches Dots: Always-On AI Agents That Work Around the Clock

OpenAI has introduced dots, a new category of always-on AI agents designed to work on complex tasks beyond a traditional chatbot interaction.

Powered by the GPT-6 Astra model, dots can be assigned goals and continue working independently using their own cloud computers and connections to more than 4,000 applications. Users can check progress, provide additional instructions and interact with their agents while tasks are underway.

The idea is to move from AI that responds when prompted to AI that can continue working toward a goal over time. Potential use cases include software development, research, scheduling, data analysis and other multi-step workflows.

OpenAI says dots can also learn user preferences over time, potentially reducing the need to repeatedly explain how tasks should be completed. Organizations can create specialist dots tailored to particular business functions and workflows.

The company is initially rolling out dots to ChatGPT Pro, Business Premium and Enterprise users, with the first dot available at no cost under the launch offering.

Alongside dots, OpenAI has introduced additional developer-focused capabilities, including the GPT-6.1 Sol model, Codex Cloud and an Agents API designed to help developers build and deploy agentic workflows.

The launch reflects a broader shift across the AI industry toward autonomous systems that can use software, access information and complete multi-step work rather than simply generate responses.

For businesses, always-on agents could eventually handle repetitive research, monitoring, coding and operational tasks while humans focus on decisions and oversight.

Source: OpenAI

Akash Takes: Dots represents a shift from “ask AI and get an answer” to “give AI a goal and let it work.” If always-on agents become reliable, the next generation of AI productivity may be measured less by how quickly a model responds—and more by how much meaningful work it can complete while you are away.