Using AI to Prevent Ecommerce Fraud and Secure Payments

As e-commerce continues to grow exponentially, fraud within these channels is also on the rise. In fact, credit reporting agency Experian reports that there’s been a whopping 56% increase in ecommerce breaches since 2016. The prevalence of digital payments and transactions, coupled with the number of consumers gravitating to apps and mobile connectivity, has led to criminals devising newer, more sophisticated methods of stealing money.

Fortunately, there is a solution. Artificial intelligence (AI) and machine learning are increasingly alleviating the fears of merchants, PSPs and ecommerce companies who are being plagued by an array of cyber attacks. In fact, these technologies have become critical tools in the fight against fraud in the continuously evolving payments and transactions industry.

Most anti-fraud systems that flag suspicious behavior (for example, unusual payments to remote suppliers, or credit card purchases that take place outside a customer’s country of residence) are “rules-based.”  This means that they detect fraud by measuring transactional activity against several pre-determined rules that humans have created by combining data about previous fraud with intuition about what constitutes “normal” buyer behavior. Although effective to some degree, this approach can be costly and slow, with high false positive rates and no way to identify new, emerging fraud patterns.

Machine learning, on the other hand, uses self-learning algorithms to integrate and analyze massive amounts of evolving, fast-moving and unstructured data.  These algorithms can detect fraud in real-time, learn from trends, automate tedious tasks, and effectively identify new fraud patterns.

While AI and machine learning are important developments in the fight against fraud, the role of humans in securing the omnichannel ecommerce space should not be underestimated. Machines can identify signs of fraudulent activity, but it’s up to analysts to act on them. This is especially important in today’s omnichannel retail environment, where chargebacks caused by fraudulent activity can have a negative impact on the touchpoints that connect buyers with sellers.

Today’s cyber-criminals know the ins and outs of payment processes and can easily locate vulnerabilities through distributed networks and the dark web.  They then employ multiple sophisticated tactics to exploit these vulnerabilities, including but not limited to identity theft, phishing and account takeover. According to Nielsen Report, fraudsters steal about 5.65 cents for every USD 100 spent!

As online fraud continues to evolve, machine learning is proving to be the most effective method that ecommerce constituents can use to protect themselves. Robust and user-friendly, it secures vulnerabilities by monitoring real-time customer behavior, and helps companies with better and more effective decision-making. From identity verification and payment authorization to checkout scoring and merchant underwriting, its applications are limitless. The underlining result is a significant reduction in fraud loss and chargebacks.

7 Tips to Prevent Chargebacks

Chargeback management – not quite the two words that invoke happy thoughts and feelings among merchants. There are many reasons chargebacks arise and, in some cases, it is not always something the merchant can control.  Examples of chargeback triggers include:

  1. Fraud
  2. Customer disputes
  3. Processing errors (e.g: duplicate transactions)
  4. Authorization issues (e.g: debiting cardholder)

It is particularly difficult for merchants to balance the prevention of chargebacks that are related to fraud – they are often afraid that velocity checks and fraud controls will affect legitimate consumers and prevent them from buying on the merchant’s website.

According to a recent CyberSource Fraud Report, chargebacks are a key, if not the key, indicator in managing fraud. The majority of survey respondents pinpointed chargeback rates as the most important KPI for fraud management, followed by other types of fraud rates:

  • Chargeback rate (62 percent)
  • Fraud rate by value (47 percent)
  • Confirmed fraud rate (35 percent)
  • Fraud rate—order volume (27 percent)

While it is unrealistic for merchants to think they can completely avoid chargebacks, they can feel empowered by the knowledge that there are definite measures they can take to better manage and, ultimately, reduce the number of chargebacks and their negative impact, such as financial damages that chargebacks have on their businesses.

Here are several tips I recommend merchants take to better manage their chargebacks:

1.   Pinpoint Reason Code: As mentioned above, there are several reasons for chargebacks. First and foremost, it is crucial merchants monitor their chargeback per reason code to pinpoint the problem.

2 .  Know Your Customer:  Yes, merchants continue to hear ad nauseum the importance of KYC, and I am here to reiterate this point for the umpteenth time. They must continue to take this extremely seriously and not rely solely on their acquirer or PSP to protect them from fraudulent customers. The more proactive merchants are, the higher their chances will be where they do not have to be reactive when something goes wrong.  If something feels ‘off’ with a customer, it is important for them to follow their gut feeling and react accordingly.

3.  Be proactive to prevent chargebacks: Prevention is the best defense. There are many things a merchant can do proactively: offering clear product descriptors, providing a customer service number, emailing consistent and accurate notifications to customers and generally maintaining good customer dialogue.

4.  Provide excellent customer service: Be informative and transparent regarding pricing, promotions, subscriptions, refunds, cancellation and shipping policies. Merchants should follow best practices, including processing and shipping orders before completing the transaction, tracking and recording shipping and/or service delivery, and last, but not least, communicating with customers regarding any questions or concerns.

5Employ security measures: Merchants must implement advanced security measures.  They have several choices: implementing 3-D Secure, which protects against chargebacks, and another option is an Address Verification System (AVS). It is best practice to both perform fraud checks on cardholders before processing orders and to check end-consumer behavioural patterns.

6. Do not automatically issue a refund if you received a chargeback: Pick the right battle and fight back if you have compelling evidence, such as communications, receipts, etc.

7. Work with the right provider (PSP or acquirer): If you received a chargeback that you believe should be represented you should expect a proactive and consultative approach towards chargeback from your provider.

A good provider should:

  • Handle and represent your chargebacks
  • Work with you on the chargeback presentment results/count and improve stats
  • Pinpoint chargebacks due to technical reasons, and represent them automatically

I hope this has been helpful. Credorax representatives are available to explain more about our chargeback management and representation services 24/7 at