Legacy modernization with Gen AI for leading global bank

About the client

The client, one of the world’s largest financial institutions, provides a diverse portfolio of financial services, including retail banking, credit cards, loans, and wealth management. Operating globally, it serves millions of retail and commercial banking customers and has consistently aimed to enhance operational efficiency while delivering superior customer service.

Business goals and objectives

To address critical business needs and support their growth strategy, the client initiated a global transformation program with the goal of legacy platform modernization with Gen AI. Their key objectives included

• Replace the legacy collections technology stack with a modern, customized, and cloud-native platform featuring an integrated decision engine.
• Achieve measurable savings, reduce credit losses, and enhance customer service through advanced digital capabilities.
• Use technologies, such as Natural Language Processing (NLP) and generative AI, to improve the efficiency and effectiveness of collections operations.
• Streamline retail banking processes through its banking application’s productivity optimization.
• Transition from traditional mainframe systems to a robust, future-proof technology stack that improves performance and flexibility.

Challenges

The client faced several pressing challenges with existing legacy systems, which reduced efficiency and the ability to innovate, including

• High maintenance costs: Legacy systems required significant resources for upkeep, resulting in increased operational expenses.
• Scalability limitations: The legacy systems struggled to keep pace as transaction volumes and business complexity increased, slowing their ability to adapt to market demands.
• Lack of flexibility: The platform was rigid and unable to integrate with modern technologies or adapt to evolving business needs.
• Shortage of skilled resources: Maintaining the legacy systems required niche skills that were increasingly difficult to source and retain, impacting platform maintenance and operations.

LTM’s solution

LTM designed and implemented a customized agency management solution, leveraging generative AI and advanced technologies, to address the client’s challenges. The agency management platform was tailored to optimize their collections and recovery processes across wealth and private banking, enabling scalable growth. Key highlights of the solution include:

• Accelerated code development : Transformed high-level descriptions into well-structured and efficient code for the agency management platform’s features using generative AI. This reduced development cycles while ensuring consistency and alignment with industry best practices.
• Enhanced coding standards : Advanced machine learning algorithms identified inefficiencies and redundant code, enabling seamless enhancements. Developers effectively reduced maintenance time, elevating overall code quality and reliability.
• Automated code revisions : Gen AI analyzed existing codebases to address inefficiencies and ensure alignment with the latest standards. Automated implementations streamlined updates, reducing the time and effort for code revisions.
• JUnit test case automation : Generated comprehensive and detailed test cases tailored to application requirements. This ensured robust test coverage, accelerated development, and minimized errors.
• Document generation with Copilot : Leveraging GitHub Copilot, we generated documents directly from Angular-based code, enabling seamless migration from Angular to ReactJS during the client’s ongoing technological transformation.
• Code assistance for compliance : Gen AI provided actionable recommendations and code updates to meet Continuous Integration/Continuous Deployment (CI/CD) pipeline standards without impacting programming logic, ensuring seamless compliance with coding guidelines.
• End-to-end test automation : GitHub Copilot generated test scenarios, cases, and data for automation tools like Behavior-Driven Development (BDD) Cucumber, Selenium, and REST APIs. This enabled efficient test automation across multiple processes, enhancing overall quality and speed.

Technology stack

Key technologies employed in the solution include:

• Generative AI for code automation, document generation with Copilot, and JUnit testing frameworks
• ReactJS and Microservices for front-end and API development
• Google Cloud Platform (GCP) BigQuery (BQ) for cloud enablement
• Selenium and BDD Cucumber for testing automation and creating a reusable framework

Business outcomes

• Improved productivity by 5-10% across critical collections and recovery modules such as litigation, bankruptcy, and administration.
• The modular design included 17 reusable frameworks that enabled sustained scalability.
• Enabled expansion into 13 additional regions requiring minimal regional-specific customizations.
• Reduced costs associated with system maintenance through automation and simplified coding processes.
• Enabled seamless integration of advanced Gen AI capabilities and cloud technology.
• Enhanced the quality and efficiency of collections operations, driving better customer experiences with digital capabilities.

Client Testimonial

Thanks to you and your team for all the efforts that helped us go live last weekend. Additionally, you encouraged the team to adopt Copilot, which has greatly benefited us, especially in unit testing, code coverage, test case preparation, and other areas. Thanks very much for your efforts and support in ensuring wider adoption by the teams.

I look forward to continuous support in pursuing further opportunities to utilize AI tools effectively to improve productivity.

Regards

Vice President, Global Technology Practice (Banking customer of LTM)

Contact us at BFS.AI@LTM.com to reimagine your collections and recovery operations with future-ready agent management solutions.