Financial crime and fraud are recurrent problems for banks and intermediaries, causing losses worth billions every year – and these sums are only increasing. It is predicted that the global cost of fraud by 2027 will be USD $40.62 billion dollars, 25 percent higher than the fraud losses in 2020. The rapid adoption of digital payments – including those on real-time rails – also challenges banks to keep up with evolving threat types. Social engineering scams that target impressionable account holders, for example, are steadily rising. Identifying and preventing fraud at the right time is imperative for all financial institutions; here are four important things to consider when tackling fraud in real time.
1. Creating a smooth user experience
Payments modernization presents a huge opportunity for banks in terms of their business model and increasing profit margins, but the rapid adoption of real-time payment technologies poses a simultaneous threat to consumer privacy and data security. Payments that are made on real-time rails in place of cash create more opportunities for profit for banks. But as consumers begin to leave crumbs of their digital identity over the internet when making digital payments, or use public WiFi while entering personal data online, they leave themselves vulnerable to attack. With increasing demands from consumers for convenience, fraud management solutions need to be capable of identifying risk signals before fraud takes place – without being detrimental to the user experience.
2. Keeping pace with new methods of fraud
Social engineering scams have been steadily rising since the uptick in real-time payments spurred by pandemic-related measures. These scams occur when the victim is unwittingly tricked into transferring funds into a fraudster’s account as they pose as a legitimate payee. This could be a fraudster convincing a homeowner that they are indebted to their energy company, or persuading an employee that they are their boss and asking them to make a large corporate transaction. Due to current regulations, it is unlikely in either scenario that the account holder will successfully recover funds. Given that real-time payments are settled instantly, if they are not stopped in real time, the money is lost.
3. Focusing less on fraud and more on growth
This is where machine learning steps in. It is an indispensable weapon in the fight against fraud. In an increasingly real-time world, where fraudsters are becoming ever more sophisticated in their methods, banks and intermediaries need effective, reactive decision-making from their fraud management tools. Machine learning can limit false positives by allowing through good transactions while denying the fraudulent ones — with the most reliable information in real time.
In particular, incremental learning technology – an industry-first approach to machine learning – is a formidable contender for fraudsters. While traditional machine learning models need to be “re-trained” as fraud patterns change, models using incremental learning make small adjustments on an ongoing basis, allowing the model to adapt itself in production when new behaviors are observed. Banks and intermediaries should see less strain on fraud analysts and increased operational efficiency despite the greater number of transactions on real-time rails. This creates a smoother experience for customers, for example by eliminating instances of entering their PIN unnecessarily and reducing fraud alerts for their own transactions. Banks can therefore increase focus on core competencies, generating more value for customers and strengthening trust.
4. Saving fraud losses through collaboration
Machine learning insights are only as strong as the data received. Systems must be supplied with the widest possible view of risk in real time, and today that requires access to internal and external data – essentially enabling machine learning to distribute, exchange and consume risk signals. Network intelligence, a proprietary technology in ACI Fraud Management, allows banks, processors, acquirers and networks to securely share and consume industry-wide fraud signals to feed their machine learning models alongside proprietary data.
This collaborative approach enables financial institutions to complement their machine learning with signals exchanged within the community and from third-party fraud intelligence sources. This data could detect, for example, when a customer sends funds to an unusually dormant bank account. As real-time transactions are irreversible, cross-industry collaboration can help prevent funds being lost through social engineering scams by holding the transaction on the receiving end.
Download our eBook, Strategies for Fighting Fraud in a Real-Time World, which discusses the essential components of an effective fraud management strategy, including:
- Adaptive machine learning
- Network intelligence approach
- Real-time payments fraud screening
- Integrated payment engines