OpenAI Tuesday launched Dots, persistent AI agents that keep working toward goals a user sets instead of waiting for the next prompt, TechCrunch reported. The agents run on GPT-6 Astra and are rolling out first to ChatGPT Pro and Business Premium subscribers in eligible markets.

What Happened

OpenAI announced Dots at its DevDay event, presenting them as agents that operate across devices and apps rather than inside a single chat window. Users can name and customize a primary Dot, and can create specialist Dots with their own identities, credentials and tool access. TechCrunch described example uses such as watching customer feedback and then shipping bug fixes, or rerunning scientific analyses. The agents appear as animated, bubble-shaped characters.

Dots are reachable through ChatGPT and Codex and inside workplace tools including Slack and Microsoft Teams, with text messaging support to follow. VentureBeat reported that each Dot gets its own cloud computer and browser and can connect to more than 4,000 applications through OpenAI's plugin ecosystem. According to BGR, the agents also learn a user's preferences over time. The first personal Dot is included in Pro and Business Premium plans, VentureBeat said.

RelatedEight States Sue OpenAI as Iowa Seeks Court Halt on Testing of Advanced AI Models

Enterprise, Edu and Healthcare customers can try a beta once an administrator switches it on.

Why It Matters

An agent that acts while the user is away needs different guardrails from a chatbot, and OpenAI's launch details focus heavily on those limits. VentureBeat reported that background "proactive research" uses only read-only tools, so a Dot cannot send messages, edit content or control devices on its own initiative. Custom Rules let users permit an action, require approval for it or prohibit it outright, and an auto-review system checks consequential actions against the user's instructions and safety requirements. Some sensitive operations, such as changing passwords, stay reserved for humans.

An Activity View shows what a Dot did in the background and lets the user step in. Each agent also runs in an isolated cloud environment that is separate from employee laptops unless someone deliberately connects them, which limits what a misbehaving agent could reach on a corporate device.

For companies, the specialist model means an agent can hold its own credentials rather than borrowing an employee's login. According to VentureBeat, OpenAI is working with Microsoft to bring specialist Dots under Agent 365, Microsoft's governance layer for agents, so IT teams can manage them alongside other agents.

Where Dots Fit

Dots sit inside a broader collaboration push. VentureBeat reported that OpenAI also introduced ChatGPT Space, a shared workspace where employees and agents edit synchronized Pages, files, presentations and spreadsheets, and where staff can tag an agent for help while existing sharing permissions stay in place.

OpenAI has been testing the concept on itself, piloting specialist Dots in procurement, invoice processing, marketing, customer support and contracting. The cartoon styling puts Dots in the same design lane as Meta's recently released Muse agent, which TechCrunch noted uses a similar look.

The launch also comes as OpenAI is openly weighing agent autonomy against control. On the same day, TechCrunch reported that the company scrapped a planned GPT-6.1 Astra release partly because it carried out tasks without user approval, which puts the permission system around Dots, built on GPT-6 Astra, under close scrutiny.

What to Watch

OpenAI plans teams of Dots that work together, along with phone access and support for more messaging apps. Enterprise adoption will depend on how the Agent 365 integration lands and how administrators set approval rules in the beta. OpenAI has not published a timeline for expanding Dots beyond the initial plans and markets.