OpenAI has introduced a new generation of autonomous AI agents called “dots,” expanding the company’s push from conversational artificial intelligence toward systems capable of carrying out longer-running work across applications.
The agents were unveiled Tuesday at OpenAI’s Developer Day and are powered by GPT-6 Astra.
Unlike conventional chatbot interactions that wait for individual prompts, dots are designed to continue pursuing defined objectives over time and coordinate work involving different applications and tools.
That potentially expands the commercial role of AI from assisting employees with individual tasks toward operating parts of persistent digital workflows.
OpenAI is positioning the technology for enterprise use, where businesses are experimenting with autonomous systems for software development, research, communications and other multi-step processes. Dots can interact with workplace platforms and use OpenAI products including Codex and ChatGPT Work.
OpenAI’s deployment-safety documentation describes the systems as always-on agents built on existing model and agent capabilities and powered by GPT-6 Astra.
The move intensifies competition for enterprise AI spending.
Meta has also expanded its enterprise strategy, while software and cloud companies are racing to build agents capable of performing increasingly complicated digital tasks with less continuous human intervention.
That autonomy also raises a separate issue: control.
OpenAI has introduced administrative controls governing access, computer capabilities and connected applications for eligible enterprise workspaces. Workspace owners can determine who may use dots and which capabilities are available.
Those controls matter because the economic value of autonomous agents depends partly on their ability to access real business systems. Greater access can make agents more useful, but it also increases the importance of permissions, monitoring and authorization boundaries.
For businesses, the next test will be whether always-on agents can reliably perform meaningful work at scale rather than simply demonstrate more autonomy.
For the AI industry, dots mark another step toward competition over systems that do work rather than merely generate responses.
