CTOS IDGuard Prevents RM300M+ Fraud Losses in Banking Sector

The national fraud bureau, CTOS IDGuard, has released its 36-month results today, with a total of RM319 million in potentially fraudulent applications being flagged by the system to alert banking consortium members since 2020.

CTOS IDGuard is part of CTOS’s suite of fraud and identification services which leverage on CTOS’s unique and extensive database, such as CTOS eKYC, CTOS Multi-face ID and CTOS Digital Footprint. To date, CTOS IDGuard has screened nearly 7 million applications from member banks, which cover over 65% of Malaysia’s banking assets.

With online fraud on the rise globally and the recent pandemic having accelerated online transactions, concerns are being felt by the expansion of digital solutions such as instant loans, real-time payments, pandemic-fueled digital onboarding and contactless services. Accelerated digitization has been accompanied by increased fraud and cyber risks, evidenced by Malaysia witnessing over 55,000 cybercrime cases with losses amounting to RM1.8 billion from 2021 until July 2023.

“Recently, CTOS IDGuard has rolled out several improvements, including improved fraud detection, increased efficiency in the fraud investigation process and additional intelligence from new sources. We have also deployed machine learning models which will bring increased precision and detection rates,” explained CTOS Digital Group CEO, Erick Hamburger. “These latest improvements are vital for the sector and our economy as digital finance continues to grow.”

CTOS IDGuard uses an industry-leading fraud and financial crime prevention engine by GBG. The system undergoes constant advancements to substantially increase the system’s effectiveness in fraud detection accuracy, lower false positive rates, and improve operational efficiency.

“In addition to the three machine learning models for credit card and mortgages within CTOS IDGuard, GBG has added two new models for auto and personal loan applications that have resulted in a 27% and 31% reduction in false positives, respectively. This frees up operational and manual review time for fraud analysts, allowing them to focus on larger fraud prevention challenges. With GBG’s ongoing machine learning model updates, we enable adaptive model training with data from new fraud patterns and investigation outcomes to improve performance and mitigate deterioration risk. This supplements existing rule-based systems to enhance fraud detection accuracy, which in turn reduces false positives and improves operational efficiency,” added Dev Dhiman, APAC Managing Director of GBG.

Over the past three years, IDGuard has been instrumental in screening applications for car loans, credit cards, home loans, personal loans and SME loans.

In terms of total applications, credit cards make up the significant majority, contributing two million of the total applications screened. Around half of the flagged applications are confirmed to be suspicious or fraudulent, and with the CTOS State of Consumer Credit Report 2022 showing that around two thirds of Malaysians have a credit card, the statistics demonstrate how important the national fraud bureau has become to the member banks.