OpenAI has turned the idea of an AI assistant that never switches off into a product called dots. TechCrunch's Tim Fernholz published an interview with the OpenAI executive behind it on 26 October 2026, and the details matter for any business thinking about AI agents. Here is what dots are, what is confirmed, and what to do about it.
What dots are
OpenAI announced dots on 29 September 2026 and describes them as "always-on agents built to handle everything." A dot is powered by GPT-6 Astra, has its own cloud computer and browser, learns from your feedback and can work towards your goals around the clock. Through plugins, OpenAI says a dot can connect to over 4,000 apps.
You talk to your dot in ChatGPT on desktop, web and mobile, in Slack and Teams, or by voice call. It can also message you with progress, questions or decisions. OpenAI describes it as one ongoing conversation instead of many separate chats.
What makes it "always-on"
Most AI tools wait for a prompt. A dot also works when you are not asking. OpenAI calls this proactive research: on a regular cadence, the dot thinks about what it knows about you and the apps you connected, researches how it could help, and then proposes it.
OpenAI's own examples include a dot that noticed an early tester had forgotten to invoice a publication, prepared the invoice and sent it after his approval. Alexander Embiricos told TechCrunch about an engineer at a drone delivery startup whose dot spotted and analysed an outage before the engineer noticed it. TechCrunch says that in its own testing the ability to think ahead was apparent and often useful, but not always timely: it flagged calendar conflicts and deadlines, though not necessarily early enough.
Who can get dots, and what it costs
- Available now: OpenAI says dots are rolling out in ChatGPT to Pro and Business Premium users in eligible markets, and Enterprise users can try a beta once their workspace admin enables it.
- Price: the first dot is included in a Pro or Business Premium plan at no extra cost, per OpenAI. TechCrunch reports that qualifying plans start at $100 a month.
- Later: OpenAI says you will be able to add more dots and scale each one's speed or monthly workload.
- Specialist dots: a preview for organisations, with their own identity and credentials, aimed at defined jobs such as procurement, invoice processing, customer support and contracts. OpenAI is starting with focused enterprise pilots and is working with Microsoft on Agent 365 governance.
The coding-tools connection
The TechCrunch headline says OpenAI "turned its coding tools into dots." Embiricos said you could achieve much of what dots do with OpenAI's Codex coding app, but it takes more work and hand-holding. TechCrunch notes that OpenAI had already spun ChatGPT Work out of the agentic tools it built for software engineers. In Embiricos's words, much of the product work was not new technology but "how you frame it to the user and how it feels."
The privacy and control questions
An assistant that is useful needs access to your calendar, email and files. That is the central trade-off, and OpenAI and TechCrunch both address it.
- What OpenAI says: each dot works on its own cloud computer, and your computer stays separate unless you choose to connect it. Proactive research uses read-only tools that cannot send messages or change content. Custom Rules let you allow, require approval for, or block specific actions, and certain sensitive tasks, such as changing a password, always stay with you.
- Training data: OpenAI does not use content from Business, Enterprise or Edu workspaces to improve its models by default. On personal plans you can control whether dots' conversations are used.
- What TechCrunch found: the controls are not yet granular, and the app refused read-only access to cloud drives. Embiricos said OpenAI needs to build "super, super, super granular controls" and that early defaults were conservative, so the dot asked for permission often.
- Payments: Embiricos said OpenAI is in conversations with payment processors, but letting an assistant spend money is not available today.
OpenAI's own caution applies: dots can still make mistakes, so always review consequential work.
What businesses should do now
- Pilot one narrow job. Pick a task with a clear definition of done, such as drafting invoices or triaging a support inbox.
- Keep approvals on. Start with read-only access and require sign-off for sending messages, editing records or anything involving money.
- Decide what data it may see. Connect only the apps the task needs, and check your plan's training and data-retention settings.
- Know when a custom build is better. A general assistant is flexible. A fixed process that runs hundreds of times a day, such as order updates or lead routing, is usually cheaper and more predictable as a purpose-built workflow automation with logging you control.
- Expect humans on the other side. Embiricos told TechCrunch he has a human assistant because OpenAI's corporate partners want to deal with a person. Customer-facing work still benefits from a human option.
For a plain-language primer on the category, read AI agents for business explained. If you want an assistant that answers customers on your own site, see how to build a website chatbot and our AI automation service.
Talk to us
Susraj Tech builds AI agents, chatbots and automations that connect to the tools a business already uses, with approvals and audit trails designed in. Tell us the process you want to hand off and we will suggest whether a ready-made assistant like dots or a custom build fits.
Sources: OpenAI, "Introducing dots" (29 September 2026) and TechCrunch, "How OpenAI turned its coding tools into dots, an always-on personal assistant" (Tim Fernholz).
