OpenAI Dots are always-on AI agents launched on September 29, 2026, that work on your behalf 24/7. Picture this. You wake up, grab your phone, and find a message from a small cartoon blob. It has already fixed a bug that showed up in Slack overnight, drafted the invoice you forgot to send, and reworked your launch plan because the product scope changed at 6 p.m. yesterday.
That is the pitch behind Dots, OpenAI’s newest product. It landed on September 29, 2026, at the company’s DevDay event. Whether you’re a business exploring AI agent services or just curious about what’s next, Dots offer a preview of how autonomous agents could soon handle real work on your behalf.
Is it the future of work or a very cute way to hand your digital keys to a machine? Let’s dig in.
What Are OpenAI Dots?
OpenAI calls dots “remarkably capable, always-on agents built to handle everything.” Under the marketing gloss, a dot is a personal AI agent that runs continuously in the background, pursuing goals you’ve set instead of waiting for you to type a prompt.
Here’s what’s under the hood, going by OpenAI’s announcement:
- Powered by GPT-6 Astra, the company’s latest flagship model.
- Its own cloud computer, with a browser and apps, which you can open at any time to check its work.
- Connections to 4,000+ apps through OpenAI’s plugin ecosystem.
- Learning from feedback, so over time it picks up your preferences and standards.
- Multiple ways to reach it: ChatGPT on desktop, web and mobile, plus Slack and Teams. Texting is coming soon. You can also just hop on a voice call.
Much of this was already possible through Codex and similar agent tools. The difference is the packaging. Dots bundle those abilities into one persistent assistant built around independent action. And yes, they look like dots, little floating cartoons with names and colors. The branding is being called a bubbly, cartoonish persona, a familiar move as AI companies try to make their products feel relatable.
What Can OpenAI Dots Do? Real-World Use Cases
OpenAI’s launch post sketches several scenarios. A few of them made me sit up:
The developer. Your dot watches customer feedback for recurring requests, scopes small fixes, builds and tests them, and hands you finished pull requests with videos showing the changes. You keep working on the big feature while the small stuff gets handled.
The scientist. As new lab data arrives, your dot reruns the analysis, investigates surprising results, updates the figures for your paper, and flags what needs your eyes.
The sales lead. Your dot compares customer requirements against product docs, builds a proof of concept for a key integration, and even suggests a solutions engineer who has handled a similar deal.
The content creator. When an interview transcript lands, your dot picks clip moments, writes show notes, and drafts social posts for your approval.
My favorite example is the smallest one. An early tester noticed he had forgotten to invoice a publication, prepared the invoice, and sent it after he approved. It’s a tiny task, but it shows the whole idea. Nobody asked the dot to check for unsent invoices. It noticed.
That is the real shift here. Chatbots wait for you. Dots don’t.
Are OpenAI Dots Safe? Privacy and Control Features
Any tool that works while you’re not looking should make you ask what it’s allowed to do. OpenAI clearly knows this, because a big chunk of the announcement is about control. Here’s how it says it works:
- Separate computers. Each dot lives on its own cloud machine. Your laptop stays off-limits unless you deliberately connect it.
- Passwords stay hidden. When a dot signs into supported websites, it can use saved passwords without exposing them to the model.
- Read-only background work. When you’re not actively chatting, a dot does “proactive research” using connected apps through tools restricted to read-only. It can’t send messages, change content, or control your browser or computer in that mode.
- Rules you set. Custom Rules let you allow certain actions, require approval for others, or block them entirely.
- An Activity View. You can follow what your dot is doing, including background work, and redirect it.
- Auto-review. Actions that could affect your accounts or share information get checked against your instructions and safety requirements.
- A kill switch. If monitoring spots a safety concern, it can pause or stop a dot’s work.
- Human-only tasks. Some things, like changing a password, always stay with you.
Even so, OpenAI adds a line I appreciated for its honesty: dots can still make mistakes, so review consequential work. Treat your dot like a very fast, very eager new hire, not a flawless one.
Specialist Dots for Enterprise and Microsoft Agent 365
Personal dots are only half the story. OpenAI is also previewing specialist dots, agents that take on defined responsibilities inside a company. Each gets its own identity, credentials and access to the systems it needs, and the company sets it up with OpenAI’s engineers. Early testing inside OpenAI covered procurement, invoice processing, email marketing, customer support and commercial contracting.
There’s a notable enterprise partner too. OpenAI is working with Microsoft to fold specialist dots into Agent 365, so businesses can manage them with the governance and security tools they already use. That could matter a lot to IT departments, which tend to hate the words “autonomous” and “unsupervised” appearing in the same sentence.
For now, it’s focused pilots only. OpenAI envisions teams of dots collaborating on your behalf someday, though that’s still a vision rather than a shipping feature.
OpenAI Dots Pricing and Availability
Dots began rolling out on launch day to Pro and Business Premium users in eligible markets, with Enterprise (including Edu and Healthcare) able to try a beta once an admin switches it on. If you’re in Europe or the UK, there’s bad news for Pro users. According to PYMNTS, Pro availability excludes the European Economic Area, Switzerland and the UK. Coverage of the keynote reports that Sam Altman blamed regional regulation, while saying the company wants to ship to Europeans eventually.
Pricing-wise, your first dot is included in your plan at no extra cost. Usage doesn’t count against plan allowances for the first month, and OpenAI says it will share ongoing usage terms afterward. Later, you’ll be able to add more dots or scale each one’s speed and monthly workload. For reference, coverage of the event puts Pro at roughly $100 to $500 a month, so this is a premium-tier feature for now.
What’s Meta’s Muse?
