Enabling Long-Running Autonomous Workflows

OpenAI is acquiring Ona to evolve its Codex ecosystem from a session-based tool into a platform for persistent, long-running agentic work. As Codex usage has grown by 400% this year, OpenAI identified a critical need to move beyond tasks that require a user to remain actively connected to a machine. By integrating Ona’s technology, OpenAI aims to enable agents to operate autonomously over hours or days, allowing users to delegate complex, multi-step projects that continue even when local devices are offline.

Secure, Customer-Controlled Execution

For enterprises, the transition from experimental AI to production-grade deployment requires more than just model intelligence; it demands strict security, governance, and operational control. Ona’s technology provides a secure, reproducible cloud environment that allows agents to function within an organization’s own infrastructure. This architecture ensures that:

• Data Sovereignty: Agents operate within the customer's cloud boundaries.
• Granular Control: Organizations retain authority over credential scoping, tool access, and activity logging.
• Workflow Integration: Work can be managed through established review and approval processes, making it suitable for high-stakes software lifecycle tasks such as vulnerability remediation, application modernization, and automated testing.

By combining OpenAI’s orchestration capabilities with Ona’s execution model, the company intends to provide a trusted workspace that meets the rigorous security standards required for enterprise-scale AI deployment.