What OpenAI’s new always-on agents can do, who can use them, and the risks you should know about

OpenAI Dots are here, and they mark a big shift in how ChatGPT works. Announced by CEO Sam Altman at OpenAI’s annual DevDay in San Francisco on September 29, 2026, Dots are always-on AI agents that keep working in the background, across thousands of apps, even when you’re not chatting with them. Instead of answering questions, they’re designed to get work done on your behalf. Here’s what they are, what they can do, who can use them today, and why they’re already raising questions.

Quick Answer

OpenAI Dots are always-on AI agents powered by OpenAI’s GPT-6 Astra model. Each Dot runs on its own cloud computer, can connect to more than 4,000 apps, learns from your feedback, and can be reached through ChatGPT, Slack, or Microsoft Teams. They’re rolling out now to ChatGPT Pro, Business Premium, and Enterprise users in eligible markets. At launch, each user gets one Dot.

What Are OpenAI Dots?

Until now, using ChatGPT meant typing a question and waiting for an answer. Dots work differently. OpenAI describes them as extensions of the user rather than chat assistants: you give your Dot a goal, and it keeps pursuing that goal in the background with minimal supervision.

Each Dot has:

• Its own cloud computer and web browser, so it can carry out tasks without using your device
• Connections to 4,000+ apps through OpenAI’s plugin ecosystem
• Memory across channels, so context from a Slack message can carry over into ChatGPT
• The ability to learn from feedback and get better at your tasks over time

You start with a single “primary” Dot, give it a name, and customise it. OpenAI says you’ll be able to add more Dots, and buy extra speed or capacity, later on. Its longer-term vision is “teams of Dots” working together.

What Can Dots Actually Do?

OpenAI’s examples show agents that notice things and act on them, usually asking for your approval before anything important happens:

• Spotting work that needs doing: In one example, an agent noticed a client needed billing, prepared the invoice, and sent it once the user approved.
• Reacting to alerts: A Dot can see a software bug report arrive in Slack and start investigating it.
• Keeping documents up to date: For instance, updating launch materials when a product’s scope changes.

You can message your Dot through ChatGPT on web, mobile, and desktop, or through Slack and Microsoft Teams. Text message support is coming soon. Developers can also launch Dots from Codex, OpenAI’s coding tool.

OpenAI Dots at a Glance

Who Can Use OpenAI Dots, and Is It Available in India?

OpenAI Dots are rolling out first to paying ChatGPT customers on the Pro, Business Premium, and Enterprise plans. Free and Plus users don’t get access yet.

There are regional limits. According to TechCrunch and other launch coverage, Business Premium users can use Dots in all supported ChatGPT regions, while Pro access excludes the European Economic Area, Switzerland, and the UK. India isn’t on that exclusion list, so Indian Pro and Business Premium subscribers should be able to try Dots, though OpenAI has said it will share detailed usage terms for each plan separately.

Specialist Dots for Businesses

For companies, OpenAI also previewed “specialist Dots”: agents given their own identity, login credentials, and system access to handle one defined job, such as research or operations. OpenAI is starting with focused enterprise pilots and is working with Microsoft to bring these agents under the security and governance controls of Microsoft Agent 365, so IT teams can manage them with tools they already use.

The Concerns: Safety, Privacy, and Control

Always-on agents that can act across thousands of apps raise obvious questions, and the launch hasn’t been entirely smooth:

• Demo hiccups: Reuters reported that live demonstrations at DevDay hit snags, with agents repeatedly failing to deliver voice updates when asked.
• A rough stretch for OpenAI’s agents: The launch comes while OpenAI is still investigating earlier incidents involving its agents, including a breach at Hugging Face in July.
• Mistakes are possible: OpenAI itself warns that Dots can still make mistakes and that users should always review consequential work.

OpenAI says Dots run on GPT-6 Astra, which it calls its most aligned model, with safety and privacy protections switched on by default, and users can set rules for what needs their approval. Even so, it’s wise to start with low-risk tasks and keep approvals on for anything involving money, customers, or sensitive data.

Why This Launch Matters

OpenAI made chatbots mainstream, but it was late to always-on agents. Meta’s Muse agent launched about three weeks earlier and quickly topped the app store charts. OpenAI Dots are the company’s answer, and they signal where the whole industry is heading: from AI that answers questions to AI that does work on your behalf, around the clock.

Also read: ChatGPT Hits 1 Billion Users

Frequently Asked Questions

What are OpenAI Dots?
OpenAI Dots are always-on AI agents in ChatGPT, powered by GPT-6 Astra. Each one runs on its own cloud computer, connects to more than 4,000 apps, and works on your goals in the background.

Are OpenAI Dots free?
No. At launch, Dots are available only on ChatGPT’s Pro, Business Premium, and Enterprise plans.

Can I use Dots in India?
India isn’t among the excluded regions for Pro, and Business Premium covers all supported ChatGPT regions, so Indian subscribers on those plans should have access as the rollout continues.

How are Dots different from regular ChatGPT?
Regular ChatGPT responds when you ask it something. Dots keep working in the background, take actions across connected apps, and can be reached through Slack and Microsoft Teams as well as ChatGPT.

Are OpenAI Dots safe to use?
OpenAI enables safety and privacy protections by default and lets you set approval rules, but it also warns that Dots can make mistakes. Review any important work before it goes out.

Conclusion

OpenAI Dots are the company’s boldest step yet beyond the chatbot: personal agents that work around the clock, across thousands of apps, on your behalf. For now they’re limited to premium plans, one Dot per person, and come with real questions about safety and control. But if they deliver on OpenAI’s promise, they could change how professionals handle routine work.

My take: I’d be really interested in trying OpenAI Dots, but only for regular, low-risk tasks that wouldn’t cause any harm if something went wrong. For example, drafting invoices or meeting notes for me to quickly review and approve, or monitoring incoming bug reports. These are the repetitive background jobs where automation could genuinely save time. In short, I’d start with very small tasks, and only once the agent had earned my trust would I expand its responsibilities.

Over to You

Would you let an AI agent like Dots work on your behalf around the clock, or is that a step too far? Share your thoughts in the comments below.