OpenAI has launched OpenAI Dots agents, a new class of autonomous AI assistants powered by GPT-6 Astra, designed to work independently in the background on user-defined tasks with minimal oversight. Announced at the company’s DevDay event on Tuesday, Dots represent a shift toward always-on, goal-oriented agents that operate across multiple platforms and interfaces rather than being tied to a single application.
The OpenAI Dots agents are now available for ChatGPT Pro and Business Premium users in eligible markets, with access through both ChatGPT and Codex. Users can create their first dot, name it, and begin delegating tasks immediately. The company envisions teams of Dots working together on behalf of users over time, each handling specific responsibilities autonomously.
How OpenAI Dots Agents Work in Practice
Unlike earlier agentic tools, OpenAI Dots agents are built to pursue objectives continuously without constant user direction. OpenAI has outlined several real-world scenarios: a software developer might deploy a dedicated dot to monitor customer feedback and automatically implement bug fixes and requested features, while a scientist could use a dot to rerun analysis and investigate unexpected results as new experimental data arrives. Users can message their Dots through Slack, Teams, and other organizational platforms, with text message support coming soon.
The system also supports “specialist Dots” that take on specific roles within organizations. Individual agents can be provisioned with custom identities, credentials, and tools through existing systems. OpenAI is already collaborating with Microsoft to integrate Dots into the company’s Agent 365 security controls, ensuring enterprise-grade protection and governance.
Design and the Broader AI Agent Landscape
The Dots interface features a bubbly, cartoonish visual design—small floating animated characters that give the agents a more relatable personality. This branding approach mirrors similar efforts by other companies to humanize AI; as covered earlier, Perplexity launches local AI agent for high-end NVIDIA hardware demonstrated comparable moves toward making autonomous agents feel more approachable. In a related development, Google AI Mode adds flight price tracking and hotel booking tools showed how other AI platforms are expanding agent capabilities into practical user workflows.
While much of Dots’ underlying functionality was already possible through Codex and similar agentic frameworks, the new product bundles these features into a cohesive package with a clearer focus on independent, background-running actions. The result is a tool designed to reduce friction between intent and execution, letting users define goals once and let their Dots handle the rest.
المصدر: TechCrunch