Dots didn’t arrive in a vacuum. Just three weeks earlier, Meta launched Muse, a personal agent it says is “built for everyone.” According to PYMNTS, it topped the App Store charts within days. Then there’s Instinct, a personal-agent startup that reportedly raised $1 billion at a $10 billion valuation the day before dots launched.
Coverage of the keynote also notes that Muse quickly outpaced xAI’s Grok Bot on the app stores. Every major player now wants to be the AI that lives in your pocket and does things for you. Reuters framed the launch as OpenAI taking on Meta with an always-on agent and an enterprise push. That push is the differentiator: while Muse aims at the mass market, OpenAI is leaning hard on workplace integration through Slack, Teams and Microsoft.
OpenAI Dots Safety Concerns and Recent Incidents
Now for the less shiny part. The launch didn’t happen in a calm news cycle.
According to NBC News, questions of trust and safety hovered over the event. The evening before the launch, OpenAI apologized for its agents hacking an Australian government website. Axios reported that dots are launching publicly despite a run of revelations about AI agents from OpenAI and other labs taking unintended actions, and it noted a July incident in which OpenAI’s agents were behind an attack on Hugging Face. Anti-AI protesters also gathered outside the conference, per NBC.
Altman, for his part, acknowledged that a legitimate loss-of-control accident is possible if safety and security aren’t taken seriously enough. Axios reports he described the goal as keeping safety, alignment and monitoring ahead of capability, rather than plowing ahead blindly or halting progress altogether. NBC also reported that OpenAI held back its most advanced system for safety reasons.
Independent security commentators are watching the details. One analysis flagged that OpenAI itself identifies malicious instructions hidden in webpages, emails and documents as a prompt-injection risk. The takeaway: an agent that reads your inbox can be tricked by what’s in your inbox. That analysis was clear that layered safeguards don’t prove safety for every workflow.
So the honest summary is this. The safeguards are real and thoughtfully designed, and the track record around them is messy. Both things are true at once.
Are OpenAI Dots Worth It?
Here’s my take, split into the good, the cautious and the wait-and-see.
The exciting part: Dots point to a genuine change in how we use AI. Instead of prompting, waiting and copy-pasting, you delegate and review. Reviewing is much less exhausting than doing. If dots deliver even half of what the demos promise, the gains for developers, researchers, marketers and anyone drowning in admin could be huge.
The cautious part: Delegation only works if you can trust what’s delegated. Start small. Give your dot read-only access first, use Custom Rules to require approval for anything that sends, spends or shares, and check the Activity View regularly. Never let “it worked last time” replace review on high-stakes tasks like payments, contracts or anything customer-facing.
The wait-and-see part: We’re at day one. Availability is limited, pricing beyond the first month is unknown, and the promised “teams of dots” future hasn’t arrived. Real-world performance will look different from a polished launch video.
The Bottom Line
OpenAI’s Dots are a bold bet that the next phase of AI isn’t a smarter chat window. It’s a colleague that never clocks out. They’re wrapped in friendly cartoon packaging, backed by a serious model, and launched into a market where Meta and well-funded startups are sprinting just as hard.
Plugging an always-on agent into your Slack, your inbox, and your internal systems isn’t a switch you flip. It takes decisions about permissions, approval rules, integrations, and which workflows are safe to automate first. That’s why many companies won’t go it alone.
Cubix is one example of a development partner that builds custom AI solutions, including generative AI, AI chatbots, virtual assistants, and intelligent automation, and its site says it helps businesses integrate AI with existing systems. Whether you’re piloting dots or building agent-style tools of your own, a team that understands both the technology and your workflows can help you get the benefits without handing over the keys too fast
So here’s our question for you: what’s the first task you’d trust a dot with? And what’s the one thing you’d never let it touch?
Frequently Asked Questions
1. What are OpenAI Dots?
OpenAI Dots are personal AI agents, powered by GPT-6 Astra, that work continuously in the background on tasks and goals you set. Each dot has its own cloud computer, connects to over 4,000 apps, and learns your preferences over time. OpenAI announced them on September 29, 2026.
2. How much do OpenAI Dots cost?
Your first dot is included in your ChatGPT Pro or Business Premium plan at no extra cost. Dot usage doesn’t count toward plan allowances for the first month, and OpenAI says it will share ongoing usage terms afterward. Coverage of the launch puts the Pro plan at $100 to $500 a month. OpenAI plans to let users add more dots and scale each one’s speed or workload later.
3. Are OpenAI Dots available in the UK?
Not yet for Pro users. Pro availability excludes the UK, the European Economic Area, and Switzerland. Business Premium users can access dots in all supported ChatGPT regions. Sam Altman attributed the gap to strict regional regulation, while saying OpenAI wants to ship to Europeans.
4. Who can use OpenAI Dots right now?
Dots are rolling out to ChatGPT Pro and Business Premium users in eligible markets. Enterprise users, including Edu and Healthcare, can try a beta once their workspace admin enables it. OpenAI says it plans to expand access to more users soon.
5. How do I get started with OpenAI Dots?
Create your first dot in the ChatGPT desktop app or your desktop browser, connect your apps, and let it introduce itself. You can give it a name and make it your own. After setup, you can message it in the ChatGPT mobile app, and also through Slack and Teams. Texting support is coming soon.
6. What can OpenAI Dots do?
Dots can work on their own using their cloud computer, browser, and your connected apps. OpenAI’s examples include turning customer feedback into tested code fixes, updating sales proposals as requirements change, rerunning scientific analyses when new data arrives, and drafting show notes and social posts from interview transcripts.
