OpenAI's dots Transforms Coding Tools into Proactive AI Assistant

OpenAI launched "dots" on September 29, 2026, at its DevDay, transforming its Codex coding tools into an always-on AI personal assistant led by Embiricos, co-founder of Multi. This move signals OpenAI's strategic shift toward persistent, proactive AI agents rather than traditional chatbots.

From Codex to Proactive Assistant

The foundation of dots originates from OpenAI's Codex coding app, which has been reframed to serve as an always-on personal assistant. This evolution highlights OpenAI's vision for the future of consumer AI, moving beyond reactive chatbots towards agents capable of anticipating user needs and taking initiative.

A key feature of dots is its "proactive research" capability. This allows the assistant to periodically review user data and propose helpful actions without direct prompting. This functionality aims to integrate AI more deeply into daily workflows, offering assistance before a specific request is made.

Privacy Considerations and Granularity

OpenAI has implemented privacy controls for dots that are similar to those found in ChatGPT, including options for training opt-out and access management. However, these controls currently lack granularity. A notable limitation is dots' refusal of read-only access to cloud drives, which could impact its ability to perform certain proactive tasks while maintaining user control over data access.

The introduction of proactive features raises questions about data privacy and the extent of AI autonomy. While OpenAI has provided some control mechanisms, the ongoing development suggests that more refined options may be necessary as the agent's capabilities expand.

Enabling Transactions and Future Outlook

OpenAI is actively engaged in discussions with various payment processors to enable dots to conduct financial transactions on behalf of users. This capability would allow the AI assistant to manage purchases and other financial activities, further embedding it into the user's digital life. This potential feature underscores the company's ambition for dots to become a comprehensive personal agent.

Embiricos, the product lead for dots, has publicly acknowledged that he still relies on a human assistant and does not foresee a future where he would conduct meetings solely with other AI agents. This perspective indicates current limitations in agent-to-agent interaction and highlights the ongoing development required for AI agents to achieve full autonomy and sophisticated collaborative capabilities.

The Shift to Persistent AI Agents

OpenAI views the competitive landscape for consumer AI as moving towards persistent, proactive agents rather than remaining centered on traditional chatbots. This strategic direction positions dots as a central component in this evolving ecosystem. The company's focus on agents that can continuously learn and act independently contrasts with the session-based interactions typical of many existing AI tools.

While dots shares some conceptual similarities with other AI assistants like Meta's Muse and Instinct, its emphasis on proactive, always-on functionality and potential for transactional capabilities sets it apart. The ongoing development and integration of such features will likely shape the future of AI news and product updates in the coming years.

Conclusion

OpenAI's launch of dots represents a significant step towards a future dominated by proactive AI agents. While the platform offers advanced features like proactive research and potential transactional capabilities, questions regarding privacy granularity and agent-to-agent interaction remain areas for future development. As OpenAI continues to refine dots and explore its full potential, the industry will be watching how these persistent AI assistants reshape user interaction with technology.

Sources

• What Is OpenAI Dots? What It Is and How to Access It
• DevDay 2026 announcements and developer resources - Announcements - OpenAI Developer Community
• How OpenAI turned its coding tools into dots, an always-on personal assistant | TechCrunch