ひとまいる has begun a large-scale modernization of its core systems. The project started without defining requirements, and this report follows the process of organizing and rebuilding its vast existing assets in less than 2 years, with AI utilization as the central focus.
カクヤス（ひとまいる） is undertaking a modernization of its core system, which had remained largely unchanged for many years. The current system comprises more than 2,000 screens and 1,200 stored procedures, while retaining the challenges of a client-server architecture. The company has launched a legacy modernization project estimated at 900 person-months.
One distinctive feature of the project is that it was launched without first finalizing the requirements. The project adopted an approach that combines the use of generative AI with work carried out manually in parallel. The estimated scale is 900 person-months, and the project is expected to be completed in less than 2 years.
The article discusses the specific role of AI and the importance of articulating requirements, and is positioned as a series exploring the practical essence of AI-driven development. Across the first through third installments, it examines how requirements should be defined and how the two workstreams can be advanced in parallel.
The series as a whole follows the practical application of AI and the challenges involved in modernizing カクヤス’s core systems. It focuses on the company, the project background, and specific numerical details on its progress.