Fraud is on the increase, with a massive 76% of South African respondents to Experian’s survey for its Defeating Fraud Report having seen an increase in fraud losses over the past year.

The report surveyed 308 fraud managers in financial services, telecom, and retail organisations across 10 countries – including South Africa – to gain insights into their perspectives on fraud prevention measures. The research delves into the alarming increase in fraud losses that continues to impact many organisations’ bottom line and provides valuable insights on how businesses can effectively counteract this rising threat.

The findings reveal that the volume of fraud attacks has gone up in almost every category, with account takeover attacks showing the highest growth for Financial Services and Telcos. This is closely followed by synthetic identity theft attacks and bot or credential stuffing attacks. In the e-commerce sector, synthetic identity attacks have grown the most for merchants followed by friendly fraud attacks.

In the South African context, the research reveals that the two biggest challenges limiting businesses’ ability to prevent fraud are an inability to align fraud prevention and revenue growth strategies (59%); and the increased costs associated with using multiple types of fraud prevention software (59%).

“As businesses expand their fraud prevention capabilities with multiple fraud services, connecting them via a single platform and Application Programming Interface (API) is crucial,” says Mark Wells, chief customer officer at Experian Africa. “This capability was highlighted in the research where ‘reducing the number of platforms across the business’ (65%) was shown to be the top fraud-related priority.”

More than half (56%) of the SA respondents believe that the future of fraud prevention will be driven by artificial intelligence (AI)- and machine learning (ML)-powered solutions.

This is important, considering that 61% of businesses find that false positives cost their business more than fraud losses. Additionally, the research shows that the most important factor when considering AI/ML-based fraud prevention – for 85% of respondents – is the ability to automate more decisions for legitimate customers.

The main benefits of using ML fraud solutions are an increase in acceptance rates, reduced losses through greater fraud detection accuracy, and a decrease in the volume of manual reviews and false positives.

“Our survey results highlight how ML has become essential for fraud prevention as it allows for improved detection of both fraudsters and legitimate customers,” says Wells. “ML also provides the analytical backbone for behavioural biometrics and device fingerprinting which enables businesses to continuously and passively monitor customers without impacting the user experience – this is the key to balancing revenue growth with adequate fraud prevention.”