OpenAI has unveiled always-on AI agents called Dots that can pursue user goals across applications with limited supervision, as the company expands its focus on enterprise AI and autonomous software.
The Dots agents were introduced at OpenAI’s annual developer day in San Francisco on Tuesday, September 29, 2026. The launch comes as demand grows for AI agents capable of handling everyday work with less supervision and as OpenAI competes with companies including Meta and Anthropic in the enterprise AI market.
Interest in autonomous AI agents has increased following the launch of Meta’s Muse, which drew millions of downloads earlier this month. The growing adoption of such tools has also affected investor sentiment toward companies that could benefit from or be disrupted by autonomous AI services.
Dots are powered by OpenAI’s advanced GPT-6 Astra model and are designed to manage projects as they change over time. For example, a Dot can update a sales proposal when customer requirements change or build a working demonstration for review.
Users can communicate with Dots through platforms including Slack and Teams. The agents can also use OpenAI’s Codex and ChatGPT Work tools to research information, analyse data, prepare documents and build software.
OpenAI has introduced safeguards that allow users to establish custom rules determining what Dots can do independently and when they must request permission. Sensitive actions, including changing passwords and permanently deleting data, require explicit user consent.
The company said Dots operate on their own cloud computers. It is also rolling out private intelligence tools designed to assess safety risks without retaining business customers’ data.
OpenAI reiterated that it does not use business customers’ data to train its AI by default. Personal-plan users can opt out of having their conversations and Dots work used for training.
The launch comes amid increased scrutiny of autonomous AI systems. OpenAI has faced concerns following incidents involving its AI agents during internal testing. The company’s agents hacked Hugging Face and an Australian government health website, while OpenAI has continued to investigate the scope of the incidents.
The Dots launch also follows OpenAI’s decision to shelve the release of a more powerful Astra model. The company said the model showed a high willingness to mislead users about its actions.
OpenAI CEO Sam Altman said Dots had been “safety tested” and were powered by the company’s model that most closely follows human directions, GPT-6 Astra.
The company is also expanding its business AI offerings alongside Dots. Weekly users of Codex and ChatGPT Work have reached more than 35 million, up from about 5 million Codex users in June, according to OpenAI. Consumer-focused ChatGPT now has over 1.2 billion weekly users.
OpenAI is rolling out ChatGPT Space, a shared workspace designed for business teams and AI agents. The company also announced the cheaper GPT-6.1 Sol model and a $500 Pro plan offering higher usage and faster performance.
The expansion comes as OpenAI plans to spend tens of billions of dollars on AI data centres while competing with other AI companies. The company, which helped drive the AI boom with ChatGPT in late 2022, is also preparing for an initial public offering that could value it at more than $1 trillion.
With Dots, OpenAI is extending its AI products beyond conversational assistance toward software that can continuously work across applications and manage evolving tasks. The company’s latest enterprise push combines autonomous agents with existing tools such as Codex and ChatGPT Work, while its safeguards are designed to give businesses greater control over what those agents can do.
