This intelligent software solution applies advanced predictive analytics to evaluate and quantify subrogation potential of incoming claims. It uses what it has learned from the claims characteristics and recovery outcomes of more than 35 million past claims to recognize patterns that signal subrogation in current claims. Neural network-based predictive models look at complex, nonlinear relationships between multiple pieces of claim data to find hidden patterns. Context Vector technology is also used to interpret notes and detect subtle subrogation signals that might not be recognized until later in the claim life cycle or never at all. Claims are scored for recovery potential, enabling insurers to focus their investigative resources where recoveries are likely to be substantial.