The Customer
Spring Point develops integrated ERP and quality management software designed for the industrial sector, combining MotorBase® and QM Wizard™ to enhance training, consistency, and quality while bridging the skills gap across workforces of any size.
The Project
Modernization BA Support Chatbot on Legacy ERP
Overview
Improving collaborated with Spring Point Solutions, a small but impactful ERP provider for motor repair shops that was recently acquired by ABB. The project aimed to accelerate modernization of their existing on-premises ERP application by eliminating organizational bottlenecks and augmenting their off-shore development team.
Given the complexity and long history of the application, which had evolved over several decades, the critical challenge was to extract and understand the hidden business rules and requirements embedded in the legacy code. To address this, we developed Code Explorer, a business analyst support chatbot that could interact with the legacy ERP code, helping to streamline and expedite the modernization process.
The Business Challenges
Spring Point Solutions faced the daunting task of modernizing a decades-old ERP system to make it cloud-compatible. This system, originally designed for on-premises deployment, had accumulated layers of updates and modifications over many years. The core developer who built significant portions of the ERP was the sole repository of intricate system knowledge, creating a bottleneck.
They needed a way to extract and document the business rules and functionalities from the complex legacy codebase without monopolizing the developer's time, to facilitate a smooth and efficient modernization process.
Our Solution
We decided to leverage the capabilities of our proprietary tool, Code Explorer. Code Explorer uses advanced machine learning models to perform a comprehensive analysis of the legacy codebase. This tool was configured to run within Spring Point’s chosen cloud environment, ensuring data security and compliance. By utilizing an LLM (Large Language Model) to interpret and summarize the code, we provided a user-friendly chatbot interface that allowed their business analyst and offshore development team to query and understand the code's functionality. This reduced dependency on the core developer and allowed for a faster, more accurate extraction of requirements.
The Business Benefits
Reduced Modernization Timeline: Cut down the expected modernization time from over two years to feasible within a year.
Enhanced Development Efficiency: Enabled offshore team to explore and understand the legacy application much faster, reducing back-and-forth communication and human errors.
Risk Mitigation: Minimized the risk of losing critical functionalities during modernization by accurately extracting business rules from the legacy code.
Cost Savings: Leveraged offshore resources effectively, reducing the overall cost of the modernization project.
Scalability: Facilitated the transition to a scalable cloud-based ERP system, positioning the company for future growth and integration within ABB’s ecosystem.
Improved Documentation: Created a detailed and searchable knowledge base of the ERP’s business rules and functionalities, enhancing future maintenance and updates.
Technologies & Methodologies Used
Python: Used for developing the core of Code Explorer due to its robust libraries and frameworks for AI and data processing.
Lang Chain and Lang Graph: Utilized for creating and managing the analysis workflows.
Antlr: Employed to generate abstract syntax trees for deeper code analysis.
Docker: Containerized the application to ensure smooth deployment across different cloud environments.
React: Developed the front-end interface for interacting with the chatbot.
Neo4j: Used as the database to store and query the knowledge base of extracted information.
Partnerships
While the bulk of this project was handled internally, some groundwork for future capabilities was laid in collaboration with external partners. These partnerships aimed at enhancing the functionalities of Code Explorer for its next iteration, ensuring it remains at the cutting edge of software modernization tools.
Lessons Learned
Client Education: Consistently educate the client about the process and findings to build trust and ensure transparency.
Tool Flexibility: Ensure the tools and technologies used can be easily adapted to different environments and requirements.
Abstract Syntax Trees: While useful, these can be resource-intensive; alternative methods might be more efficient depending on the scenario.
Effective Query Design: Training users to ask precise questions significantly enhances the effectiveness of AI tools.
Data Security: Ensure that sensitive data is adequately protected by configuring analysis tools to run within client-controlled environments.
Iterative Feedback: Regular interaction with the client during the development process helps refine the tool and its outputs, leading to better overall results.
Conclusion
The Modernization BA Support Chatbot project for Spring Point Solutions showcases Improving's capability to deliver innovative solutions for complex legacy systems. By leveraging Code Explorer, we significantly reduced the modernization timeline, enhanced development efficiency, and minimized risks associated with business rule extraction. Our approach not only provided immediate benefits but also positioned Spring Point for seamless integration and future growth within ABB’s ecosystem. This case study underscores our unique advantage in combining advanced AI with strategic consulting to solve our clients' most challenging problems.
