Designing the future of fraud prevention
Showing the right signals from thousands of data points
Sift enables fraud teams to investigate fraud cases and provide an 'opinion' of whether or not they are observing fraud - the Sift score. To gain an accurate score, Sift utilizes machine learning and AI to observe, learn and combat fraud from over 5000 data points coming from our customers (companies who fight fraud at scale). Those data points are also available in the console for users to use during manual testing. Let me also be obvious here and say that “5000 is a lot of data points to show” - In this context, our team's ongoing challenge is to surface essential data points from the thousands of data points available to the fraud agent (our main target persona).
Variable information density
To achieve clarity of data while keeping our users informed with minimal intervention, we had to create a UI system and patterns that allow for surfacing the top data points, while maintaining depth and information available within a click or in context. This has become one of our design processes - allowing every feature to have variable information density (high-level glances > details on context > and a fully detailed view within a click). This allows users at any proficiency level to easily dive in and out of the data with speed.
