OpenAI has launched Dots, a new class of always-on AI assistants designed to work in the background across connected apps, follow user preferences and take on tasks that once required a person to click through multiple tools. The company announced the product during its DevDay 2026 keynote, positioning Dots as a direct answer to Meta’s Muse, the customizable AI assistant that arrived only weeks earlier and quickly became one of the most talked-about consumer AI products of the year.

Unlike a conventional chatbot that waits for a prompt and returns a single response, a Dot is built to operate continuously. OpenAI describes the assistants as able to do nearly anything across supported services, using their own cloud computer to open a web browser, interact with more than 4,000 apps and carry context from one session to another. The pitch is simple: instead of asking an AI a question, users can assign it a goal and let it work while they do something else.

Key facts

• OpenAI announced Dots during its DevDay 2026 keynote as always-on AI assistants.
• Dots are powered by OpenAI’s GPT-6 Astra model.
• Each Dot uses its own cloud computer and can access a web browser plus more than 4,000 supported apps.
• Users can interact through a text-message-like interface or a voice call from ChatGPT on web, desktop or mobile.
• Dots can connect to Microsoft Teams and Slack and carry context across apps and devices.
• OpenAI plans to add text-message communication with Dots soon.
• Users can create one Dot at launch; multiple agents are planned.
• Dots message users with updates and questions while completing tasks.
• Users can watch progress through the Dot’s cloud computer.
• Dots learn from their work and can suggest additional tasks before the user asks.
• OpenAI is testing specialist Dots for organizations.
• Dots are rolling out to ChatGPT Pro, Business Premium and Enterprise users.
• Conversations with a Dot do not count toward ChatGPT usage limits.
• Dots include built-in rules, custom permissions and an auto-review safety feature.

A direct response to Meta’s Muse

The timing of the launch is difficult to ignore. Meta’s Muse has become a cultural and commercial phenomenon, giving mainstream users a friendly, customizable AI companion that can browse the web, make purchases, generate documents and track goals. Muse’s cute avatars and approachable interface helped make agentic AI feel less like enterprise software and more like a consumer product. OpenAI’s Dots borrow some of that same energy: each Dot is represented by a customizable avatar, and the primary interaction is a text-message-style thread that feels closer to chatting with a helpful contact than configuring a workflow automation.

But OpenAI is not simply copying the look and feel of Muse. Dots are being presented as more persistent and more deeply integrated. A Dot can connect to workplace tools such as Microsoft Teams and Slack, where it can carry context from sessions in other apps and on other devices. That means a user could start a task on a phone, continue it on a desktop and let the Dot coordinate with colleagues in a shared workspace. OpenAI is also introducing ChatGPT Space, a collaborative workspace where users and colleagues can work with a Dot together.

How Dots work across apps and devices

The core of the product is the cloud computer. Each Dot runs in its own environment, which gives it a browser and access to supported apps without taking over the user’s device. This is important for both capability and safety. A cloud computer allows a Dot to keep working when a laptop is closed or a phone is put away, and it gives users a window into what the assistant is doing. OpenAI says users can follow along with a Dot’s progress by accessing that cloud computer, turning the agent from a black box into something closer to a shared desktop.

Interaction is designed to be flexible. Users can send messages in a text-message-like interface, similar to the experience offered by Muse, or they can hop on a voice call with a Dot from ChatGPT on the web, desktop or mobile. OpenAI says Dots will soon be able to communicate over text message as well. The combination of text, voice and background work suggests that OpenAI wants Dots to feel present across a user’s day rather than confined to a single app window.

Task delegation, updates and learning

Once a user gives a Dot a task, the assistant does not disappear silently. It messages the user with updates and questions while it works. That back-and-forth is meant to keep the human in the loop without forcing them to supervise every step. OpenAI offered two examples that show the range of intended use cases. A developer working on an app could ask a Dot to use customer feedback to build and test updates, then present a video showing the changes. A content creator could have a Dot review an interview transcript, find moments to clip, create show notes and write social posts.

Dots also learn while they work. According to OpenAI, that learning allows them to personalize their output and come up with other tasks to complete before the user even thinks to ask. The idea is that a Dot becomes more useful over time as it accumulates context about a person’s preferences, projects and working style. For now, users can create only one Dot, but OpenAI plans to let users deploy multiple agents at once. The company is also testing a feature that lets organizations create specialist Dots for specific roles inside the company.

Safety rules and auto-review

The rise of always-on agents raises obvious questions about control. An assistant that can browse the web, connect to thousands of apps and act in the background needs clear boundaries, especially when it is handling workplace data, purchases or personal information. OpenAI says Dots come with built-in rules for when to act independently, and users can set custom rules that block certain actions or dictate when a Dot should ask for permission.

Dots also include an auto-review feature. OpenAI says the feature verifies whether an action complies with custom rules and safety requirements. It can also help a Dot determine which tasks should be handed off to the user, such as changing a password. Those safeguards will be closely watched. Meta’s Muse has already faced concerns about safety, including a report from one user who said the assistant handed out their address to a buyer on Facebook Marketplace without their knowledge. The incident does not prove that all agentic assistants are unsafe, but it illustrates the stakes as AI products move from answering questions to taking actions.

Pricing, availability and future scaling

Starting today, OpenAI is rolling out Dots to ChatGPT Pro, Business Premium and Enterprise users. The company says conversations with a Dot will not count toward ChatGPT usage limits. That detail could make Dots more attractive to heavy users who already rely on ChatGPT for work and do not want background agent activity to eat into their existing quotas. OpenAI also says that in the future users will be able to add more Dots and scale the output of each Dot by increasing its speed or the total amount of work it can take on per month.

The product arrives at a moment when AI companies are racing to turn chatbots into agents. The first wave of generative AI was defined by conversation: users typed a prompt and received text, an image or a piece of code. The next wave is defined by delegation. Agents are being designed to plan, browse, use software, coordinate with other tools and complete multi-step tasks with limited supervision. OpenAI’s Dots, Meta’s Muse and similar products from other technology companies are all chasing the same promise: an assistant that can do real work in the background rather than wait for the next prompt.

Competitive context and open questions

For OpenAI, Dots also serve a strategic purpose. ChatGPT is already one of the most recognizable AI products in the world, but the market is moving quickly, and competitors are experimenting with persistent, personalized assistants that feel less like tools and more like companions. By bundling Dots into ChatGPT for paid users and connecting them to workplace apps such as Teams and Slack, OpenAI can position its assistants as both consumer-friendly and enterprise-ready.

The biggest questions are execution and trust. Always-on agents must be reliable enough to leave alone, transparent enough for users to understand what they are doing and constrained enough to prevent costly mistakes. The cloud computer helps with transparency, and auto-review is meant to enforce boundaries, but users will ultimately judge Dots on whether they save time without creating new risks. The competition with Muse will play out not only on features but on confidence: whether people are willing to hand over real tasks to an AI assistant that keeps working after they close the tab.

Source: The Verge News