IBM Japan Develops "ALSEA," a Standard Context Solution for Enterprise Specification-Driven Development

Making Specification-Driven Development the New Standard for Enterprise System Development in the AI Era

Now Available for Early Adopter Projects

IBM Japan, Ltd. (hereafter IBM Japan) has been developing "AI Lifecycle Shared Engineering Artifacts (ALSEA)," a standard context*1 solution for the full-scale application of specification-driven development in large-scale enterprise system development. Starting today, it will be available for early adopter projects.

◾️ Background of Development With the evolution of generative AI, its application is rapidly expanding in fields such as code generation and test automation. In development environments, the use of generative AI as a means to enhance individual work efficiency is becoming commonplace.

Since 2024, IBM Japan has been systematizing its "AI for IT" solutions to improve IT speed and quality by leveraging advanced technologies like AI. We have been supporting our clients' IT transformation by providing solutions that cover almost all development types and project phases, including code generation, test automation, advanced IT operations, and project management.

Furthermore, to enable the integrated use of these individual products and solutions and to complement areas that have not been adequately addressed, IBM Japan has launched "IBM Bob," an AI agent-driven development support partner for enterprises, usable by all engineers and non-engineers alike. This allows "AI for IT" to cover all directions, all situations, and all development types, shifting system development from being "human-led" to "AI-led," where humans focus on design and supervision.

Generally, large-scale enterprise system development presupposes an organizational structure, consensus-building with explainability, and standardized quality assurance. In mission-critical projects that span several years and involve hundreds of people, it is difficult to rely on individual skills or impromptu AI use. Since decision-making and the quality of deliverables directly impact business continuity, it is said that building a system that eliminates personal dependency is essential. Based on these enterprise-specific prerequisites, the full-scale application of AI requires a mechanism for the organization to control development decisions, instructions, and deliverable generation with high reproducibility.

◾️ ALSEA: A Standard Context Solution for Enterprise Specification-Driven Development IBM Japan has systematized the methods, know-how, standard processes, deliverable templates, rules, and guides for large-scale system development accumulated over many years into a "context" that IBM Bob can understand and utilize. We have developed ALSEA as a standard foundation for enterprise AI (specification)-driven development. By providing the standard context necessary for development as ALSEA, we enable AI to be correctly operated as an organization, allowing AI to take the lead comprehensively even in large-scale projects, with humans focusing on design and supervision.

ALSEA has the following features:

Standardizes Deliverable Quality: By providing IBM's unique large-scale development know-how (development guidelines, standards, rules, etc.) as a common context for IBM Bob to reference, it generates explainable and highly reproducible deliverables that do not depend on differences between developers or teams. This helps ensure the uniform quality required in enterprise environments.

Reduces Human Workload: By minimizing individual and ad-hoc instructions and standardizing on the minimum necessary deliverables, "AI (IBM Bob) takes the lead" in generating design documents, program code, and test deliverables. Humans can concentrate on review and judgment, reducing the burden and coordination costs in the development field by preventing rework and ensuring thorough reviews.

Applicable to Large-Scale Development Projects: Even in enterprise environments where multiple projects are running concurrently, it enables subsystem division and support for hybrid cloud environments, centered on standardized processes and project management.

Leveraging these features, IBM Japan helps resolve issues pointed out in the Ministry of Economy, Trade and Industry's DX Report's "2025 Digital Cliff," such as the black-boxing of legacy systems and human resource constraints.

◾️ Future Developments IBM Japan aims to establish specification-driven development using ALSEA, while leveraging IBM Bob, as the standard foundation for enterprise system development. It will be available for early adopter projects starting today, with general availability planned for the second half of 2026. To lead industry transformation with AI specification-driven development, we aim to achieve quantitative effects such as a 35% reduction in man-hours and a 30% reduction in project duration for the entire system development project from 2027 onwards. We will strongly support maximizing the strategic value of systems and achieving results directly linked to management.

End

*1: "Context" = Systematized and structured language that allows AI to understand the explicit knowledge accumulated by the systems department, the tacit knowledge of experts, and the work status. It includes: ① Prompts (user instructions), ② RAG (Retrieval-Augmented Generation), ③ System Prompts (instructions to control AI's autonomy behind the scenes), ④ Short-term memory (session information retention), ⑤ Long-term memory (cross-session information retention, utilization, and updating), ⑥ External tool integration, and ⑦ Structured output (AI-interpretable format).

IBM, the IBM logo, and ibm.com are trademarks or registered trademarks of International Business Machines Corporation in the United States and/or other countries. Other product and service names may be trademarks of IBM or other companies. A current list of IBM trademarks is available on the web at ibm.com/trademark.