RWE Modernizes Legacy App in Two Days

RWE Generation Ltd, one of Europe’s leading energy producers, faced a growing operational risk: a number of aging, undocumented applications were running on outdated technology stacks. These systems—many over two decades old—were essential to operating power plants, but impossible to update, scale or understand without extensive cost, risk, effort and downtime.

To prove the potential of AI-assisted modernization, Publicis Sapient and RWE chose one particularly stubborn application: Tube Tracker. The goal was to show how generative AI, combined with engineering expertise, could transform a critical legacy asset into a modern, maintainable foundation in just days.

Tube Tracker, a visual interface used to manage pipe systems in RWE’s power plants, was vital for finding damaged infrastructure quickly. However, it was over 24 years old, written in Java and had no accessible source code, documentation or experts left to maintain it. In its current form, the application posed not just a technical challenge but a serious business continuity risk.

Publicis Sapient used Sapient Slingshot, our AI platform for automating and accelerating the software development lifecycle, to accelerate Tube Tracker's modernization through a five-step process in just two days:

1. Decompilation: Using open-source AI tools, we converted binary files into readable Java source code—a critical step that modernization simply couldn’t happen without.
2. Rebuild: We created a modern development environment with Java 17 and PostgreSQL 16, enabling the app to run on current systems for the first time in years.
3. Refactoring: Slingshot was used to clean up and restructure the old code, reducing over 7,000 lines to a sleek, readable 5,000, with updated syntax, naming conventions and unit tests.
4. Business logic extraction: Slingshot analyzed the code to produce entity relationship diagrams and data flow sequences, revealing the app’s core functionality, something no one at RWE could previously access.
5. Documentation generation: With AI assistance, we created full inline documentation and external README files so future developers can quickly understand and extend the codebase.

At nearly every step, generative AI was paired with human oversight, ensuring code quality, clarity and correctness with an approach that drastically reduced effort and risk.

By modernizing Tube Tracker, RWE eliminated a critical operational risk while proving that even the most stubborn legacy systems can be transformed quickly and safely. What was once an opaque, untouchable application is now a fully documented, maintainable asset that RWE teams can deploy, evolve and extend with confidence across sites.

Beyond the technical outcome, the project sparked a broader shift in how modernization is approached. With AI paired with human oversight, RWE gained full transparency and control—building trust in a new, faster way of modernizing without replacing engineering expertise.

The results speak for themselves:

• Application revived: A former black box is now a living, annotated codebase ready for maintenance and reuse across RWE sites.
• Development in days, not weeks: One engineer modernized the application in two days—compared to an estimated two weeks of manual effort.
• Significant time and efficiency gains: 35–45 percent time savings in automated code generation and 30–40 percent efficiency gains in test creation and setup.
• Risk eliminated: Compliance, security and upgradeability concerns were fully addressed.
• Cleaner, modern codebase: Lines of code were reduced from ~7,000 to ~5,000 through refactoring and modern syntax.
• Increased confidence and scalability: The application is now deployable across additional sites with zero rework.

Together, these outcomes show how RWE turned AI-assisted modernization into a practical, repeatable model—making engineers faster, smarter and more consistent while safeguarding critical operations.