A Tier 1 Global Bank was preparing to embark on an enterprise-wide legacy modernization initiative to address the challenges posed by aging mainframe systems. In support of this transformation, Capco conducted a comprehensive Proof of Value (POV) to demonstrate how its proprietary Legacy Modernization Toolkit could be leveraged to modernize a complex sample codebase effectively.
The POV encompassed application analysis, intelligent documentation, forward engineering, and automated testing, showcasing the potential of GenAI to accelerate and de-risk legacy modernization efforts.
The client’s key objectives included:
• Reverse engineering source code and generating comprehensive, user-friendly technical documentation
• Extracting embedded business rules and gaining deep insights into data behaviour
• Defining a future-proof target state architecture
• Performing these activities at scale across a large, complex codebase
• Generating high-quality modern code at acceptable velocity
• Proving functional parity between the legacy and modern code stacks
• Demonstrating a developer-friendly, human-in-the-loop experience throughout the process
This initiative highlighted how AI can dramatically improve modernization velocity, accuracy, and maintainability while reducing the burden on engineering teams.
High-Level Approach
To support the bank’s modernization objectives, Capco leveraged its proprietary Legacy Automation Toolkit, specifically designed to streamline and accelerate the most labour-intensive aspects of legacy transformation. The toolkit focuses on delivering tailored, scalable solutions that enable organizations to modernize with confidence and control.
The approach included three core components:
1. Reverse Engineering
Capco’s platform generated persona-specific, contextualized knowledge views to support developers, architects, and business analysts. This was augmented by an "Ask-AI-Anything" chatbot, capable of performing real-time impact analysis and answering complex technical queries to improve traceability and planning.
2. Target State Discovery
Capco defined optimized target services that aligned to modern architecture standards. This involved consolidating duplicate functionality and ensuring that each service was built for performance, scalability, and maintainability laying the foundation for a sustainable, future-ready environment.
3. Forward Engineering
Using advanced GenAI techniques, the solution generated high-quality, production-grade code that was virtually indistinguishable from human-developed code. This included unit tests, design assets, and seamless GitLab integration. Outputs were designed to be easily testable via in-memory database environments, ensuring rapid validation and deployment.
This structured methodology enabled high-fidelity modernization at scale, with a strong emphasis on developer usability and technical excellence.
Key Benefits
• Over 2 million lines of COBOL code were successfully analyzed and restructured, with AI used to deeply examine around 150,000 of those lines.
• The project involved reviewing and modeling 2,400 jobs, 1,400 COBOL applications, and 31,000 individual code blocks.
• The automation and efficiency gains are expected to free up around 150,000 hours of work each year.