The digital world moves at an unrelenting pace, constantly demanding innovation and agility. Yet, many enterprises find themselves tethered to the past, relying on complex, decades-old legacy systems that, while functional, have become inhibitors rather than enablers of progress. The challenge of understanding, maintaining, and modernizing these monolithic structures is immense. But what if artificial intelligence, specifically generative AI, could offer a powerful new lens through which to view and transform these critical assets?
Today, March 30, 2026, marks a pivotal moment in this ongoing struggle. Fujitsu, a global leader in technology and services, has officially launched its groundbreaking generative AI service, 'Fujitsu Application Transform powered by Fujitsu Kozuchi.' [1, 2] This innovative offering is set to redefine legacy system modernization by automating the historically arduous process of source code analysis and design document generation, promising to dramatically cut down the time and resources traditionally required. [1, 2]
Legacy systems are the backbone of countless organizations across industries, from finance and healthcare to government and manufacturing. Built in an earlier era, often on now-obsolete programming languages like COBOL, these systems handle mission-critical operations. However, their age brings a host of formidable challenges:
- Mounting Technical Debt and Maintenance Costs: Maintaining legacy systems consumes a significant portion of IT budgets. A 2019 GAO report highlighted that the U.S. government spent $337 million maintaining just ten legacy systems, some over 50 years old. Federal agencies, in 2019, reportedly spent 80% of their IT budgets on operations and maintenance, primarily for legacy systems. [6] Research indicates that the average technical debt cost can be as high as $361,000 per 100,000 lines of code. [7] These figures are only growing, with 70% of IT budgets still allocated to maintaining these outdated systems. [8]
- Security Vulnerabilities: Older systems often lack modern security protocols, making them prime targets for cyber threats and leading to potential data breaches and compliance failures.
- Lack of Documentation and Institutional Knowledge: Over time, original architects retire, and detailed documentation becomes outdated or non-existent. This leaves current teams struggling to understand intricate code logic, making modifications risky and slow. [13, 10]
- Skills Shortages: Expertise in legacy programming languages like COBOL is dwindling. A Deloitte report noted a 23% decline in the mainframe workforce over five years, with 63% of positions unfilled, leading to extended downtimes and delayed issue resolution. [12]
- Slowed Innovation: Monolithic architectures and the complexity of integrating old systems with new technologies stifle agility and prevent businesses from rapidly adopting modern solutions like cloud computing, AI, and mobile applications.
- Resistance to Change: Despite these known risks, many organizations delay modernization, with a 2025 survey revealing that 50% cited "the current system still works" as their primary blocker, alongside budget limitations and fear of disrupting operations.
In fact, a 2025 survey found that 62% of organizations continue to rely on legacy software systems. This widespread reliance, coupled with the inherent difficulties, underscores the urgent need for more efficient and less disruptive modernization strategies.
Generative AI has emerged as a truly transformative force across various domains, and its application in software development, particularly for legacy modernization, is proving revolutionary. Unlike traditional rule-based systems, generative AI, powered by large language models (LLMs), can understand, interpret, and even generate human-like content – including code. [7, 14]
Its capabilities in code analysis and transformation include:
- Automated Code Refactoring and Translation: GenAI can analyze legacy codebases and rewrite them into modern languages (e.g., COBOL to Java or Python) while preserving crucial business logic. Studies have shown AI-driven modernization of COBOL to Java achieving 93% accuracy, reducing complexity by 35%. [13]
- Documentation Generation: By analyzing complex code, generative AI can automatically create detailed, easy-to-understand documentation, bridging the knowledge gap left by outdated or missing records.
- Performance Optimization and Security Enhancements: AI can identify inefficiencies, bottlenecks, and security vulnerabilities within legacy code, suggesting improvements and automatically patching issues.
- Test Case Generation: Understanding the intent behind legacy code allows GenAI to generate comprehensive test cases, reducing risk and improving accuracy.
- Accelerated Development Cycles: Automating repetitive and time-consuming tasks allows developers to focus on higher-value innovation. GenAI tools can accelerate writing new code by 47%, documenting code functionality by 50%, and refining existing code by 63%. [15]
The global generative AI in software development market is experiencing rapid growth, projected to reach a value of approximately USD 287.4 billion by the end of 2033, growing at a CAGR of 21.5% from 2024. Specifically, the market for generative AI in the Software Development Lifecycle (SDLC) is forecast to increase by USD 1.7 billion, at a CAGR of 38.7% between 2024 and 2029. [19] This highlights the immense demand and potential for AI-driven solutions in this space.
Fujitsu's 'Application Transform' service represents a significant leap forward in addressing the legacy modernization challenge. Powered by Fujitsu Kozuchi, the company's dedicated AI platform, this service leverages Fujitsu's extensive system development expertise and proprietary AI technology. [1, 2]
At its core, 'Fujitsu Application Transform' is designed to:
- Analyze Complex Source Code: It thoroughly examines source code from existing legacy systems, including challenging languages like COBOL.
- Automatically Generate Design Documents: The service then automatically creates high-quality, easy-to-read design documents, providing a clear understanding of the system's intricate content.
What sets 'Fujitsu Application Transform' apart is its reliance on Fujitsu Knowledge Graph-Enhanced RAG for Software Engineering. This proprietary technology significantly enhances the service's capabilities:
- Preventing Omissions and Hallucinations: By linking large volumes of source code and leveraging the knowledge graph, the system prevents common generative AI pitfalls like omissions and factual inaccuracies.
