PayRank started with a series of conversations between Michael and his lifelong friend Brian, who has spent more than 15 years working in the collections industry. As Michael was stepping away from a traditional actuarial career, they began comparing notes on the analytical tools used in their respective fields. Michael described work he had done in healthcare call-center operations, where propensity models helped determine which members warranted an in-home visit, which could be served by telephone, and how frequently outreach should occur. The underlying idea was simple. Limited resources should be directed toward the people and actions most likely to produce a meaningful result and ROI should be tightly calibrated and measured.

Those conversations exposed a similar problem in collections. Many agencies still rely on balance, placement age, bureau scores, static rules, or collector judgment to decide which accounts receive attention. Bureau scores can help, but they are expensive and were not designed to optimize collection recoveries. Larger organizations can build sophisticated prioritization systems internally, but doing so requires expensive analytics talent and infrastructure. That gap became the starting point for PayRank. Early development was tested against real collections portfolio data, where the separation between high-value and low-value accounts was compelling enough to justify building the concept into a full production platform, including portfolio-specific modeling, automated scoring workflows, redundant infrastructure, and a client-ready analytics environment.