CyberAgent AI Integration: Leveraging ChatGPT Enterprise and Codex

CyberAgent, a Japanese internet company specializing in advertising, media, and gaming, has integrated ChatGPT Enterprise and Codex as foundational technologies to drive business growth and operational efficiency. By prioritizing security and a culture of voluntary adoption over mandates, the company has reached a monthly active user rate of 93% across nearly all departments.

Secure AI Adoption via ChatGPT Enterprise

ChatGPT Enterprise provides the security and governance framework necessary for CyberAgent to scale AI usage across its workforce without compromising sensitive data. Before adopting the enterprise version, the company faced challenges with uncertainty regarding data safety and inconsistent usage across departments.

ChatGPT Enterprise addressed these issues by providing:

• Enterprise-grade security and access control: Enabling employees to use AI confidently in daily tasks.
• Customizable data handling: Allowing the company to manage input data according to specific contracts, configurations, and operational policies.
• Internal guidelines: Complementing the tool with clear rules for handling confidential information.

Ken Takao, Manager in the Data Technology Department, notes that these features allowed the company to support the business use of a wide range of information, excluding confidential data, thereby expanding the scope of AI use across the organization.

Driving Voluntary Adoption and Cultural Integration

Rather than enforcing tool usage, CyberAgent fostered adoption through culture building and targeted support. The 93% monthly active user rate was achieved through several strategic mechanisms:

• Knowledge Sharing: Implementing systems to share successful prompts and use cases.
• Usage Visibility: Creating internal rankings (for personal view only) to help employees understand their own AI tool usage.
• Proactive Follow-ups: Using Slack bots to reach out to inactive users to understand roadblocks or suggest practical AI applications for their specific roles.
• OpenAI Training: Participating in over ten OpenAI-hosted workshops, ranging from "ChatGPT Enterprise 101" to custom GPTs and Codex hackathons, with over 100 employees attending each session.

Enhancing Technical Decision-Making with Codex

CyberAgent utilizes Codex not only for code generation but as a tool for "upstream" work—design, alignment, and evaluation—to reduce rework during implementation.

Key Use Cases for Codex

• Design Pressure-Testing: Reviewing and evaluating design proposals from multiple perspectives.
• Code Review Optimization: Generating improvement suggestions and selecting from multiple options during reviews.
• ** Codex-driven Documentation**: Building and maintaining knowledge documents (e.g., AGENTS.md) to provide agents with richer context.

Quantifiable Benefits

According to Ken Takao, the use of Codex has resulted in better design quality, faster alignment before implementation, and clearer rationale behind proposals, which accelerates decision-making.

Impact on Development and Product Velocity

Codex is being used across both engineering and non-developer roles to write specifications, create mockups, and structure product-adjacent work. It has also been used to build internal systems, such as the company's usage ranking system.

In specific product divisions, the impact is evident:

• AI Business Division: Sou Yoshihara uses Codex via MCP in Cursor for the design and implementation planning of Kiwami Prediction AI, noting that Codex produces higher quality proposals than other coding models.
• Game Development (GOODROID): Hidekazu Hora used Codex to develop the game WormEscape, which reached soft launch in approximately one month. Hora describes Codex as a "reliable partner" that helps resolve knowledge gaps and increases both development speed and quality.

Long-term Strategic Vision

CyberAgent views AI as a fundamental turning point for the internet industry rather than a temporary trend. The company continues to leverage its AI Lab, established in 2016, and its AI Operations Office, launched in 2023, to ensure AI remains a foundational technology for both long-term technical advancement and practical business application.