OpenAI's Dots: Helpers That Keep Working While You're Away
OpenAI launched Dots, a helper that keeps working on its own, on September 29, 2026. TechCrunch
The launch is aimed at business use and at Meta. Dots are always-on agents that chase a user's goal across apps by themselves. Think of a helper who keeps doing chores after you ask once. Reuters The point is steady work across apps, not just one chat reply.
OpenAI shared the news on openai.com in a post called 'Introducing dots'. It calls Dots "remarkably capable, always-on agents built to handle everything." OpenAI
You make a dot in the ChatGPT desktop app or in ChatGPT on desktop web by following the setup steps. OpenAI Help OpenAI says a dot keeps work moving across tools and projects while you can still talk to it. OpenAI Developers Setup stays inside ChatGPT. You talk by chat while it does several jobs at once.
Dots has a bubbly, blobby cartoon face. The name is crowded.
The web address dot.com belongs to Elon Musk's xAI. Right now it sends visitors to the download page for xAI's Grok chatbot app. Records show dot.com changed hands in July, as reported on September 29, 2026. The address dots.com belongs to a company that closed long ago. TechCrunch
Separately, old copy with no date on dot.com talks about "Bots like teammates" and a product called Grok Bot. That copy says Grok Bot logs in once and uses your apps and websites the way you would. It also says Grok Bot comes with certain Cursor, SuperGrok, or Teams plans. That wording is separate from the dated news about the redirect and should be treated as background until xAI explains current plans.
The broader context here is a move from helpers that only answer to helpers that do ongoing work for you. For tech teams that run systems, the practical questions are clear. How does the helper log into other tools for you. Where are the saved logins kept. What written record shows what it did while you were away. What happens when goals overlap or it needs to try again or fails.
Looking at what this means for business use, always-on changes how much trust is needed. A chat helper that writes drafts is easy to control. A helper that keeps working across work tickets, documents, code, and admin tools while you talk to it needs clear limits, speed controls, and stop rules. In my view, this kind of nonstop help only earns real trust if each background step can be checked, traced to the helper, and undone.
Worth flagging for builders is the name clash itself. OpenAI released Dots without owning dot.com or dots.com. One is held by a direct rival in helpers, the other by a closed company. That does not stop sharing through the ChatGPT desktop app, but it makes finding, linking help pages, and controlling the brand harder. Watching my own children learn new tools, the name they remember is the address they type. Exact web addresses still shape that habit.
Looking ahead, if the safety controls work, the benefit is one we have seen with past tech changes. Boring, long jobs leave people's to-do lists. People check results instead of clicks. That is the promise OpenAI is selling with Dots, and why Meta and xAI are building rival helpers at the same time.
