We analyzed and documented the workings of an ERP system. We combined AI-assisted source code analysis with expert validation and regular consultations with the client.
Sector
Service & Technology
The Challenge
Our client, an agricultural commodity trading company, engaged us to comprehensively document its ERP system and carry out a technical and IT security review. At the start of the project, the system was unfamiliar to us, and alongside its source code we had only a rudimentary specification covering approximately 20% of its functionality.
Some of the missing information was known to the client’s employees, but was available primarily through verbal discussions and meeting notes. Our task was therefore not simply to supplement existing documentation, but to reconstruct the operation of the entire system and develop the business system design and technical assessment documentation from the ground up.
The Solution
We based our investigation on an analysis of the available source code. Using AI-assisted reverse engineering methods, primarily GitHub Copilot and ChatGPT, we accelerated our understanding of the unfamiliar codebase and the preparation of documentation.
We continuously cross-checked the technical information extracted from the source code against the business knowledge provided by the client. Regular consultations were particularly important, as not every business rule and operational relationship could be reliably reconstructed from the code alone.
The findings of the AI-assisted analysis were processed under expert supervision. This combination of technical investigation and client-side knowledge formed the basis for the business system design and technical assessment documentation.
Technology
GitHub Copilot, ChatGPT
The Result
As a result of the project:
- we produced the business system design and technical assessment documentation for the grain-trading ERP system;
- we reconstructed the operation of the previously only partially documented system using its source code and information provided by the client;
- the client accepted the completed documentation and the findings of the technical assessment.
The project demonstrated that AI-assisted reverse engineering can accelerate system analysis, while expert validation and ongoing dialogue with the client remain essential.
