CyberAgent Moves Faster with ChatGPT Enterprise and Codex

CyberAgent has made ChatGPT Enterprise the foundation of its in-house AI environment, naturally weaving AI into work across advertising, media, and games. It established AI Lab in 2016, and in 2023 launched the AI Operations Office, treating AI not as an experiment but as a core technology for organizational operations.

The starting point of adoption was security and governance. After the spread of generative AI, there was anxiety about what information could be entered, and usage levels varied by department, making management difficult. CyberAgent used ChatGPT Enterprise's account management, usage visibility, and enterprise-grade security features together with internal guidelines to build a safe usage environment, which as a result allowed AI to be confidently incorporated into everyday work.

Adoption was not mandatory, but the pace of spread was fast. Even though teams and subsidiaries each chose tools according to their own needs, the company recorded a 93% monthly active user rate. Internally, prompts and use cases were shared, individual usage rankings were provided, and for those who hadn't used the tools for a certain period, a Slack bot directly asked about their usage status and alternative tools, pushing learning and habit formation.

OpenAI's educational support also backed the spread.

• Even with a short notice period, more than 10 sessions were held, with over 100 employees participating in each session.
• Level-based programs were run, from the introductory ChatGPT Enterprise 101 to custom GPTs, hands-on Codex sessions, and internal hackathons.

Codex is establishing itself not merely as a code generation tool but as a means of improving quality at the upstream stage. It is used for reviewing design proposals, comparing improvement options in code review, and maintaining knowledge documents like AGENTS.md, helping teams make better judgments before implementation. Non-developer roles also use it for writing specifications, creating mockups, and structuring tasks, and there was even a case where the internal usage ranking system itself was built using Codex.

Feedback from the field is concrete. Sou Yoshihara of the AI Business Division uses Codex through Cursor's MCP, applying it to the design and implementation planning of Kiwami Prediction AI, while Hidekazu Hora of GOODROID said that in developing the game WormEscape, the flow from design to implementation and verification became smoother. CyberAgent sees AI not as a short-term trend but as the next standard for the internet industry, and is redesigning the way work itself is done, beyond simply adopting tools.