OpenAI Dots Always-On AI Agents Launch: GPT-6 Astra-Powered, Connects 4,000+ Apps, Enterprise 'AI Colleague' Era Begins

Introduction: From Chatbots to Always-On AI Colleagues

On September 29, 2026, OpenAI unveiled a product at its annual DevDay conference that promises to fundamentally reshape how enterprises work: Dots. These are not another iteration of chatbots, but an entirely new category of AI agent — always-on, equipped with their own cloud computers, and capable of continuing complex tasks even after users disconnect.

Powered by OpenAI's latest flagship model, GPT-6 Astra, Dots connect to over 4,000 applications and support major enterprise platforms including Slack, Microsoft Teams, and email. This launch marks a critical inflection point in the evolution of AI agents from "tools" to "colleagues," signaling a profound transformation in how enterprise work gets done.

Core Capabilities: The Always-On AI Colleague

Persistence and Autonomy

The most revolutionary aspect of Dots is their persistence. Each Dot has a dedicated cloud computer, meaning it can continue working after users close their browsers or applications. While traditional AI assistants "disappear" when a conversation ends, Dots function like genuine colleagues, continuously advancing tasks in the background.

Specifically, Dots offer these core capabilities:

• Persistent memory: Remembers user preferences, work habits, and project context across sessions
• Proactive research: Actively gathers information within user-authorized boundaries (read-only mode)
• Multi-step execution: Automatically completes complex workflows requiring multiple steps
• Cross-platform monitoring: Continuously monitors Slack channels, email inboxes, and other platforms, proactively notifying users when important information appears

Broad Application Integration

Through a plugin system, Dots connect to over 4,000 applications spanning:

Learning and Personalization

Dots learn how users work over time. By analyzing user feedback and preferences, each Dot gradually becomes more attuned to a specific user's thinking patterns and work habits. This personalized learning enables Dots to provide increasingly precise assistance rather than generic, one-size-fits-all responses.

Technical Architecture: The Power of GPT-6 Astra

Model Capabilities

GPT-6 Astra, the engine behind Dots, is OpenAI's most powerful model to date, featuring:

• Extended context understanding: Processes and retains large volumes of historical conversations and task context
• Multimodal capabilities: Simultaneously handles text, images, documents, and other formats
• Advanced reasoning: Demonstrates exceptional planning ability in complex, multi-step tasks
• Low-latency responses: Responds quickly to user queries even while executing complex background tasks

Companion Tool: GPT-6.1 Sol Ultrafast

Launched alongside Dots, GPT-6.1 Sol Ultrafast is a service tier optimized for high-speed responses. Priced at $12 per million input tokens and $60 per million output tokens via API, it delivers speeds up to 8x faster than the standard version, designed specifically for agent applications requiring real-time interaction.

ChatGPT Space: A New Dimension of Collaboration

OpenAI simultaneously launched ChatGPT Space, a collaborative environment where colleagues and their respective Dots can share context, knowledge, and documents. This enables multiple Dots to work together on the same project, forming genuine "AI teams."

Safety and Governance: Defining the Boundaries of Autonomous Agents

Multi-Layer Safety Mechanisms

As agent autonomy increases, safety becomes paramount. OpenAI designed multiple safety layers for Dots:

User-defined rules: Users can set explicit behavioral boundaries for each Dot, requiring approval before executing sensitive operations such as sending emails, changing settings, or publishing content publicly.

Auto-review mechanism: A secondary agent reviews each action against user instructions before execution, catching potential errors or policy violations.

Human-in-the-loop checkpoints: For high-stakes operations like password changes or financial transfers, the system requires users to take direct control, ensuring critical decisions always remain with humans.

Known Limitations

OpenAI's internal testing shows Dots are highly resistant to prompt injection attacks, but exhibit a moderate tendency toward "scope drift" as tasks become more complex or chain-dependent — meaning agents may deviate from the original intent of instructions. This finding is driving ongoing improvements to agents' boundary awareness.

Market Positioning and Business Model

Subscription Plans

Dots are currently available to:

• ChatGPT Pro subscribers ($100+/month)
• Business Premium plan subscribers

Notably, due to regulatory requirements, Dots are currently unavailable on Pro plans in the European Economic Area, Switzerland, and the UK, though Business Premium workspaces in those regions retain access.

Usage Limits

Standard conversations are unmetered. For background tasks, OpenAI provides a one-month grace period (expiring late October 2026), with long-term usage terms still being finalized.

Industry Impact: 86% of Large Enterprises Already Exploring Human-Agent Collaboration

According to the latest survey data, 86% of enterprises with over 500 employees are already operating systems involving human-agent collaboration. Dots' launch perfectly aligns with this trend, providing enterprises with a standardized, secure, and controllable always-on agent solution.

Competitive Landscape

Dots' launch intensifies competition in the AI agent market:

• Microsoft Autopilot: Positioned as an employee's "chief of staff" with tenant-level identity and memory
• Salesforce Hunter: A sales agent deeply integrated with CRM data
• HubSpot Breeze: A rebuilt CRM context-aware agent

Asia-Pacific Perspective: Opportunities and Challenges for Enterprise Adoption

For Asia-Pacific enterprises, Dots' launch presents unique opportunities. As APAC organizations continue increasing digital transformation investments, mature AI agent technology offers new pathways to operational efficiency.

However, APAC enterprises face specific challenges when adopting Dots:

1. Data sovereignty concerns: Some APAC countries have strict regulations on cross-border data transfers
2. Language support: Dots' support for Asian languages is still being refined
3. Regulatory compliance: AI regulatory frameworks vary significantly across countries, requiring careful compliance assessment

Conclusion: A New Era for AI Agents

OpenAI Dots' launch is not merely a product milestone — it represents a critical turning point in the history of AI agent development. From passive response to proactive execution, from single conversations to persistent collaboration, AI agents are integrating into enterprise workflows in unprecedented ways.

With the powerful capabilities of GPT-6 Astra and broad integration across 4,000+ applications, Dots is poised to become core infrastructure for enterprise digital transformation. For business decision-makers across the Asia-Pacific region, now is the critical moment to evaluate and plan AI agent strategies.

---

Sources: OpenAI DevDay 2026, Codersera, Wired, 9to5Google