CyberAgent appears in this wiki as an adopter of AI-driven development practice rather than as a vendor of tooling. All the accounts available here come from inside the company — posts on its own developers’ blog, by engineers and by the executive in charge of technology — so what they establish is how the organization describes itself, not an outside assessment of it.
The engineer’s post states three organization-level facts. The company is described as having a culture of promoting AI adoption irrespective of job function. Its annual investment in introducing development agents is given as ¥400 million. And internal AI usage is made visible across the company, concretely as a ranking of tool usage: that post’s author reports having placed second company-wide for Claude Code usage, while noting that the denominator was masked.
A third post, 約2,000人が使うClaude Codeと向き合う。, gives the scale of that adoption for one tool: about 2,000 people inside the company were using Claude Code in June 2026, by which point use had moved from individual engineers’ experiments to an organization-wide phase in which governance, cost management, security, education and knowledge sharing had become design concerns, not only productivity. The same post mentions initiatives spreading across the CyberAgent group to make AI utilization visible and to give non-engineers Claude Code training.
The company’s stated direction is set out by its executive officer in charge of technology in 2028年「開発プロセス完全自動化」を見据えた、エンジニア組織の構造改革: at an internal event in October 2025 it adopted the vision of aiming for full automation of its development process, equated with AI maturity Level 4, by 2028, which that post frames as engineers and AI agents working together to reinvent development rather than as automation to save effort.
Background
The engineer’s post is written from the partnered retail-media division (協業リテールメディア div.) of the company’s AI business unit (AI 事業本部), which builds and operates a reward-type advertising service for retailers’ own apps — the stated aim being to lead customers to a purchase once they understand the product.
That service is described as spanning ad delivery servers, in-app web views styled to each retailer’s brand, an SDK for integrating into an app, an admin console for submitting ads, and an analytics platform over delivery results and retail point-of-sale data. The author gives that spread as the reason the work demands backend, frontend, native and data-science skills alongside domain knowledge of retail, and names Sapporo Drug Store and Yamada Denki as retailers it has been deployed with.
That post also supplies the organizational condition its practice was a response to: the service launched with six engineers and three business staff, and at the time of writing ran with two engineers and one business staff member, without a corresponding reduction in the advertising campaigns and initiatives it had to deliver.
Activities
The executive’s post, 2028年「開発プロセス完全自動化」を見据えた、エンジニア組織の構造改革, also describes the organizational measures the company attaches to that vision. Its engineer evaluation system, the JB Career Program, is to be fully renewed for the first time since 2019, around four career ladders: a renewed tech-lead engineer track, a new business-lead engineer role spanning business problem discovery through implementation and operation, a strengthened specialist track with 25 defined focus areas, and engineering managers. It also names stronger training for junior engineers, an engineer-specific AI ranking that evaluates on engineering quality and technical maturity, and an AI Driven Promotion Office, set up in August 2025, to support across the company an autonomous development structure in which engineers and AI agents work together.
Pages in this wiki written from CyberAgent’s developers’ blog: