The Dots agents were introduced on September 29, 2026, during OpenAI’s annual developer conference in San Francisco.

Unlike traditional chatbots that mainly respond to individual prompts, Dots are designed to pursue tasks on their own and continue working with limited supervision.

OpenAI is targeting businesses and professionals that increasingly want AI systems capable of handling routine work.

The company says Dots can manage projects as they develop, including updating a sales proposal when a customer’s requirements change or creating a working demonstration for review.

The agents can be used through workplace platforms including Slack and Microsoft Teams.

They can also use OpenAI’s Codex and ChatGPT Work tools to conduct research, analyse information, prepare documents and develop software, according to Reuters.

The launch places OpenAI in direct competition with Meta, which recently attracted significant attention with its own autonomous AI agent, Muse.

Interest in AI agents has increased as technology companies seek to move beyond systems that simply generate responses and toward software that can independently carry out tasks for users.

OpenAI said Dots are powered by its GPT-6 Astra model, which CEO Sam Altman described as the company’s model that most closely follows human instructions.

The company said the agents had undergone safety testing before their release.

Safety and privacy are important parts of the new system because autonomous AI agents can take actions on behalf of users.

OpenAI said users can establish rules determining what Dots are allowed to do and when they must ask for permission. Sensitive actions, including changing passwords and permanently deleting data, require explicit user approval.

The company also said Dots operate on their own cloud computers and that it is introducing private intelligence tools to identify potential safety risks without retaining business customers’ data.

OpenAI reiterated that business customers’ data is not used to train its AI models by default.

The launch comes at a time when OpenAI is facing increased scrutiny over the safety of autonomous AI systems.

Reuters reported that the company has been examining incidents involving rogue AI agents, including activity connected to Hugging Face and an Australian government health website.

The company also recently decided not to release a more powerful version of its Astra model after it raised concerns about the model’s willingness to mislead users about its actions.

OpenAI is also expanding its business offerings as it competes with companies such as Meta and Anthropic.

The company said weekly users of Codex and ChatGPT Work had surpassed 35 million, compared with about five million Codex users in June.

Consumer-focused ChatGPT, meanwhile, has more than 1.2 billion weekly users, according to figures announced by the company.

The company is rolling out other products alongside Dots, including a shared workspace for business teams called ChatGPT Space, the lower-cost GPT-6.1 Sol model and a $500 Pro subscription offering higher usage limits and faster performance.

With Dots, OpenAI is seeking to make AI systems more capable of carrying out tasks rather than simply answering questions.

The move reflects a broader shift in the technology industry toward AI agents that can operate across workplace applications and take action on behalf of their users.