Fraud Management Tool

Fraud Management Tool is an application based on neural networks, the most advanced technology for pattern recognition. Our process analysis is made by specialized resources with risk and antifraud certification and training.

Areas of functionality:

• Application fraud detection for retail lending module
• Transactional fraud detection module

Application Fraud Module

Uses advanced database analysis with learning capabilities to detect inconsistencies in the data provided by a loan applicant. The application is based on neural networks technology, which represents complex systems of simple processing elements, highly connected and working in parallel, with advanced pattern recognition capabilities. It is an indiscriminative and non-subjective method of fraud detection because it discovers relationships between variables through learning, not through validation of hypothesis.

On the basis of the neural technology determined coefficients, applicants with false documents or not intending to reimburse the loan will be accurately profiled, and a pattern for “good” applicants will be modeled. Fraud detection is made in real time with a very high percentage of fraudulent applications being detected in real time by neural scoring.

How it works:

• Calculates risk scores for each application Fraud scoring calculates the risk for a client to be fraudulent and is complementary to application scoring (indicative of reimbursement capacity – not willingness)
• Prioritizes the suspicious cases by degree of risk

Main benefits

• Minimizes fraud, bad debt and write-off loss
• Maintains the “fast, effective and simple” standards for credit requests processing
• Reduces operational costs for fraud cases verification
• Continuously optimizes fraud detection
• Refocuses bank’s employees on sales and promotion of credit

Transactional fraud detection module

Aiming to detect 100% frauds, our team develops a specific highly accurate scorecard for each bank, using a sample of transactions, both valid and fraudulent, from the bank’s database. The transactional fraud module detects in real time the transactions with a high degree of risk and triggers a configurable set of actions for effective fraud combating.

The fraud module gives a score for each card operation, classifying the transactions in 3 zones:

• white zone (no fraud)
• grey zone (to be further investigated)
• black zone (fraud)

The implementation of the fraud module leads to a precise detection of fraudulent transactions, minimizing the value of losses and assuring the security of transactions for the bank’s clients.