OpenAI Launches Round-the-Clock Agent Dot, Marking a New Era for AI Collaboration

Overview

OpenAI officially launched the 24/7 agent Dot on September 29, 2026, marking a new stage in human-computer collaboration. Powered by the latest GPT‑6 Astra model and equipped with a dedicated cloud computer, Dot can stay online around the clock to handle various complex tasks. By integrating over 4,000 app plugins, Dot aims to deeply understand user preferences and proactively share important workloads, freeing users from tedious execution so they can focus on more creative core tasks.

Core Capabilities and Work Model

Dot has the ability to autonomously advance projects while maintaining efficiency even when handling multiple tasks simultaneously. It can interact with users in real time through multiple channels such as ChatGPT, SMS, Slack, or Teams, and automatically syncs context across different platforms to ensure a seamless workflow.

• Continuous Learning and Adaptation: Dot can continuously evolve from long-term feedback, accurately grasping users' work standards and thinking habits, and even completing anticipated tasks before they are explicitly requested.
• Omnichannel Collaboration: Users can initiate conversations anytime on mobile or desktop, and Dot will proactively report progress, ask questions, or seek approval for key decisions.
• Coding and Testing: It possesses the ability to write and test code, converting customer feedback into verified code fixes accompanied by video demonstrations of the changes.

"Your Dot can take over a project and advance it autonomously, without requiring you to switch back and forth between different conversations or provide step-by-step guidance." —— OpenAI

Security Mechanisms and Permission Control

OpenAI emphasizes that users have complete control over Dot, which includes built-in layers of security and privacy safeguards. Each Dot runs in an independent cloud environment, keeping personal devices and data isolated unless the user actively authorizes access.

1. Tiered Permission Management: Users can customize the scope of applications Dot can access and set clear operational rules. For sensitive operations such as sending messages or modifying app content, Dot must undergo review and receive approval before execution; for background "proactive research," its permissions are strictly limited, preventing it from directly controlling browsers or sending messages.
2. Security Interception and Pausing: The system includes built-in defense mechanisms against malicious commands. If potential security risks are detected, Dot will automatically pause or stop working.
3. Data Privacy Protection: For enterprise users, content within their workspace is not used by default to improve models. Individual users can choose whether to allow Dot's work content to optimize models, but proactive research notes are generally not included in training data.

Enterprise Deployment and Prospects

In addition to agents for individuals, OpenAI also introduced the Dedicated Dot, designed specifically for designated responsibilities within organizations. These Dots have independent organizational identities and system access permissions and have undergone internal testing in scenarios such as procurement, invoice processing, and customer support. OpenAI is collaborating with Microsoft to integrate Dedicated Dots into Agent 365 and enterprise identity management systems, enabling IT teams to govern and monitor them using existing tools.

Currently, eligible Pro and Business Premium users in qualified markets can gradually experience the Dot service, with their subscription plans already including usage credits for the first Dot. In the future, users can pay monthly to add more agents or upgrade their workload capacity.