OpenAI Dots

• Watching customer feedback, scoping small fixes, and bringing back tested pull requests to review
• Keeping launch materials in step with a changing product scope and preparing revised drafts
• Rerunning an analysis and updating figures when new data arrives
• Preparing an invoice or follow-up it noticed was missing, then sending it after your approval
• Checking your calendar each morning and messaging you what is coming up
• Picking up a bug report in Slack and starting the investigation

Dots are OpenAI's answer to the persistent personal agent, and their main advantage is reach. They inherit ChatGPT's plugin connections and can be messaged in Slack and Teams, so teams already using ChatGPT get an agent without new infrastructure. The controls are more layered than most consumer agents offer: read-only background research, Custom Rules, a separate Auto-review step, and hard hand-offs for passwords and payments. The weak spots are access and memory control. Dots are limited to top plans, Pro users in the EEA, Switzerland, and the UK are left out, and clearing what a dot has learned means resetting it. Treat the launch month as a trial, since the longer-term usage terms are not yet known.

About OpenAI Dots

OpenAI announced dots at DevDay on September 29, 2026 as a separate way of working inside ChatGPT. A regular chat answers a request and stops. A dot takes on an ongoing responsibility, keeps making progress between conversations, and brings results back for review. Each dot is powered by GPT-6 Astra and works on its own cloud computer with its own browser, which you can open at any time to watch or take over. It reaches your tools through ChatGPT's plugin ecosystem, which OpenAI says covers more than 4,000 apps, and plugin permissions are shared with ChatGPT, ChatGPT Work, and Codex. You can message or call your dot in ChatGPT on web, desktop, and mobile, or message it in Slack and Microsoft Teams. Texting is a limited beta for some US Pro users. It can also start Codex cloud tasks and, if you allow it, work on your own laptop through the ChatGPT desktop app. When you are not working with it, a dot does what OpenAI calls proactive research: it reads the apps you have connected with tools restricted to read-only and saves notes, so it can suggest work before you ask. Control rests on several layers. Built-in rules decide when a dot acts alone and when it asks. Custom Rules let you allow, require approval for, or block specific actions. Auto-review checks actions such as sending an email against your instructions before they run, and the most sensitive actions, such as changing a password or transferring money, are handed back to you. Activity View shows ongoing and delegated work. The tradeoffs are access and transparency. At launch dots are limited to Pro (not in the EEA, Switzerland, or the UK), Business Premium, and an admin-enabled Enterprise beta. You start with one primary dot. You cannot view or delete individual memories, only reset the whole dot. And usage terms after the first month have not been published. OpenAI is also previewing specialist dots for organizations, which get their own identity and credentials and are starting as focused enterprise pilots.