Contact us
Generative AI in bank: How MBH Bank Leverages GenAI to help its review processes
Generative AI in bank: How MBH Bank Leverages GenAI to help its review processes
Client:
Region:
CEE
Industry:
Fintech
MBH Bank aims to be a market-leading group of companies in the Hungarian banking sector, prioritizing sustainability and innovation. As part of its modernization strategy, the company has embraced cloud technologies, particularly AWS, to optimize its operations and software development practices. To further enhance its development process, MBH Bank partnered with SnapSoft to implement a generative AI-based static code analyzer. This custom solution, built on AWS Bedrock, was designed to provide automated linting and code analysis for pushed code. Unlike traditional tools, it helps identify issues that were previously difficult to detect, streamlining code reviews and improving overall development efficiency. This initiative reflects MBH Bank's commitment to leveraging advanced technologies to support its development teams and ensure software quality.
Our partner said
Customer Challenges
The development of a generative AI-based code analyzer introduced several complex challenges. Extensive research was required to select the most suitable AI models for analyzing code effectively. Crafting prompts that ensured accurate, context-aware feedback was critical, as was addressing the issue of hallucination, where the AI might provide irrelevant or incorrect suggestions. Supporting multiple programming languages added further complexity, requiring careful design and implementation. The system also needed to deliver high-speed performance to seamlessly integrate with existing workflows, which demanded significant optimization efforts. SnapSoft relied heavily on developer feedback to refine the tool, continuously improving its effectiveness based on real-world experiences.
SnapSoft’s contribution to the solution
SnapSoft designed and implemented the solution entirely from scratch, tailoring it to meet the specific needs of MBH Bank. The team built a robust application framework, trained and tested the AI on extensive volumes of existing code, and iteratively enhanced the solution through real-world testing and developer input. These refinements ensured the tool could handle edge cases effectively and provide meaningful suggestions. The final solution was seamlessly integrated into MBH Bank's CI/CD pipelines, automating linting and accelerating the development process.
Why AWS?
The project utilized AWS because of its seamless compatibility with MBH Bank's existing cloud infrastructure. AWS Bedrock provided access to cutting-edge generative AI models, including Claude, which proved particularly effective for code analysis. The platform's reliability and scalability ensured smooth operation even under demanding conditions. Moreover, AWS allowed the solution to be configured securely, enforcing policies that ensured it operated exclusively under predefined accounts. This approach aligned with MBH Bank's rigorous security and compliance requirements. The collaboration between MBH Bank and SnapSoft highlights how generative AI and cloud technologies can revolutionize development practices, delivering faster, more reliable tools that empower developers and enhance software quality.