OpenAI's dots: an always-on assistant grown from its coding tools

OpenAI has put an assistant in ChatGPT that keeps working when you are not talking to it. Called a "dot", it runs on GPT-6 Astra, has its own cloud computer and browser, and can work towards your goals around the clock, according to OpenAI ^[1]. A new interview shows how the idea grew out of the company's coding tools ^[2].

What a dot does

OpenAI describes dots as always-on agents that learn from feedback and connect to more than 4,000 apps through plugins ^[1]. You can reach yours in ChatGPT on desktop, web and mobile, or in Slack and Teams, with texting listed as coming soon ^[1]. You can open the dot's computer at any time to inspect its work, and you can also allow it to use your laptop ^[1].

When you are not actively working with it, a dot looks for ways to help in the background. OpenAI calls this "proactive research" and says it uses connected apps through read-only tools, so it cannot send messages, change content or control your browser or computer ^[1].

One example from OpenAI: an early tester's dot noticed he had forgotten to invoice a publication, prepared the invoice and sent it after his approval ^[1].

From coding tool to assistant

Alexander Embiricos, who leads OpenAI's enterprise product team, told TechCrunch's Disrupt event that you could achieve much of what dots do with Codex, OpenAI's coding app released at the start of this year, but that it takes more work and hand-holding ^[2]. TechCrunch reports that OpenAI had already spun ChatGPT Work out of the agentic coding tools built for software engineers, and dots continue that path ^[2]. OpenAI's launch material says a dot can start tasks in Codex and ChatGPT Work, and those tasks count towards usage limits as usual, while simply talking to a dot does not ^[1].

TechCrunch's own testing found the dot's ability to think ahead "apparent and often useful", but not quite there yet: it flags calendar conflicts and deadlines, though not always in a timely way ^[2]. Embiricos said his own dot once warned him of a conflict three minutes before it started, and expects the timing to improve ^[2].

Control and privacy

OpenAI says users can set Custom Rules to allow, require approval for or block specific actions, and that sensitive tasks such as changing a password always stay with the user ^[1]. It also warns that dots can still make mistakes, so consequential work should be reviewed ^[1].

TechCrunch notes that the privacy controls are not granular and that the app can be pushy; it refuses to accept read-only access to cloud drives, for example ^[2]. Embiricos said OpenAI wants to build "super, super, super granular controls" ^[2].

Who can get it

Dots are rolling out to Pro and Business Premium users in eligible markets, and Enterprise users can try a beta once an administrator enables it ^[1]. The first dot is included in the plan at no extra cost ^[1]. TechCrunch puts the entry point at the Pro tier, with plans from $100 a month ^[2].

What it means for business

For leaders, the practical question is not the model but the permissions. An assistant that watches your email and calendar and acts unprompted can save time on recurring work, but it needs clear rules about what it may do alone. OpenAI is also piloting specialist dots with their own identity and access for tasks such as procurement and invoice processing, and working with Microsoft on managing them through Agent 365 ^[1]. Embiricos also said OpenAI is in conversations with payment processors, though he did not give details ^[2].

Source: OpenAI — Introducing dots