- High Accuracy and Comprehensiveness: Compared to analysis solely by general generative AI, Fujitsu's technology generates consistent design information without omissions from existing system source code, even for complex COBOL. This has led to a 95% improvement in comprehensiveness. [1, 2]
- Improved Readability: The service boasts a 60% improvement in the readability of design documents compared to conventional methods, ensuring that the generated output is not just accurate but also user-friendly.
The impact of 'Fujitsu Application Transform' on modernization efforts is projected to be substantial:
- 97% Reduction in Work Time: The service dramatically reduces the time-consuming process of understanding programming languages and generating design documents, a task that previously demanded extensive human effort.
This immense efficiency gain means that even without deep expert knowledge, organizations can rapidly gain insights into their black-box legacy systems.
Fujitsu's vision extends beyond just documentation. The company plans to sequentially introduce additional features starting in fiscal year 2026, including: [1]
- Support for rebuilding existing source code.
- Automatic rewriting of source code.
- Support for operation and maintenance.
This phased approach underscores Fujitsu's commitment to a holistic, AI-driven transformation of the software development lifecycle.
'Fujitsu Application Transform' is not an isolated offering but rather a crucial component of Fujitsu's broader, ambitious AI strategy. The company has been at the forefront of integrating AI into its operations and customer solutions:
- Earlier Initiatives: In February 2025, Fujitsu launched a software analysis and visualization service, leveraging generative AI to visualize application structures and generate design documents, projecting a 50% efficiency improvement for retail customers. This paved the way for 'Application Transform.'
- AI-Driven Software Development Platform: In February 2026, Fujitsu unveiled its 'AI-Driven Software Development Platform,' designed to automate the entire software development process – from requirements definition and design to implementation and integration testing. This platform leverages Fujitsu's "Takane" large language model and agentic AI technology. [22, 25] Internal implementation of this platform reportedly achieved a staggering 100-fold boost in productivity, reducing a task that would typically take three person-months to just four hours. [22, 25]
- Internal AI Adoption: Fujitsu has actively promoted the internal use of generative AI since May 2023, for tasks like programming code generation and internal report generation, across its 124,000 global employees.
- Fujitsu Kozuchi Platform: The underlying Fujitsu Kozuchi platform itself was announced in January 2026, designed to manage the entire generative AI lifecycle, including optimal model development, operation, and continuous improvement.
This continuous evolution demonstrates Fujitsu's deep commitment to becoming a leader in AI-driven digital transformation, aiming to shift the paradigm of software development from a conventional person-month-based approach to a customer value-based approach.
The introduction of 'Fujitsu Application Transform' brings significant implications for businesses struggling with legacy systems:
| Feature/Benefit |
Traditional Modernization Approach |
Fujitsu Application Transform (AI-Powered) |
| Code Analysis Time |
Months to years, highly manual, prone to human error. |
~97% reduction in work time. [1, 2] Rapid and automated. |
| Documentation Quality |
Often missing, outdated, or inconsistent. |
High accuracy, 95% improved comprehensiveness, 60% better readability. [1, 2] |
| Understanding Complex Systems |
Relies heavily on scarce subject matter experts and reverse engineering. |
Automated interpretation of COBOL and other legacy code, even without expert knowledge. [1, 2] |
| Risk of Errors/Omissions |
High, especially with manual analysis and lack of documentation. |
Significantly reduced, prevents omissions and hallucinations through Knowledge Graph-Enhanced RAG. [1, 2] |
| Resource Allocation |
Developers spend significant time understanding and documenting old code. |
Frees developers to focus on higher-value tasks, innovation, and new feature development. |
| Cost of Modernization |
High, due to extensive manual effort, long timelines, and specialized skills. |
Potentially significantly lower due to automation and efficiency gains. |
| Speed to Value for Modernization |
Slow and incremental. |
Accelerates initial understanding phase, paving the way for faster modernization. [13] |
This service allows organizations to gain an unprecedented understanding of their existing systems, making informed decisions about modernization strategies – whether it's refactoring, re-platforming, or complete re-architecting. It mitigates critical risks associated with legacy transformation by providing clarity and accuracy from the outset.
Fujitsu's launch of 'Fujitsu Application Transform' today, March 30, 2026, marks a pivotal moment in the battle against technical debt and the pursuit of true digital transformation. By harnessing the power of generative AI, particularly its proprietary Knowledge Graph-Enhanced RAG technology, Fujitsu is not just offering a tool; it's providing a strategic advantage. [1, 2]
Enterprises can now look forward to a future where understanding and modernizing their legacy systems is no longer a monumental, high-risk undertaking, but a streamlined, efficient, and highly accurate process. This dramatically reduced time-to-insight and documentation generation—a staggering 97% reduction in work time [1, 2]—will free up invaluable resources, allowing businesses to accelerate innovation, enhance security, and ultimately, stay competitive in a rapidly evolving digital economy. Fujitsu's 'Application Transform' is more than a service; it's a testament to how intelligent automation can rewrite the rules of software development, paving the way for a more sustainable and agile technological future.
- marketscreener.com
- global.fujitsu
- electronicsmedia.info
- itvoice.in
- global.fujitsu
- abtglobal.com
- nttdata.com
- softura.com
Featured image by Ben Neal on Unsplash
