OpenAI employees are using the company's Codex agentic coding app to handle an increasing amount of work [1].
This trend demonstrates the practical application of AI coding agents in a corporate environment. By allowing both technical and non-technical staff to build their own tools, the company is effectively testing the productivity potential of its own technology in real-time.
According to reports, workers are using the tool to create simple dashboards and automate repetitive tasks. This approach allows employees to function as if they have their own chief of staff—a role typically reserved for senior executives—by offloading complex data management to the AI agent.
While the internal use of Codex is widely reported, some reports suggest a conflict in how the technology is being integrated. One report indicates that OpenAI staff are using the tool internally [1], while another suggests the company has developed a plugin to run its Codex agent inside Anthropic's Claude Code agent [2].
The ability for non-technical employees to generate code and build functional dashboards without deep programming knowledge is a central part of this internal shift. It allows the company to move faster by removing the bottleneck of relying solely on on-site software engineers for every small internal tool.
By treating the AI as a personal assistant for technical tasks, OpenAI is creating a blueprint for how future corporate workflows will operate. The shift from manual data entry and manual dashboard creation to agentic AI tools is changing the way the company's internal operations are managed.
“OpenAI employees are using the company's Codex agentic coding app to handle an increasing amount of work.”
The internal adoption of Codex at OpenAI highlights a shift toward 'agentic' AI, where tools do not just suggest text but actively execute tasks. This suggests that the boundary between technical and non-technical roles is blurring, as employees can now leverage AI to build their own software solutions without traditional coding skills, potentially redefining productivity metrics in the tech industry.
