WBT Pricing Engine

Estimating web-based training and explainer-video projects becomes difficult when business rules live in several places and are interpreted differently by sales, production and software. The WBT Pricing Engine addresses that problem by turning the pricing logic into one explicit, testable system.

I designed the architecture around a Markdown playbook as the human-readable source of truth. A mirrored TypeScript configuration powers the calculation engine and React inquiry interface. During the build, an automated validation step compares documented values with the implementation; if the two diverge, the build fails before inconsistent pricing logic can be published.

This was an intentional design decision rather than a documentation convenience. The goal was to make the rules understandable to non-developers while still giving the software a deterministic configuration that can be validated automatically.

The same playbook also provides structured instructions for Gemini-supported inquiry handling. That allows human users, the application and AI-assisted workflows to operate from the same pricing model instead of maintaining separate interpretations.

The project is a good example of my systems-oriented work: I was not only building an interface, but designing the relationship between documentation, business rules, validation, software behavior and AI assistance so the overall process remains consistent and maintainable.