Overview

SberBank Anti-Fraud System is a platform aimed at countering cyber fraud. All customer service channels are connected to a single anti-fraud platform and evaluate their operations in one multichannel anti-fraud system (transaction scoring). Currently, all customer service channels are connected to this platform, which enables having a risk profile of customer behavior in all the channels. One of the most important achievements of SberBank Anti-Fraud System is the best performance indicators for detecting cyber fraud.

Key features
• 100+ data sources are used to monitor 4.6 billion financial transactions monthly.
• More than 100 models & algorithms are used to analyze information obtained from the data sources.
• In-memory data enables checking more than 450 million events daily, with a unique short response time of 1 ms.
• BigSecurityData (a separate analytical system) is used to enrich the fraud monitoring system models with data from internal and external sources.
• Machine learning models are used to detect anomalies and transactions scoring.
• Collaboration with the 5 largest telecom operators enables real-time exchanging information on risk events such as virus infection, change of SIM cards, change of owner, etc.
• Feedback form on the website allows clients to report suspicious phone numbers and websites.

Recognition
• In 2020, VISA and Mastercard confirmed the efficiency of SBER in risk management and prevention of acquiring and emission fraud.
• According to VISA, the relative indicators of the ratio of the amount of fraud to the turnover of SBER are 10 times higher than the average indicators in Russia, 15 times higher than the indicators of CEMEA, and 46 times higher than the world indicators.
• In 2020, the audit by KMPG showed high maturity of the anti-fraud processes (4.1 points, higher than the average of 4.0 of Fortune 500 banks).
• Patents were obtained for 5 models used in the platform (1 model was included in the top 100 patents for 2019-2020 in Russia)