Case Study
Fintech company transformed legacy fraud detection application with cloud-based microservices architecture
OUR CLIENT
Founded nearly two decades ago, this U.S. software company helps businesses identify fraud and manage compliance. The company serves large enterprises and government agencies using the latest technologies, like artificial intelligence (AI) and machine learning (ML).
THE CHALLENGE
Needed application development expertise to modernize legacy fraud detection system
The client’s fraud analytics platform needed an overhaul. The application’s legacy technologies couldn’t scale to keep pace with growth, and the IT team couldn’t easily integrate the system with more modern third-party systems. Also, the dated, monolithic architecture delayed development cycles, so the company couldn’t rapidly deliver new revenue-generating features and functionality to customers. The company needed a technology partner with extensive software engineering experience to help design, develop, test, and deploy a modern version of the fraud analytics application.
THE TRANSFORMATION
Deployed innovative cloud-based application using DevOps best practices and leading tools
Drawing on extensive fraud analysis product expertise, the project team assessed the company’s existing technologies, goals for the new solution, and go-to-market strategy to transform the legacy application into a modern, component-based cloud architecture. From that assessment, the project team kicked off the development effort by migrating the monolithic application to a web-based infrastructure with reusable microservices. The application featured optical character recognition (OCR), using AI, ML, and data analysis to rapidly process purchase receipts to pinpoint potential fraud. To ensure efficient continuous development processes, the project team:
- Implemented shift left and shift right approaches to testing—to identify quality control issues as early as possible in the development cycle under real-world transaction load conditions.
- Automated unit, smoke, and regression testing—using a cloud-agnostic test automation framework.
- Streamlined security testing—with automation tools like OWASP Zed Attack Proxy and Burp .
- Deployed cutting-edge technologies, like Kratos, Dynatrace, and a web page analyzer—to improve API performance, monitor application up-time, and increase client-side application performance, respectively.
- Improved accessibility for disabled users— with tools like, JAWS , the leading screen reader, and WAVE , an accessibility evaluation tool, as well as a web accessibility toolbar.
THE IMPACT
Delivered tremendous application development improvements and customer benefits
The new modern, cloud-based fraud analysis application helped the software engineering team:
- Automate 90% of code testing.
- Enhance application performance by 70%.
- Improve API performance by 40%.
- Decrease downtime incidents by 30%.
- Accelerate new functionality time-to-market by 25%.
- Reduce release deployment costs by 15%.
- 70% reduced P-Card policy violations.
From a fraud detection standpoint, the new application found $3 million in previously unidentified duplicate payments in just one year of historical data analysis. This indicates the solution will significantly improve the company’s core mission to help businesses detect fraud and financial errors. With the new automated OCR functionality, the company’s customers can now track, monitor, and analyze 100% of their payables data without manual intervention. The new AI and ML-based entity extraction capabilities have helped customers reduce manual audit efforts by 50%.
RESOURCES
Product Engineering Services and Solutions | UST
Software Product Engineering | UST
Financial Consulting Services and Transformation Solutions | UST