Johannesburg, 21 May 2025 – Experian, a global leader in data and technology, today hosted its annual Fraud Forum in collaboration with the Institute of Commercial Forensic Practitioners (ICFP). The forum highlighted key findings from Experian’s latest research report conducted by Forrester Consulting, based on a survey of 500 senior fraud leaders across eight countries, including South Africa. The report underscores the need for businesses to adopt Machine Learning (ML)-based, orchestrated fraud decisioning to combat escalating fraud losses and increasingly sophisticated attacks.
“Machine Learning (ML) is the top fraud priority for organisations in 2025,” says Bree Caprin, Experian South Africa’s Chief Customer Officer. “Our research reveals a growing concern about escalating fraud losses and the struggle to keep pace with increasingly sophisticated attacks. Businesses that fail to embrace ML-powered solutions risk falling further behind.”
The study found that 61% of Financial Services and 48% of Telcos expect their fraud losses to increase in the year ahead. A similar percentage (59%) admit that their organisations are struggling to keep up with the rapidly evolving fraud threat. This highlights a significant gap between current fraud prevention capabilities and the complex attacks being deployed.
The report outlines five areas where businesses should invest and innovate to effectively combat fraud in 2025:
ML is Critical, but Implementation is a Challenge: While 71% of respondents agree that ML is critical to staying ahead of the growing fraud threat, 53% admit they struggle to implement it effectively. Pre-trained and verified ML models offer an off-the-shelf solution that can greatly reduce the time to positive ROI, compared with developing a successful model in-house. Implementing ML fraud detection models enables businesses to automate more applications and transactions, significantly reducing the need for manual reviews and lowering associated investigation costs.
Explainability is Essential: Transparency in ML decision-making is crucial for building trust and ensuring compliance. Experian’s ML fraud models offer full transparency with reason codes that break down the value of each contributing feature for every decision. As regulations around the use of ML-based decisioning tighten, fraud leaders are increasingly focused on ensuring the explainability of the models they use.
Fraud Orchestration is a Must: While 76% of businesses use multiple fraud solutions from different vendors, over half (58%) do not connect these signals into a unified decision. Fraud orchestration platforms improve fraud detection rates, reduce false positives, enhance customer experience, and lower costs by dynamically calling fraud services based on risk levels. Two-thirds (66%) of organisations plan to increase the number of fraud solutions they use in the next twelve months, highlighting the need for orchestration to manage complexity.
Cloud Migration Offers Agility and Scalability: Cloud-based fraud detection offers numerous benefits, including automatic software updates, enhanced security, increased processing power, and elastic storage scalability. Cloud also acts as an enabler for ML by facilitating access to and connection with large amounts of data in a secure way. The plug-and-play functionality of cloud solutions allows organisations to achieve more accurate fraud detection quickly, no matter what their current IT capacity is.
Reduce False Positives to Improve Profitability: False positives remain a significant challenge, with 61% of respondents reporting that they cost more than actual fraud losses. Combining ML, device data, and behavioural data can significantly improve fraud detection accuracy, minimising misclassifications and reducing the volume of manual reviews. Inaccurate rules-based prevention systems often flag both good and bad customers, increasing manual reviews and impacting profitability.
“The key to effective fraud prevention lies in accuracy and efficiency. By leveraging the power of ML, device data, and behavioural insights, we can significantly improve fraud detection rates, ensuring legitimate customers experience minimal friction while fraudsters are stopped in their tracks. This also allows fraud investigators to focus on fine-tuning strategies and reducing operational costs,” concludes Caprin.