Every bank’s nightmare-ish dream: real-time transaction monitoring

real time

In an increasingly globalized and automated world, the demand for real-time payment processing or the ability to handle instant payments—is growing ever larger. Many clearinghouses have already set up real-time payment schemes. But just because real-time transaction processing is possible doesn’t mean that banks are ready to invest in it. After all, there are many risks to this type of payment processing. What are banks’ major concerns and how can these be alleviated?

Less control

One of banks’ main concerns is a loss of control. This is especially worrisome because many banks are barely able to control their compliance now, without having to be able to process payments in real-time. If payment batches were to be processed immediately instead of overnight or even within 3 to 5 business days, there would be very little time to check those payments thoroughly to ensure they conform with compliance rules. Banks have a hard enough time keeping up with alerts as is. Imagine how difficult this would be if it needed to occur within (milli-)seconds.

When the UK implemented its faster payments scheme in 2008, the number of financial crimes tripled within the first two years after implementation. These were mainly crimes in which criminals misled clients into making instant, irreversible payments to non-existing companies. Even if banks can’t be held accountable for this type of crime, it still can cost them their reputation.

Making money-laundering easier

Real-time payment processing also makes it more difficult to detect money-laundering payments. To launder money, criminals want a payment to move as quickly as possible through a number of accounts at different international banks. This is done to obscure the trail of where the money has been and is going. There’s no faster method of moving money than using real-time payment processing, in which a payment arrives at its destination in a matter of seconds. Currently, real-time payment processing isn’t available for cross-border transactions, but it’s still possible to thoroughly obscure the path a payment has taken domestically before it’s ever offered to the international payment network.

Why offer real-time payment processing?

Despite the issues that can arise with real-time payment processing, it need not be entirely avoided. Banks that offer this service will gain a competitive advantage over banks that don’t provide it. Clients want their payments to be processed quickly because for them it increases efficiency, transparency, convenience, and financial control. For small and medium-sized companies, this form of payment processing helps alleviate liquidity stress. And, in general, people have grown accustomed to things moving fast, so they have little patience and understanding when payment processing is slow.

How to be compliant while offering real-time payment processing

Any bank looking to invest in real-time payment processing solutions will also need to adjust their compliance systems to handle this task. Their transaction monitoring systems will need to be equipped for real-time alert processing. Rule-based approaches have become outdated, not only for this type of advanced payment processing, but also for traditional payment processing. The conventional systems still in use at the majority of banks can’t handle the volume of payments that are processed every day, the regulations that keep on changing and evolving, or the speed of payment processing as is. If real-time payment processing is to be added to the equation, conventional systems will clearly no longer be adequate. By turning instead to an advanced transaction monitoring system, your bank will be able to achieve a number of tangible objectives — objectives that are extremely important in order for a bank to be able offer real-time payment processing safely and securely. These objectives are: to reduce workload and to increase speed and accuracy.

Reduce the number of false positive alerts

At the core of mitigating compliance risk lies reducing the number of false positive alerts being generated. The more false positive alerts there are, the weaker the compliance agent’s ability to oversee the full system. When every alert seems important, none really stand out, and therefore all alerts become unimportant. With real-time transaction monitoring, it will be even more critical to reduce the volume of false positives since this will give compliance officers more time to spend on rapid response to the true positives.

Using machine learning is the best way to reduce the frequency of false positive alerts. Machine learning is the technology that allows computers to learn from feedback they receive from humans. By learning from compliance officers’ knowledge and experience, such smart systems learn to distinguish the characteristics of actual positive alerts from those of false positive alerts. This results in a reduction of the number of alerts for transactions that might appear suspicious but are actually benign and thus a decrease in the overall number of alerts.

Machine learning is also valuable for real-time transaction monitoring because it allows for enrichment of the signals generated by the monitoring. These signals are enriched with behavioral information and network information, along with information from other sources. The system can cross-check with various additional sources such as solved cases, sanctions lists, and big data. By drawing upon so much data, the system becomes able to recognize patterns that would otherwise remain undetected, making the alerts it generates more accurate and harder to miss.

User-friendly screens

Besides reducing the frequency with which false positive alerts are generated, it’s imperative to have a system that displays all relevant information about true positive alerts on a single screen. If a compliance officer can see a payment’s entire context in a single view, they’ll be able to handle that alert much more quickly than they would if they had to open multiple screens to do so.

In addition, there are always alerts which cannot be immediately categorized as true or false positives. These alerts require assessment by humans, and it’s best if this can be accomplished in an extremely user-friendly way. The interactive assessment of these alerts should also be used to train the system how to handle similar alerts in the future. This will increase the number of alerts that can be handled automatically. In this way, the system will become more and more efficient, while also building upon the expertise of the compliance agent.

Since real-time payment processing is all about speed, it’s also vital to monitor transactions as quickly as possible. If false positive alerts can be eliminated, the bank will be able to process the majority of payments in real-time. The result: delighted customers, compliance officers who are less stressed, and lower compliance risks for banks.

About BusinessForensics