Fighting payment fraud

Ralf Gladis at Computop advises that retailers should deploy AI tools to fight back against payment fraud

Fraud is a major problem for retailers, and one that they are publicly trying to address. A recent survey found that returns fraud was regarded by 64% of UK brands as the trend most impacting their company.  However, it is just one among many, with e-commerce brands facing promo abuse, chargeback fraud, account takeover and triangulation fraud all too frequently.

Where once retailers were reluctant to discuss fraud openly, it has now become a pressing issue that requires the collective actions of the entire industry to combat and particularly when it relates to payments.

Criminals have taken advantage of the expansion in e-commerce and combined this with the basic digital skills needed to commit frauds such as those carried out during card-not-present transactions, and even sharing tips with each other on social media sites. 

The notion that retail entities can easily absorb the costs of payment fraud has only exacerbated the problem and all too often, retailers lack the resources to log and report every fraudulent transaction. Criminals know they can get away with it and to make their attempts more successful they add complexity into the mix.

The impact is huge. In April this year, research showed the global retail sector had lost $429 billion to payments fraud in 2023. Businesses that predicted 100% growth in revenues also lost the highest amount to fraud in the past year, and it found that over a third (35%) of consumers had become a victim of payments fraud. Despite this, only two-thirds said they had effective fraud prevention systems in place. 

In the physical world, retailers are adopting innovative methods to combat payment fraud. Carrefour, the French supermarket is trialling a system that allows customers to pay using the palm of their hand. This is a technique already in use at Amazon’s Whole Food shops in the U.S. but it’s not something that can be as easily deployed for online sales. 

Preventing payment fraud

Fortunately, retailers are not alone in their determination to tackle the payment fraud challenges that threaten not only their financial security but also customer trust and satisfaction. 

The best approach they can take is to assess the cutting-edge technologies that are now hitting the market, often courtesy of their payment service provider.

A good example is an AI fraud score, which uses machine learning to provide an automated indication of risk. The score is a self-learning solution that examines each payment transaction in isolation and evaluates the relationships between various data points across transactions.

An AI fraud score takes a holistic approach so it can uncover even the most complex fraud patterns that might elude human detection or simpler automated systems. By continuously learning from new data, it can adapt and improve its own accuracy over time, leading to steadily decreasing fraud rates.

Enhancing customer experience and compliance

One of the standout features of an AI fraud score is its ability to ensure higher sales and improved customer satisfaction by reducing false positives—these are the instances where legitimate transactions are mistakenly flagged as fraudulent.

Traditional blacklisting methods, which often rely on blocking certain IP addresses or BINs (Bank Identification Numbers), can also lead to genuine customers being unable to make purchases, leading to lost sales for retailers.

Using an AI fraud score avoids these exclusions by focusing on nuanced transaction patterns rather than broad, potentially discriminatory, criteria. This also ensures compliance with EU geo-blocking regulation which prohibits unjustified geographical discrimination in e-commerce and applies to any UK retailer selling in the EU. 

Customisation and flexibility

In many cases retailers can customise the AI fraud score provided by their PSP to align with their individual risk tolerance and customer demographics. This flexibility means that the solution is not a one-size-fits-all model but instead provides a more tailored approach that meets the unique needs of each retailer’s business.

A retailer dealing with high-risk transactions can adjust the tool to be more stringent, while another might set it to be more lenient. Control is entirely in the hands of the retailer. 

Risk management integration

There are a range of other risk management tools that retailers can utilise alongside, and in conjunction with, the AI Fraud Score. These include static paygate risk assessments that have proven effective in the past or used to support advanced modules that limit transaction volumes within specified time frames.

It’s also important that all risk management tools work with 3D Secure 2.0, the security protocol that shifts liability for fraudulent card transactions to the issuing bank and away from the retailer. 

Payment fraud is too big a problem to be dealt with overnight. Ironically, however, the solution to what is essentially a technology challenge, also lies in technology. Combining the power of machine learning and AI with customisable, adaptable features has the power to detect fraud and enhance security, lowering the risk for retailers and boosting customer satisfaction.

By integrating an AI fraud score and other preventative tools into their fraud prevention strategies, retailers can begin to win this battle and protect their online businesses for the future. 

Ralf Gladis is CEO at Computop

Main image courtesy of iStockPhoto.com and coldsnowstorm