OpenAI introduced Dots, a new personal agentic assistant powered by GPT-6 Astra, at its DevDay event on September 29, 2026. These distinctive agents are designed to operate independently of specific hardware or interfaces and continuously pursue user-defined goals with limited supervision. Users can start with a primary Dot, customize it by naming, and eventually build teams of Dots that collaborate on their behalf.

Dots will be accessible to Pro and Business Premium users in eligible markets through ChatGPT and Codex platforms. OpenAI highlights versatile applications, such as developers deploying Dots to monitor customer feedback and implement changes, or scientists using them to analyze experimental results dynamically. Communication with Dots can happen through organizational tools like Slack and Teams, with text messaging support planned for the future.

What sets Dots apart from previous agent-based tools like Codex is their branding and packaging, featuring bubbly, cartoon-like avatars that aim to create a relatable and approachable AI experience. This approach aligns with similar recent launches like Meta’s Muse agents and marks a new step in making AI systems more personable and engaging.

OpenAI is also enhancing Dot functionality by enabling specialist agents with specific identities, credentials, and tools integrated via existing systems. Collaboration with Microsoft includes integrating Dots into Agent 365 security frameworks, signaling broader enterprise adoption potential. Overall, Dots promise continuous background assistance aligned with complex, evolving user objectives.