Riskified is a SaaS fraud-detection platform used by global e-commerce merchants. I worked across both the core web application and the marketing website, translating complex risk systems into clear, data-driven interfaces.
The application was designed to serve two distinct personas with different needs and workflows — each requiring a different primary view into the same underlying data.
Need a high-level overview of transferred orders, approval rates, and reasons for declines. Primarily use the dashboard view.
Handle order-level actions such as chargebacks and disputes. Primarily work within the Orders tab to review and manage individual transactions.
Declined-order view Fraud detection is an inherently abstract problem — invisible signals, probabilistic outcomes, money that exists only as data. The design challenge was making risk legible without oversimplifying it. The solution uses dots as a direct metaphor for individual orders: each one a transaction moving through the system, approved or declined, guaranteed or flagged. Motion and density do the explanatory work that numbers alone can't.