Fraud Detection:  Fraud is a significant challenge in many industries.  Perpetrators take advantage of the huge volumes of data that companies must process with limited resources.  Predictive models help by quickly scanning many thousands of cases (claims, transactions, customers, etc.) and flagging those that are questionable for further review.

Credit Application Scoring:  Predictive models have been used for many years to determine if a given applicant is an appropriate fit for the targeted risk profile of a given credit related products.

Portfolio Management:  Predictive models can identify stocks and sectors that may fly and those that are likely to crash. 

Market Timing:  Conventional wisdom says "the market cannot be  timed."  However, predictive models are used daily to pick position entry and exit points in order to maximize return and minimize risk.