FIS Memento

Check fraud accounts for an estimated $18.7 billion dollars in losses annually. Banks of all sizes have fraud teams attempting to detect and recover fraudulent transactions every day, but the tools they used were either manual or rule based. The goal of the Memento Check Fraud solution was to leverage statistical models and machine learning to determine which transactions were most likely to need specific investigation as well as maximize the money that could be recovered.

The goal for the product design itself was to use novel data visualizations to help agents find the needle in the haystack: cases that truly needed their attention.

Notable details:

• Discovery + planning There were no dedicated researchers at Memento, so I built up a regular user testing cadence inspired by Steve Krug’s concept of DIY usability testing — every two or three weeks we would set aside an afternoon to bring in users and subject matter experts for usability tests. I would host a post-mortem with the core stakeholder group after each session with the goal of agreeing on three things we would fix or enhance based on the tests. This became the foundation of my belief that design is a team effort, and anyone can be coached into the thought processes and artifacts used to create experiences.
• Code as high fidelity I sat within the front end engineering team and reported to the VP of Engineering. I participated in the team sprint and standup cadences, as well as working development stories related tothe Memento plaform’s Flex frontend.