OpenAI Debuts 'Dots' Agentic Assistants

At its DevDay event on Tuesday, OpenAI announced the launch of Dots, a series of "always-on" personal agentic assistants. Powered by the GPT-6 Astra model, Dots are designed to operate independently of any specific hardware or interface, utilizing a cloud computer to access a web browser and over 4,000 supported apps, according to The Verge.

Unlike previous tools, Dots are intended to pursue user-defined goals in the background with minimal oversight. OpenAI envisions a variety of use cases: a software developer could deploy a dot to monitor customer feedback and implement bug fixes, while a scientist might use one to investigate unexpected experimental results, TechCrunch reports. Additionally, content creators could use Dots to clip interview transcripts and write social posts, per The Verge.

Users can interact with these agents via voice calls or a text-like interface within ChatGPT, as well as through Microsoft Teams and Slack. OpenAI is also collaborating with Microsoft to integrate Dots into Agent 365 security controls, TechCrunch reports. While users can currently only create one Dot, OpenAI plans to eventually allow teams of agents to work together.

To manage autonomy, OpenAI has implemented "built-in rules" and an "auto-review" feature to ensure actions comply with safety requirements and custom user restrictions, The Verge reports. The agents are represented by bubbly, cartoonish avatars, a design choice TechCrunch notes is similar to Meta's recently released Muse agent.

Dots are available starting Tuesday for ChatGPT Pro, Business Premium, and Enterprise users in eligible markets. OpenAI is also testing "specialist Dots" tailored for specific organizational roles, according to TechCrunch.