Authors:
Hema Latha Boddupally

Addresses:
Department of Technical Architect, Information Technology, R3 Technology Inc, Hillsborough, New Jersey, United States of America.

Abstract:

Generative AI-powered code transformation techniques for legacy .NET systems address the long-standing challenges encountered by organisations that depend on ageing application stacks, tightly coupled components, and outdated development models that limit scalability, maintainability, and integration with modern architectures. The objective of this study is to examine whether generative AI can provide a structured and dependable pathway for modernising complex .NET applications while reducing the extensive manual effort traditionally required for refactoring and architectural redesign. The research problem centres on assessing the ability of generative AI to interpret the semantics of legacy code, propose optimised transformations, and preserve functional accuracy throughout the modernisation process. A mixed-methods approach was used, incorporating quantitative evaluation of code quality gains, transformation precision, and performance improvements, alongside qualitative analysis of developer experience, maintainability, and architectural alignment. Key findings show that generative AI accelerates modernisation workflows, enhances consistency across transformed modules, and supports the transition of legacy logic to modular, cloud-oriented, and testable designs. The study contributes strategically by presenting a structured framework for integrating AI-assisted refactoring into enterprise modernisation initiatives and academically by offering a comprehensive assessment of generative AI capabilities in legacy system evolution. The results indicate that generative AI can act as a powerful catalyst when combined with human oversight, validation, and governance. 

Keywords: Generative AI; Code Transformation; Legacy .NET Systems; Automated Software Engineering; Structured Framework; Modernisation Workflows; Cloud-Oriented; Testable Designs.

Received on: 13/01/2025, Revised on: 09/04/2025, Accepted on: 08/06/2025, Published on: 09/12/2025

DOI: 10.69888/FTSCS.2025.000527

FMDB Transactions on Sustainable Computing Systems, 2025 Vol. 3 No. 4, Pages: 253-263

• Views : 86
• Downloads : 10

Download PDF