OpenAI Dots Officially Launches: GPT-6 Astra-Powered Always-On Agents Connect 4,000+ Apps with Enterprise Governance Framework
From Chatbots to Always-On Digital Colleagues
On September 29, 2026, OpenAI officially launched Dots at its annual developer conference (DevDay), with a gradual rollout to users beginning in October. This is not just a new feature but a fundamental shift in the AI assistant paradigm: evolving from "question-and-answer" single-turn conversations to "always-on" agents capable of continuously running and proactively completing tasks.
The core concept of Dots is positioning AI agents as users' "resident colleagues" — not tools waiting for instructions, but digital partners capable of proactive monitoring, autonomous execution, and continuous learning.
Technical Architecture: GPT-6 Astra's Agentic Capabilities
Model Foundation
Dots is powered by OpenAI's flagship model GPT-6 Astra, the latest frontier model launched at the same DevDay. GPT-6 Astra is designed for complex agentic tasks, featuring:
- Long-term memory and context retention
- Multi-step task planning and execution
- Tool calling and external system integration
- Complex reasoning and decision-making capabilities
Dedicated Cloud Computing Environment
Each Dot has its own cloud computer, which is the technical foundation of its "always-on" capability. Even when the user is offline, the Dot can continue running in the background, monitoring new inputs, responding to triggers, and executing scheduled tasks.
4,000+ App Ecosystem
Dots connects to over 4,000 applications through a plugin ecosystem, covering:
- Project management tools (Jira, Asana, Linear)
- Communication platforms (Slack, Microsoft Teams, Gmail)
- Development tools (GitHub, GitLab, Cursor)
- Business intelligence (Salesforce, HubSpot, Notion)
- And hundreds of industry-specific applications
Core Features: End-to-End Automation from Monitoring to Execution
Persistent Goal Execution
Unlike traditional ChatGPT sessions that end after answering a question, Dots maintains a "standing goal" and can:
- Monitor: Continuously watch for new data, events, or time-based triggers
- Analyze: Intelligently analyze and categorize new inputs
- Execute: Autonomously complete predefined workflows
- Report: Inform users of execution results and anomalies
Typical Use Cases
Product Development:
- Monitor customer support tickets, automatically categorize and generate GitHub Issues
- Track competitor dynamics, regularly generate competitive analysis reports
- Monitor CI/CD pipelines, automatically handle common build failures
Marketing:
- Monitor social media mentions, automatically generate response drafts
- Track advertising performance, automatically adjust budget allocation recommendations
- Regularly generate content calendars and publishing plans
Research and Analysis:
- Continuously monitor academic databases, push relevant new papers
- Automatically organize and summarize daily news, generate industry briefings
- Track regulatory developments, automatically update compliance documents
Enterprise Governance Framework: Safe Autonomy
Three-Level Operation Classification
OpenAI is well aware of the security challenges posed by "always-on" agents and has designed a sophisticated governance framework:
| Operation Category | Applicable Scenarios | System Behavior |
|---|---|---|
| Allow | Reversible operations (drafts, local files) | Auto-execute, no confirmation needed |
| Require Approval | Shared or public-facing operations (sending emails) | Pause and request user confirmation |
| Block | Sensitive operations (billing, deployment) | Refuse execution, log attempt |
Enterprise Administrator Controls
For enterprise and education workspaces, the Dots feature is off by default and requires administrator activation. Administrators can:
- Set organization-level operation whitelists and blacklists
- Monitor activity logs for all Dots
- Set different permission policies for different departments
Availability and Pricing
Subscription Plans
| Plan | Dots Quota | Additional Cost |
|---|---|---|
| Pro | 1 Dot | None |
| Business Premium | 1 Dot | None |
| Enterprise/Education/Healthcare | Beta (off by default) | Requires admin activation |
Usage Limits
- Conversations with a Dot do not count toward ChatGPT usage limits
- Tasks performed by the Dot via Codex or ChatGPT Work count toward respective allowances
- API developers can use the "Agents API with computer use" to replicate the persistent agent pattern
Access Channels
Dots can be accessed through:
- ChatGPT (web, desktop, mobile)
- Slack integration
- Microsoft Teams integration
- iMessage/RCS (US Pro users limited beta)
Asia-Pacific Adoption Prospects
Enterprise Automation Demand
Enterprises in the Asia-Pacific region, especially large corporations in Japan, South Korea, Singapore, and Australia, are facing serious labor shortage issues. Always-on agents like Dots can significantly increase workflow automation without increasing labor costs.
Language Support Challenges
Currently, the best Dots experience remains primarily in English. For non-English markets in the Asia-Pacific region (such as Japanese, Korean, and Chinese), the completeness of language support will be a key factor affecting adoption rates.
Data Privacy Considerations
Data privacy requirements vary across Asia-Pacific countries. Enterprises deploying Dots need to carefully evaluate whether its data processing policies comply with local regulations, especially in highly regulated industries like finance and healthcare.
Industry Impact: A New Standard for Agentic AI Tools
The launch of OpenAI Dots, alongside the simultaneous release of "always-on" agents like Meta Muse Wonder and Anthropic Claude agents, marks the AI tools industry entering a new competitive dimension.
The core of this competition is no longer "which model is smarter" but "which agent ecosystem is more complete, more secure, and easier to use." 4,000+ app integrations, sophisticated governance frameworks, cross-platform access — OpenAI Dots demonstrates strong competitiveness across all three dimensions.
Conclusion: The Ultimate Form of AI Assistants?
OpenAI Dots represents an important milestone in the development of AI assistants: from passive response to proactive execution, from single conversations to continuous operation, from tools to colleagues.
For enterprises and individual users in the Asia-Pacific region, this transformation brings both enormous efficiency improvement opportunities and new governance and security challenges. How to fully leverage agentic AI capabilities while establishing appropriate oversight and control mechanisms will be a core issue in enterprise AI strategy in the coming years.
