fraud detection


OCBC Bank’s fraud detection recovers SG$8 million in losses

Fraud detection solutions continue to grow in demand as more organizations look to implement them. While cybercriminals still manage to avoid them, most fraud detection solutions today are doing their part in detecting fraudulent activities a lot faster.

According to the 2020 Global Economic Crime and Fraud Survey by PwC, the last 24 months saw a staggering US$42 billion in losses to fraud globally. 47% of the organizations surveyed have experienced fraud in the past 24 months with the company experiencing six fraud incidents.

In banking, fraud activities continue to see an increase. Many victims keep falling for scams and online fraudsters that trick them into sharing out their banking details. These details and then used by fraudsters to gain access to their bank accounts.

As part of its anti-fraud program, OCBC Bank has deployed a Fraud Surveillance System (FSS). Developed together with SAS Institute, a leading software analytics company, the FSS uses a combination of flexible data and network analytics, as well as cutting-edge surveillance techniques, to monitor customer activities and transactions.

In Singapore, banking-related scams surged in the wake of the pandemic, leaving clients vulnerable to phishing and other means of fraud. OCBC Bank and SAS collaborated in a phased deployment of FSS as part of the Bank’s Operational Risk Management program to monitor customer activities and transactions, resulting in improved customer experience and operational outcomes.

(Photo by Roslan RAHMAN / AFP)

Fraud detection with AI and ML

As the first Singaporean bank to tap AI and machine learning to combat financial crime, OCBC Bank has long been at the forefront of technological innovation. It was also the first bank to launch its own AI unit in the country.

Fraud detection solutions today help with not just detection, but also investigation remediation and analysis of customer fraud exposures. Using multiple AI-based analytic techniques including predictive modeling, text mining, and network link analysis amongst others, the system can proactively identify known and unknown patterns.

The holistic capability of FSS enabled monitoring of customers’ transactions and alerts in real-time. Monitoring was done across multiple channels, including ATM, CASA, online banking, mobile banking, wire transfers, as well as corporate transactions. With FSS, synergies were realized across OCBC’s operations in multiple countries.

The anti-fraud program was able to recover approximately SG$8 million (US$5.95 million) worth of fraudulent transactions in the first year of production use and on Singapore activities alone. OCBC also reported that more than SG$10 million (US$7.5 million) was prevented from falling into the hands of fraudsters through its anti-fraud program in 2020.

The bank’s Group Operational Risk team has extended the real-time fraud surveillance coverage to international branches, including its Hong Kong subsidiary. The system will also be made available in OCBC’s branches in Southeast Asia such as Thailand and Vietnam.

Meanwhile in India, StashFin, a Singapore-based fintech company build its own models for fraud detection. It uses computer vision, image matching and facial detection to detect fraudsters among applications of its lending services. Customers are put through several ML filters to detect their risk level and are scored based on other data in their registry.

While there is no denying that fraudsters will continue to try to run their fraudulent activities, fraud detection tools will also only continue to improve and get better in their detection. With the technology also being automated and easier to manage, more companies can apply them in their systems to ensure they have added security features to their platforms.