Customer
Big debt collection agency in Poland looking for optimzing the collection process.
Challenges
- Maximized chance for repayment is to be called by an experienced collector. However the number of phone calls is limited by human resources.
- Not calling a debtor who needs human persuasion results in further delinquency and greater risk for non repayment, but calling a debtor who doesn’t need additional persuasion is a wasted effort.
Solution
- Building a mathematical model of the collection process and calibrate its parameters using historical collection data.
- Machine Learning is applied to directly predict the eventual collection outcome from any point in time. This allowed us to estimate the value of calling a debtor at any time by calculating the difference in expected repayment with and without calling the debtor.
- Built in automatic system of recommendation and update of the proceedings path in the debtor’s case, based on current variables and data
- Optimization of settlement parameters with the debtor at an angle the possibility of keeping it
- Automation of decision making
- Personalization of the message content to the debtor
