Tackling International Revenue Sharing Fraud (IRSF) With ML
International Revenue Sharing Fraud is a major challenge for communication service providers (CSPs). The Risk and Assurance Group 2020 Survey estimated that this call forwarding fraud costs CSPs up to RM5.63bil annually, presenting a significant area of revenue leakage for many operators.
IRSF operates by fraudsters illegally obtaining mobile sim cards, then setting up automatic call forward features that redirect calls to premium rate numbers overseas. It is one of six key telecommunications fraud types operators must address, and presented a critical challenge for one major Middle Eastern CSP who relied on Neural Technologies’ advanced machine learning solutions to tackle it.
Identifying IRSF with advanced machine learning
An operating IRSF fraud system can be quickly established by experienced fraudsters, meaning rapidly identifying and tackling these threats is key.
Once an IRSF sim card system is set up, fraudsters use auto dialling services to quickly place a high volume of calls, incurring significant billing costs in a short period of time. CSPs find themselves exposed to high losses from the calls to a premium rate number, while criminals benefit from an illicit agreement to share the profits with the owner of the premium rate number.
Neural Technologies’ machine learning solutions offer a unique advantage in tackling this type of fraud. Where static fraud detection systems rely on established patterns of behavior in order to identify potentially high-risk events, the adaptive and automated solution from Neural Technologies is designed to identify new and emerging fraud types, quickly eliminating the fraud threat or flagging it to analysts as and when anomalous behavior occurs.
In the case of the Middle Eastern CSP, Neural Technologies’ Optimus Platform detected an emerging trend where subscribers within a certain country with no previous risk history had sent an SMS followed by a voice call to an international number.
This spike in activity to overseas numbers didn’t match previous traffic profiles, triggering a risk identification by the Optimus Machine Learning tools, quickly highlighting to analysts that the activity likely corresponded to high-risk IRSF. The CSP was able to quickly investigate, and block a range of numbers before losses accumulated and customer satisfaction was impacted.
Eliminating the threat of fraud
Optimus Revenue Protection solutions leverage sophisticated machine learning techniques designed to quickly identify these fraud types. In the case of this operator, a rapid spike in call events and duration of calls to overseas destinations prevented significant revenue leakage for the major Middle Eastern CSP.
Analysis by the Communications Fraud Control Association (CFCA) reveals some telecom carriers believe that as much as 18% of their revenue may be traffic generated by organised crime. That means fraud not only negatively impacts revenue and customer reputation, but can also be used to fund wider criminal activities.
Advanced machine learning solutions provide a powerful platform to tackle this threat, identifying and eliminating fraud in real-time or near-real-time, and preventing CSP exposure to significant losses through activities such as International Revenue Sharing Fraud.
Tackling fraud must be a priority for CSPs to reduce revenue leakage and business risk. Get in touch with Neural Technologies to discuss how our machine learning Revenue Protection solutions can help reduce your exposure to fraud.
