Problem

How do you bring an old manual and time-consuming process as an optimised feature inside the product?

The way we classified spending data at the time was a slog—slow, tedious, and full of mistakes. People had to do the same mind-numbing tasks over and over. Everyone had their own opinions, but most were half-baked assumptions, and nobody really knew how things actually worked. Before jumping into solutions, we needed to dig deeper—how were people handling this manually? What patterns were we missing? How much of a headache was this really causing?

Project Goal

Build an MVP that simplifies classification by automating steps, boosting accuracy, and cutting manual work to enhance FTE efficiency.

Key Metrics

• Reduce classification effort by 40% - 50% and optimize FTE performance.​

• Increase accuracy of classified data by reducing classificaiton errors.

Problem Discovery

Before diving in, I needed to figure out what user research was required to test our ideas and fill in the missing pieces.