Application Fraud Analytics services help organization prevent, detect and control fraud by building early warning mechanism to identify patterns that lead to identification of suspicious, illegitimate and fraudulent transactions.

The most effective solution is to detect fraud at the application stage. Market Equations applies predictive analytics and modeling techniques to authenticate, validate and detect fraudulent applications helping organizations avoid loses caused by bad debts and write-offs and high operational costs incurred on recovery and collections.

FREE Data Health Test and Consultation. Contact Us!

Market Equations India helps a leading Bank design and develop a predictive model that could predict the future propensity of a customer committing fraud at the application level at pre RIC stage for their unsecured lending products (Personal Loan, Business loan and Small Business loan).

Objective:

• Design and develop a predictive scorecard to help identify potential fraud accounts at the application stage.
• Optimize fraud detection efforts by building smart risk analytics and mitigation strategies by identifying customers most likely to fraud or turn bad debt.

Challenge:

The Risk Intelligence and Control (RIC) department of a leading bank was facing huge challenges in controlling fraud for their unsecured lending products (Personal Loan, Business loan and Small Business loan). Adding to this challenge was the unavailability of resources internally to scan and verify each application form making it impossible for the department to apply advanced verification systems such as CPV etc.

Approach:

Using customer profile data, credit and payment history we developed a predictive model that could predict the future propensity of a customer turning fraud at the application level. Our predictive model bucketed and defined (Good-Bad) customers having a high propensity of turning fraud and helped identify customers with high propensity to turn fraud in the first 3 months of the relationship.

• Good: Application approved
• Indeterminate: Status = 'WIP' (Work in Progress) or 'Rejected'
• Bad: Accounts provided as Fraud

A linear model was developed that could easily integrate with the client's systems that would allow the system to raise auto alerts for high risk cases.

Outcome:

The final output was presented as a Scorecard such that higher the score the more likely the account will turn fraud in the next three months. Using these scores, alerts are raised for most suspicious accounts and these cases are picked by RIC for further thorough scanning. Using this scorecard, RIC was able to increase and improve its fraud detections significantly.