OPENAI has unveiled Dots, a persistent AI assistant described as online and operational around the clock. Marketed as proactive, Dots can perform tasks in the background and even initiate actions, such as sending a forgotten invoice after the user grants permission. The concept frames Dots as a digital employee with its own cloud computer and browser, capable of handling work across a wide range of applications through plugins that connect to more than 4,000 services.
The service is integrated beyond the ChatGPT interface, notably within Slack and Microsoft Teams, allowing users to talk to a Dot and assign tasks directly within workplace chats. Dots can be accessed at any time to monitor progress and can be authorised to connect remotely to a user’s local PC, enabling collaborative work between human and AI.
Availability starts with Pro, Business Premium, and Enterprise subscribers in select markets, with each eligible user receiving one free primary Dot that supports naming and customization, alongside unlimited conversations inside ChatGPT; Codex and Work have usage caps.
OpenAI emphasises stronger safety boundaries, with Dots’ tools operating in a read-only mode during proactive background research and no autonomous messaging or browser control. Sign-in verification is designed to avoid exposing passwords to the model, and users can create custom rules to govern which actions Dots may perform autonomously versus those requiring human consent.
The article notes implications for a multi-Dot, team-based workflow and raises questions about data boundaries as AI agents move between cloud and local systems.