Financial decisions
driven by data and AI

Smart scoring

We use a narrow AI that allows risk officers to build, test and validate advanced credit scoring models.

• 01

  Fraud prevention

  We have enhanced the credit scoring capabilities with the integration o newly developed AI-based fraud-scoring engine.

  The fraud-scoring engine analyzes data related to the applicant’s authenticity and detects atypical behavior or information anomalies by using biometrics like the time spent filling in the form.

  

• 02

  Application

  The Application Scoring engine developed by using a combined stack of machine learning algorithms, compares the client’s application data with other available databases.

  The system can retrain scoring models instantly using mini-batch optimization. This prevents overloading of models when new data arrives.

  

• 03

  Marketing

  A ML-based customer retention system automatically grants each customer a  personalized loan discount by analyzing numerous parameters.

  The propensity scoring assigns the borrower a special score that evaluates his propensity to return to the service and, based on this, grants the customer a discount.

  

• 04

  Collection

  The Collection Scoring is a system that is used when servicing non-performing loans and evaluates the probability of receiving payments from borrowers by analyzing big data sets.

  It enables us to create client segments, calculate strategy and activity efficiencies and automate the information exchange within the company and external parties.

  

• 05

  Document recognition

  Currently, general-purpose computer vision models are most often used for document recognition.

  Our computer vision model is based on an ensemble of several neural networks, each of which specializes in one type of object to allow processing documents much more accurately.

  

• 06

  Client service automation

  We use a self-learning chatbot that interacts with users and gives them information on financial products, individual recommendations, and financial prudence.

  Information is processed based on statistical matches covering a wide range of FAQs. Machine learning technology enables the chatbot’s number of answers to increase daily.