OpenAI unveiled Dots, its new always-on AI agent platform, at its annual developer conference. These agents, powered by the GPT-6 Astra model, are designed to work continuously on user-defined goals, operating in the background without constant prompting. Dots can connect to over 4,000 applications, allowing them to interact with a wide range of software and information. Users can interact with their Dots through ChatGPT, Slack, and Microsoft Teams, with direct text messaging planned for the future.
The platform is positioned as enterprise software, capable of handling complex tasks such as monitoring customer feedback, identifying software bugs, and preparing fixes. For example, a software developer could deploy a Dot to monitor customer feedback and implement bug fixes, or a scientist could use a Dot to automatically revisit analyses as new experimental data becomes available. OpenAI also highlighted potential uses in sales, marketing, and content production, where Dots could process interview transcripts, draft social posts, or manage project updates.
Each Dot operates on its own private cloud computer, which users can inspect at any time to review its activities. This separation of processing aims to enhance security, with OpenAI stating that Dots build upon ChatGPT's existing security systems with additional safeguards. The company also noted that Dots have built-in protections against malicious instructions, with the AI pausing its work if it detects suspicious activity.
Dots are available to users subscribed to ChatGPT's Pro and Business Premium plans, with one Dot included at no extra cost for these tiers. The service will not be available in the European Economic Area, Switzerland, and the UK due to regional AI regulations. This launch places OpenAI's Dots in direct competition with similar AI agents from companies like Meta, which recently released its Muse agent.