OpenAI Dots Review (2026): What It Does Well and Where It Falls Short

OpenAI Dots review: what's the verdict?

Dots are the most ambitious agent OpenAI has shipped, and they're priced and built for heavy knowledge work. If you live in Slack or Teams, juggle a dozen apps and already pay for ChatGPT Pro, a dot looks like a genuine upgrade. If you just want your admin handled, it's a lot of machine for $100 a month.

A quick word on how this review works. Dots launched at OpenAI DevDay on September 29, 2026, and they're rolling out gradually. We haven't used one yet. This is a launch-week review based on OpenAI's launch materials, its help docs and the early-tester examples OpenAI shared. Where OpenAI hasn't said something, we say so. We'll update this page after hands-on use.

That's also why you won't find star ratings or scores out of ten here. Our scorecard further down rates each category as Strong, Mixed or Weak based on what OpenAI has confirmed, and says why.

What is an OpenAI dot and how does it work?

A dot is an always-on personal agent built into ChatGPT. Each one gets its own cloud computer and browser and works toward your goals around the clock, even when your chat is closed. OpenAI's pitch: "remarkably capable, always-on agents built to handle everything."

You name your dot, and it shows up as a bubbly, cartoonish floating dot. You talk to it in the ChatGPT desktop app, web or mobile app (voice included), in Slack or in Microsoft Teams. Texting is a limited US-only beta for Pro, and carrier charges may apply. It remembers context whichever channel you use.

Under the hood it connects to 4,000+ apps through OpenAI's plugin ecosystem, can use your own laptop with your permission, and can sign into supported websites with saved passwords the model never sees. It can also hand bigger jobs to Codex and ChatGPT Work. If you're weighing it against the tools you already have, see how dots differ from ChatGPT agent and Work.

OpenAI's showcase example is refreshingly ordinary. An early tester, a freelance writer, had forgotten to invoice a publication. His dot noticed, pulled the details from the email thread, drafted the invoice and sent it after he approved. Inside OpenAI, a bug posted in Slack gets dots investigating, and dots have turned a design into a working app.

Who can get one: users 18 and over on any ChatGPT Pro tier or Business Premium, in eligible markets, as the gradual rollout reaches them. Enterprise, Edu and Healthcare workspaces get dots as a beta that stays off until an admin enables it. OpenAI is also piloting specialist dots for enterprises, with their own organizational identity and access to company systems, and talks about "teams of dots working together on your behalf."

What do OpenAI dots do well?

On paper, dots get the hard parts of an agent right: persistence, reach, memory and control. Here's what stands out from OpenAI's launch materials.

Always on, on its own computer. Most assistants stop when you stop typing. A dot has its own cloud computer and browser and keeps working 24/7. Activity View lets you watch that work any time, with In progress, Scheduled and Completed tabs, and step in, pause or reset.

Huge reach. 4,000+ app connections covers most of the tools a modern team uses. Add Slack, Teams and ChatGPT Space, where several people edit one Page and can tag a dot in, and it fits into how many teams already work.

It learns you. A dot picks up your preferences, standards and working habits from feedback over time. That's the difference between an assistant you re-brief every morning and one that already knows your invoice template.

Sensible guardrails. Proactive research is read-only: while it scans your apps in the background, it can't send messages, change content or control your computer. Custom Rules let you allow an action, require approval or block it. An auto-review system checks consequential actions, sensitive tasks like changing passwords always stay with you, and a monitoring system can pause or stop a dot over safety concerns.

Included, not bolted on. Your first dot comes at no extra charge with every Pro tier and Business Premium. Chatting with it doesn't count toward your usage limits, and for the first month dot usage doesn't count toward eligible plan allowances either.

Where do OpenAI dots fall short?

The price floor, the map and the channels. None of these are about how smart a dot is. They're about who can actually get one and where they can use it.

$100 a month, minimum. Dots aren't on Free, Go or Plus. The cheapest route is Pro Standard at $100 a month, or $1,200 a year, or a Business Premium seat at $100 a month billed annually ($125 monthly). Pro More is $200 and the new Pro 500 is $500. More dots, extra speed and more workload will cost extra later, and OpenAI hasn't said how much. Tasks your dot runs in Codex or ChatGPT Work also count toward your allowance, so a busy dot can eat into the limits you pay for.

Not in Europe on Pro. Pro dots exclude the European Economic Area, Switzerland and the UK at launch, with no timeline. Business Premium does include Europe and the UK, but that's a business plan, not a personal one.

