Source: Fujitzu
Platform will be used in modifications of all 67 software packages provided to medical and governmental industry customers by end of fiscal year 2026
Fujitsu Limited. Sydney, Australia, 26 February 2026 – Fujitsu Limited today announced the development and launch of its AI-Driven Software Development Platform, a new initiative to bring software development into the AI age and contribute to the sustainable growth of its customers and society.
This platform automates the entire software development process, from requirements definition and design to implementation and integration testing.
By leveraging the Takane large language model (LLM) [1] and agentic AI technology for large-scale software development developed by Fujitsu Research, the AI-Driven Software Development Platform enables AI agents to understand complex, evolving large-scale systems owned by enterprises and public organisations.
The platform has multiple AI agents collaboratively execute each stage of software development, achieving full automation of the entire process without human intervention.
Fujitsu aims to use this AI-Driven Software Development Platform to carry out revisions to all 67 types of medical and government business software products provided by Fujitsu Japan Limited by the end of fiscal year 2026.
The revisions are necessary due to legal and regulatory changes. From January 2026, the platform has been used in Japan for software modifications made necessary by the 2026 medical fee revisions [2].
In a PoC that updated software as per the 2024 medical fee revisions, the platform demonstrated a significant reduction in development time for one of approximately 300 change requests.
Using conventional software development methods [3], the modifications would have taken three person-months. With this technology that was dramatically shortened to four hours, achieving a 100-fold increase in productivity.
By utilising this AI-driven development platform, Fujitsu will dramatically improve the speed of software modifications necessitated by legal amendments and system changes.
This will significantly reduce the burden previously required for system verification during modifications, thereby freeing up time for the planning and development of measures and services that lead to improved patient, resident, and customer services, as well as enabling customers and partners to continuously adapt to ever-changing operations and societal needs.
Furthermore, Fujitsu envisions that this technology will also help customers and partners to respond to expanding IT demand and alleviate the worsening shortage of IT professionals. In AI-driven development, Fujitsu positions AI-Ready Engineering—the process of preparing assets and knowledge to ensure AI correctly understands existing systems and achieves highly reliable automation – as crucial.
With AI-Ready Engineering and the AI-Driven Software Development Platform working in tandem, Fujitsu will accelerate AI-driven software development. Fujitsu will promote a transformation in engineers’ work styles, strengthening its Forward Deployed Engineer (FDE) complement, and shifting the paradigm of software development from a conventional person-month-based approach to a customer value-based approach.
Moving forward, Fujitsu plans to expand the application of the AI-Driven Software Development Platform to a wide range of sectors, including finance, manufacturing, retail, and public services, by the end of fiscal year 2026.
Fujitsu will also begin offering this service to customers and partner companies to enable them to rapidly and flexibly develop systems that adapt to changes in their business environments.
Through these efforts, Fujitsu aims to transform the software development process into an AI-driven model as an industry standard.
Notes
[1] Takane large language model (LLM): Jointly developed by Fujitsu and Cohere Inc.
[2] Medical fee revision: A national system that reviews public medical fees and adjusts cost allocation for medical procedures.
[3] Conventional software development methods: Development methods where quality is verified at each stage, from software requirements definition, design, and implementation to integration testing.
