AI transformation projects deliver agility & innovation
We delivered a large-scale AI transformation project to help a leading bank in Southeast Asia reimagine its Loan Origination System (LOS). Leveraging iSoftStone’s AI software engineering toolchain, our team modernized the codebase, improving critical processes across retail and corporate lending and achieving a 5x speedup in operations.
Legacy systems blunted the bank’s competitive edge
The bank had a large customer base and a complex financial services portfolio. Its LOS had operated reliably for many years but was built on a legacy stack with a large and complex codebase. Even minor fixes took significant manual effort.
Experienced engineers familiar with the technology the LOS was built on were retiring. System documentation had become incomplete or was inconsistent with the actual codebase. Maintenance costs were getting higher, development cycles were getting longer, and support response times were getting slower.
In short, as fintech was evolving rapidly in the wider world, the LOS was aging badly.
The bank needed scalability and flexibility to support new lending products and evolving regulatory requirements.
With AI supporting code generation, error correction, and documentation, development and maintenance efficiency improved by over 40%. Code defect density dropped by 50%.
AI software engineering toolchain & new features
First, we worked with the bank’s business and IT leaders to understand their businesses objectives for the next five to 10 years. We drew on our experience in the global finance industry and, together with the bank, developed an LOS modernization roadmap emphasizing agility, intelligence, and stability.
Then, we leveraged iSoftStone’s proprietary AI software engineering toolchain to support the modernization of the LOS by:
- Analyzing the legacy codebase and reverse-engineering architecture and data flow diagrams and core business documents.
- Automating modernization and fixing vulnerabilities by translating and refactoring the legacy code into modern, efficient, maintainable code.
- Automatically generating documentation, including dev documentation, API documentation, and user manuals, and then syncing them to the codebase.
After modernizing the core of the LOS with AI, the project team adopted an agile development approach, rapidly delivering new functional modules in line with the bank’s vision and strategy. An AI-powered intelligent credit approval assistant, a mobile loan application portal, and real-time interfaces with third-party credit reports were quickly deployed.
40% efficiency improvements & faster launches
The AI-modernized LOS has significantly improved the bank’s ability to respond to market changes, delight customers, and find a new edge against its competitors.
With AI supporting code generation, error correction, and documentation, development and maintenance efficiency improved by over 40%. Code defect density dropped by 50%, significantly improving system stability and reducing production risks.
System changes that previously required months of engineering can now be delivered speedily. The team is configuring new products in the LOS and launching them to the public in just three weeks.
