The AI Refactor: Upgrading 20-Year-Old Enterprise Systems to Modern JS Architecture

For many organizations, the biggest challenges aren?t new technologies?it?s the 15?20-year-old systems still running mission-critical operations. In my role modernizing large-scale manufacturing and ERP applications, I worked directly with Classic ASP pages, VBScript business rules, UniData/ODBC lookups, and printer/label subsystems that had zero documentation. Rewriting them required more than code?it demanded AI-assisted archaeology.

In this session, I will walk through how I broke down a monolithic legacy application and rebuilt it as a scalable JavaScript microservices ecosystem. Using real examples?such as migrating a VBScript-driven data collection flow into a Node.js service, converting ODBC-based ERP queries into Python micro-APIs, and using LLMs to untangle hidden business rules?I?ll show how AI speeds up the entire modernization process.

You?ll learn practical patterns, migration sequences, Strangler-Fig architecture, and how AI helps validate logic while keeping production stable. If you?ve ever been afraid to touch legacy code, this talk gives you the exact playbook I used to safely refactor a 20-year-old system into an AI-native JS platform.

Comments