ZClaim Reduces Worker Comp Claim Loss by 40%

Our predictive model ranks new claims on day 0 based on a risk score, the higher the score the more expensive the claim will be. The top 10% claims count for more than 80% of the total losses. During the life time of a claim, as new information become available, the claim will be re-scored by the model. Our product generates alerts for those top 10% claims. Analysts worked on them first. By doing so, the claim loss is reduced by 40%. The following is a partial list of variables used by the model:

  • Age
  • Sex
  • Injury Body Part
  • Injury Cause
  • Length of Employment
  • Annual Salary

Our deposit check models reduce fraud losses by 70%

Our deposit check fraud models have been running in a top 15 bank for more than 3 years. These models have reduced the charge-off due to fraud by 70% without the increase of the number of fraud prevention staff. This translates into tens of millions of dollars of savings. The input variables to the models include:

  • Deposit Amount
  • The Ratio Between Deposit Amount and Life Time High Amount
  • Number of Checks in the Deposit
  • Is This a New Account
  • Account Opening Branch