Desktop-only setup, and no WhatsApp. You have to create your dot in the ChatGPT desktop app or on desktop web; you can't set one up on your phone. After that, it's ChatGPT, Slack and Teams. SMS is a limited US-only Pro beta, and WhatsApp isn't a channel at all. OpenAI says other messaging apps are coming "soon", with no date. Dots also can't call you and don't have their own email address.

It will make mistakes. OpenAI says so plainly. That's honest, and it's why the rules matter: an agent with access to 4,000+ apps and your saved logins needs you to set approvals before it acts.

Launch-day tier confusion. Early reports from TechCrunch and BGR said dots were only on the $200 Pro tier. An OpenAI staffer then clarified that all Pro plans include a dot, $100 included. Add the Pro More cuts (Codex and Work allowance down from 20× to 10× Plus, GPT-6 Pro messages from 200 to 100 a week) and a lot of subscribers spent launch day confused about what they're paying for.

OpenAI dots scorecard: how does it rate by category?

Strong on capability and controls, mixed on access, weak on price transparency. These ratings are qualitative and based on OpenAI's launch materials, not on our own testing, so treat "reliability" in particular as an open question.

The pattern is typical of a big launch. The product is designed thoughtfully, and the business side (plans, regions, what costs extra) is still settling. Expect several of these ratings to move in the coming weeks, and we'll update them once we've run a dot ourselves.

What we'll check hands-on: whether proactive research surfaces useful loose ends or mostly noise, how often approval requests interrupt your day, how quickly a busy dot eats into your Codex and Work allowance, whether preferences taught in Slack really carry over to ChatGPT, and how clean a reset actually is. Those answers will decide whether the "Unknown" on reliability turns into Strong or Weak.

Is OpenAI Dots worth it? Who should buy and who should skip

Worth it if you already need ChatGPT Pro. Hard to justify if you'd be buying Pro just for the dot. Since the first dot is included, the real question is whether Pro itself earns its $100 or more a month.

Buy it (or switch it on) if you're already on Pro or Business Premium, your work runs through Slack, Teams and many SaaS apps, and you'd love a teammate that notices loose ends without being asked. Developers, operators, consultants and managers with too many threads are the obvious fit. Use the first month, when dot usage doesn't count toward eligible allowances, to see how much it actually does.

If you do try it, start tight. Set Custom Rules to require approval for anything that sends, pays or deletes, connect only the few apps you need, leave laptop access off unless a task demands it, and check Activity View daily for the first week. You can loosen the rules once you trust what it does.

Wait if you're on Pro in the EEA, Switzerland or the UK (you can't get one yet), if you want to know what extra dots will cost before you build around them, or if you'd rather someone else find the rough edges first.

Skip it if you're on Free, Go or Plus and your needs are everyday admin: reminders, follow-ups, chasing people, reading the odd contract. You'd be paying at least $1,200 a year for a cloud computer you don't need, and it still wouldn't live in the app where most of that admin happens: WhatsApp. For more options, see our best OpenAI dots alternatives.

What should you use instead if you skip dots?

If your admin lives in WhatsApp, use an assistant that lives there too, like Agent Sonic. Most of the "skip" group isn't short on AI. They're short on follow-through, and their follow-through happens in chats with clients, suppliers, staff and family.

Sonic is a personal AI assistant that lives entirely inside WhatsApp. No app, no dashboard, no desktop setup: you save it as a contact and message it, by text or voice note.

Here's what that looks like. "Remind me to chase the Hendricks invoice every Tuesday until I say stop." "Ask Dana if the contract is signed and tell me what she says." In your team group: "Every Sunday at 9, remind everyone to post their weekly numbers." Forward a supplier's price list and ask "Which items are over $50?" Or send a voice note on the move: "Add Friday's site visit to my calendar at 10."

It also keeps notes, to-dos and a memory of what you tell it, browses and searches the web live, creates and edits images, works with Google Calendar, Sheets and Docs, and replies in your language, from English to Hebrew to Arabic and Arabizi. It works wherever WhatsApp works, Europe and the UK included.

Honest limits: Sonic doesn't make purchases, doesn't have 4,000 integrations, doesn't run a cloud computer for hours of autonomous coding or research, and isn't in Slack or Teams. If that's your work, a dot is the stronger tool. If your work is chats, reminders, follow-ups and documents, Sonic fits. Get access at tryagentsonic.com/contact.

Frequently asked questions