Supply Chain Analytics Leader Accelerates Midrange Legacy Modernization using AI-powered Precision

600,000

Lines of Code Analyzed

1,624

RPG Files Documented

Zero

Risk using On-premises AI deployment

The Customer

A leading U.S.-based procurement and supplychain analytics enterprise.

The Challenge

The customer embarked on modernizing a large-scale IBM i (AS/400) midrange legacy system comprising more than 55 million lines of code. The initial phase focused on analyzing 600,000 lines to inform and prioritize their modernization roadmap. As part of the effort, the company required detailed Business Requirements Documentation (BRDs) and Technical Requirements Documentation(TRDs), structured and optimized for AveriSource partner, PwC's forward-engineering platform to accelerate the system’s rewrite to C#.NET.

The scope included processing 1,624 RPG files and 566 CL files, extracting complex business logic, and generating documentation to support downstream application development. Due to the large code base, the customer needed a solution that could handle significant processing volume and velocity while maintaining control over their data through a private AI infrastructure.

The AveriSource delivery team calibrated the AveriSource Platform to produce BRD and TRD output that was fit-for-purpose, aligning with PwC’s forward engineering methodology, ensuring a seamless handoff from legacy analysis to modern application development in C#.NET.

The Approach

The AveriSource Platform enabled the customer to rapidly analyze their extensive IBM i portfolio and generate forward engineering-optimized business requirements documentation, supporting an accelerated rewrite strategy from RPG and CL to C#.NET through AI-powered business rule extraction and customized requirements generation.

Inventory and Discover

Using AveriSource Platform Inventory and Discover, the customer conducted a comprehensive analysis of their IBM i application architecture. The assessment covered 1,624 RPG files and 566 CLfiles, mapping program structures, data flows, interdependencies, and relationships across the targeted 500,000-600,000 lines of legacy code.

Product Reports and Features Utilized:

Inventory: missing and unreferenced files
Business logic visualization
Dependencies
Data Lineage
Scalability
RPG and CL Language Support

Analyze

Using AveriSource Analyze with integrated on-premises AI capabilities, the team identified, analyzed, and extracted core business rules from the RPG and CL programs. The platform utilized a horizontally scaled deployment of Meta Llama 3.1 8B on the customer’s private Azure infrastructure to automatically generate comprehensive business requirements documentation.

Product Reports and Features Utilized:

Business Rules Extraction
AI-Powered BRD Generation
Code Analysis
Rule Analysis
Customized BRD Templates

The Results

The AveriSource Platform successfully delivered comprehensive, forward engineering-ready business requirements documentation to support the customer’s IBM i modernization initiative, achieving the following outcomes:

500,000-600,000 lines of IBM i code analyzed from a 55 million line total codebase
1,624 RPG files and 566 CL files processed and documented
BRD generation calibrated for PwC's forward-engineering tools
Horizontally-scaled GenAI architecture enabling high-volume, high-velocity processing through multiple parallel threads
On-premises AI deployment using Meta Llama 3.1 8B on Azure, ensuring full data security and governance control
Established an accelerated rewrite migration pathway from legacy IBM i RPG/CL to a modern C#.NET architecture
Seamless integration between legacy analysis and forward engineering processes, reducing manual translation effort and project risk