OpenAI Launches Dots: Always-On AI Agents That Work 24/7 Across Your Apps

The headline

At its annual Developer Day in San Francisco on September 29, OpenAI introduced Dots — always-on AI agents that keep working on your goals even after you've closed the chat window. Powered by GPT-6 Astra, each dot runs on its own cloud computer, connects to over 4,000 apps through OpenAI's plugin ecosystem, and learns your preferences over time so finished work arrives the way you would have done it — sometimes before you think to ask.

Dots are rolling out now to ChatGPT Pro, Business Premium, and Enterprise users in eligible markets. Your first dot is included in your plan at no extra cost, with an allowance for deeper work and extended limits for the first month after launch.

How Dots actually work

Unlike a chatbot that waits for a prompt, a dot pursues objectives across applications. It has its own cloud computer, its own browser, and access to the apps you connect. It can juggle multiple projects at once — you hand it new tasks without managing separate threads or directing every step.

OpenAI's examples from internal use are concrete: a bug appears in Slack and dots immediately start investigating. A new design arrives and dots turn it into a working app while the team focuses on customer feedback. A planning cycle gets underway and dots keep everyone in sync toward the deadline. One early tester's dot noticed he had forgotten to invoice a publication, prepared the invoice, and sent it after his approval.

Dots live where you work: ChatGPT on desktop, web, and mobile, plus Slack and Microsoft Teams, with texting coming soon. Context carries across every channel. When you're not actively working with it, a dot does "proactive research" using connected apps — but with read-only tools that cannot send messages, change content, or control your browser.

Control and safeguards

Because always-on agents raise governance questions, OpenAI built boundaries into the product: your computer stays separate unless you connect it, saved passwords can sign into sites without exposing them to the model, and Custom Rules let you allow specific actions, require approval, or block them. An auto-review layer checks consequential actions against your instructions, and monitoring can pause or stop a dot if it detects a safety concern. Activity View lets you follow background work and redirect as needed.

For organizations, specialist dots get their own identity, credentials, and access to systems of record for well-defined responsibilities — procurement, invoice processing, email marketing, customer support, and commercial contracting were OpenAI's early test areas. OpenAI is also working with Microsoft to integrate specialist dots with enterprise governance in Microsoft Agent 365.

What this means for your business

The shift from reactive chatbots to persistent agents changes the economics of automation. The expensive part of most workflows is not the execution — it's the coordination, follow-up, and status tracking between steps. An always-on agent absorbs that overhead, which means the ROI case for AI agents just expanded from "answer faster" to "operate continuously."

But persistence makes permissions, auditability, and approval rules part of the product decision, not an afterthought. Any team piloting Dots or specialist dots should define upfront which actions require human approval and which are safe to run autonomously — the same governance discipline serious enterprises already apply to RPA.

What to do next

If you're on ChatGPT Pro, Business Premium, or Enterprise, try assigning your dot one recurring responsibility this week — invoice chasing, meeting prep, competitor monitoring — and review its work daily. Measure the time it saves before expanding. For larger organizations, watch the Microsoft Agent 365 integration: governance through tools you already manage will be the trigger for enterprise-wide adoption.