OpenAI has introduced Dots, always-on artificial intelligence agents designed to continue working between conversations and handle recurring tasks for users.
The company announced the feature on September 29, 2026, during its DevDay conference in San Francisco.
Dots are powered by OpenAI’s GPT-6 Astra model and are being rolled out gradually to eligible Pro and Business Premium users aged 18 and older in supported markets.
Table of Contents
What OpenAI Dots can do
Unlike a conventional chatbot that waits for a prompt, a Dot can be assigned an ongoing goal and continue making progress after the user leaves the conversation.
OpenAI says each Dot has its own cloud computer and browser.
Users can connect selected applications and define what the agent is allowed to do independently.
The system is designed for work such as organizing information, following up on projects, reviewing files, monitoring recurring tasks and preparing materials for a user’s approval.
OpenAI’s official guidance also describes connections with services including Google Calendar, Google Drive, Docs, Sheets and GitHub, although availability and permissions can vary by account and market.
Dots can bring completed work back to the user for review rather than requiring the person to remain in the conversation while the task is running.
How the new agents differ from ChatGPT
ChatGPT traditionally responds to individual requests, while Dots are intended to work across multiple steps and return to a task over time.
That makes the feature closer to a digital work assistant than a standard question-and-answer tool.
The agent can use connected applications, retain relevant context and, in some cases, delegate parts of a job to subagents.
The approach reflects a broader shift in the artificial intelligence industry toward software agents that can operate tools and complete workflows instead of only generating text or images.
Meta has also moved into this area with its Muse AI assistant, putting OpenAI’s Dots into a growing competition over personal agents that can act on a user’s behalf.
Access and limitations
Dots are not being released universally at launch.
OpenAI says the rollout is gradual and currently applies to eligible Pro and Business Premium users in supported markets.
Enterprise access is available as a beta and is turned off by default, while some regions are excluded from the initial Pro rollout.
The company has also said that Dots usage will not count toward eligible users’ plan allowances during an initial one-month period.
OpenAI has not yet provided the longer-term usage terms for every plan.
Users cannot currently create a Dot from mobile devices, according to OpenAI’s support documentation.
Creation is available through the ChatGPT desktop application or desktop web experience.
The feature also depends on the permissions a user grants to connected services.
A Dot cannot safely or accurately manage an application without access to the relevant information and tools.
Safety questions follow the launch
Always-on agents introduce different risks from ordinary chatbots because they can interact with software, handle personal information and continue operating after the user has stopped watching.
OpenAI’s GPT-6 Astra safety documentation says Dots include safeguards designed to review actions and limit what an agent can do without additional user involvement.
OpenAI’s documentation indicates that the system is intended to keep users involved in decisions with higher consequences, rather than giving the agent unrestricted control over communications, financial activity or other sensitive actions.
That distinction will matter as users connect email, calendars, cloud files and workplace systems.
A mistaken summary is inconvenient, but an incorrect message, file change or business decision could have wider consequences.
The launch comes as researchers and technology companies continue to debate how much independence should be given to advanced AI systems.
Recent developments have increased attention on monitoring, permissions, model behaviour and the need for human review.
What Dots mean for Canadian users
For Canadians, the rollout is part of a wider expansion of AI tools into office work, education, software development and customer service.
Canada’s federal government launched its AI for All strategy in June 2026, with priorities that include public trust, AI adoption, worker training, sovereign infrastructure and stronger safeguards.
Tools such as Dots could make advanced automation more accessible to small businesses and individual professionals, particularly for repetitive administrative work.
At the same time, users will need to consider where information is stored, which companies can access connected data and whether an AI agent’s actions can be audited or reversed.
Businesses considering these tools may also need internal rules covering confidential records, customer information, employee data, approval processes and the use of AI-generated work.
What happens next
OpenAI is expected to expand Dots gradually as it gathers feedback from users and evaluates how reliably the agents perform longer-running tasks.
The company’s next major test will be whether Dots can move beyond demonstrations and handle useful, repeatable workflows without creating more oversight work than they save.
For now, the launch gives eligible users an early look at a more persistent form of ChatGPT—one that is designed to keep working after the conversation ends.
Frequently Asked Questions
What are OpenAI Dots?
Dots are always-on AI agents in ChatGPT that can continue working on assigned goals between conversations using connected applications and a cloud computer.
Who can use OpenAI Dots?
Dots are rolling out gradually to eligible Pro and Business Premium users aged 18 and older in supported markets. Enterprise access is available as a beta and is off by default.
What can a Dot do?
A Dot can manage recurring tasks, work with connected apps, organize information and return results for the user’s review.
Are OpenAI Dots available on mobile?
OpenAI’s support documentation says users currently create Dots through the ChatGPT desktop application or desktop web experience, rather than on mobile.
Do OpenAI Dots act without human oversight?
Dots can continue working independently within the permissions a user provides, but OpenAI says its safeguards are designed to keep users involved in higher-consequence actions and review.
Fact-Checked: Key launch, availability and safety details were checked against OpenAI’s official release notes, support documentation and GPT-6 Astra safety materials, along with Canada’s federal AI strategy information.
Disclaimer: Availability, features, pricing and regional access for OpenAI Dots may change as the rollout continues.