Human and artificial intelligence working together to fight payment fraud

In a rapidly evolving payments fraud landscape, it’s important that merchants have strong fraud systems and machine intelligence behind them, but that should never mean sacrificing the human eye


The prevalence of fraud has risen rapidly as criminals have sought to take advantage of a unique pandemic-induced combination of financial and health threats which have made people more vulnerable to scams. The National Cyber Security Centre, the UK’s cybersecurity agency, revealed last month that it has taken down more scams in the last year than in the previous three years combined, fuelled in particular by coronavirus and NHS-themed fraud attempts.

Though some may associate the rise of fraud with cybercriminals becoming more sophisticated, the reality is many old scams are among the most prevalent, with many people still falling for phishing attacks. Younger consumers have become well-versed on how to spot such scams, but the growth of digital activity among older generations, forced to shop online during the pandemic, has opened opportunities for fraudsters to target a far less tech-savvy demographic. A main entry point to scam consumers is when they are making a payment.

“Many consumers are starting to make payments online for the first time, going through the process of authorising a transaction via emails and messages. If they are not very cautious, they can be scammed into thinking similar-looking emails and text messages look legitimate.” says Tom Pilling, chief risk officer at Trust Payments, a global payments technology company. “It is the perfect environment for fraudsters to scam people who are new to purchasing online.

“We definitely saw a big rise in fraudsters scamming people that traditionally don’t purchase online. The modus operandi of a fraudster hasn’t necessarily changed in a significant way. The methods fraudsters use to get through fraud engines and fraud transaction monitoring solutions have certainly evolved and improved, but when you look at the transaction itself and the information held within it, it is still the case that if it doesn’t look right, it generally isn’t right.”

We continue to learn everyday with the volume of transactions we process online

The more consumers fall for these types of scams, giving away their personal details, the more it also affects business owners and shop owners, many of whom have also been forced to embrace ecommerce during the pandemic. Unable to trade from their physical stores, merchants who previously had no online presence suddenly had to create a website or web shop very quickly to survive, while click and collect options also grew significantly.

While the ecommerce industry has raced ahead, payments companies have also had to be careful to reignite some of their early education programmes for merchants which were experiencing selling on digital channels for the first time. This has particularly been the case for smaller companies, which have needed educating on what they should be looking out for in terms of typical signs of fraud and scams, including, for instance, cardholders making repeat purchases in very short periods of time, average transaction value, or velocity-type checks.

“These things are quite standard to a lot of big merchants that have their own fraud teams in place but they’re new to a lot of smaller merchants,” says Pilling. “When you give merchants the education they need, they gain that additional layer of confidence. They can always consult with us to get advice about particular transactions, but you also see their confidence grow as they learn to make some of those decisions themselves. If an order looks too good to be true, it should spark caution with merchants. That human instinct is just as important as high-tech fraud tools.” 

Merchants, and indeed payment processors and acquirers, face the difficult challenge of not only trying to protect consumers from fraud but also balancing that with the need to provide a strong customer experience without making the payment process too cumbersome. Trust Payments, which powers online payments for some of the world’s most well-established, as well as emerging, companies, has designed intelligent omnichannel payment solutions that monitor transactions for fraud while helping merchants grow, by ensuring the customer experience is seamless and convenient.

Trust Payments is dedicated to making as many decisions for merchants as possible with the smallest impact on merchant transaction authorisation versus decline. While committed to removing false positives, it is also focused on ensuring all the good transactions flow through without disruption. That means using real-time monitoring to look at different verticals, trends and types of transactions to determine what is good and what is bad. Those that sit in the middle, as potentially suspicious, will then be reviewed by a member of its fraud analyst team.

That balance is crucial. Though it’s important for acquirers to be supporting merchants with sophisticated, machine learning-powered transaction monitoring systems, it’s also vital that a human eye is maintained, both from fraud analysts working for the acquirer and also from the instincts of merchants, who often know the behaviours of their customers better than anybody.

“A lot of it is about learning,” says Pilling. “We continue to learn every day with the volume of transactions we process online. And with that we’re able to then also look at historical data to determine how we need to adjust our own rules so that we get that nice balance of good transactions, bad transactions and those transactions that sit in the middle that need further follow up and analyses. While our systems are really sophisticated from that respect, having that bit in the middle and the human side of things is key to the way we do our own analyses and educate merchants as to the type of transactions that they should be looking out for as well.

“There’s absolutely no doubt that ecommerce is a growth area, but we’re also supporting other emerging verticals. As well as our own in-house fraud solutions, we partner with different companies to provide a unique overall perspective of analysis. We look at those third parties that can do website analysis, for instance, our crypto blockchain analysis, to help create a really unique picture of transactions in the future. However, we must never forget that humans have that natural instinct, more so than machines, to know if it doesn’t look right, it probably isn’t.”

For more information, visit trustpayments.com


Sponsored by

Trust Payments