Intelligent Ways to Manage Fraud

32

Fraud continues to impact online merchants. Some experience chargebacks, while others experience various forms of return fraud. Fraud hurts not only merchants but also consumers, as merchants often increase prices to account for fraud. Some merchants even budget for fraud as a cost of doing business.

Thankfully there are multiple new methods of detecting and preventing fraud that rely on artificial intelligence, which is the simulation of human thought by computers. This includes learning, reasoning, and self-correction.

Fraud detection and prevention is an ideal use case for artificial intelligence, for the following reasons.

Extensive real-time data analysis. Multiple data points — real-time and historical — need to be considered before classifying a transaction as fraud. An artificially intelligent system can do this, with its greater computing power and ability to quickly analyze large volumes of data.

Continuously evolving fraud patterns. A system that cannot learn and identify new patterns of fraud will fail to detect new fraud patterns. Many artificially-intelligent fraud systems are built with self-learning features.

Quick turnaround. Most fraud checks must be done instantly, before the transaction is completed. Fraud vendors offering artificially intelligent systems have met this key requirement by supporting transaction times in the low milliseconds. The combination of access to fast computing resources and optimized algorithms make quick turnaround possible.

Vendor Options

Quite a few vendors now offer products based on this technology. Here are a few of them. There are many more.

  • MasterCard Decision Intelligence uses artificial intelligence to increase the accuracy of real-time approvals of MasterCard transactions and reduce false declines. To use this offering, a merchant needs to only accept MasterCard.
  • Fraugster, a startup, has developed a self-learning fraud detection engine. This technology comes with APIs that collect customer data points — such as a customer’s profile and transaction patterns — and combines those points with information from other sources. All of this data is then analyzed in real-time to return a fraud score in less than 15 milliseconds.
  • Feedzai is another startup that detects and prevents fraud in real-time by processing data from different sources. It offers capabilities beyond fraud management by also helping merchants detect policy violations — such as employees not complying with corporate policies — and preventing blacklisted users from transacting on the site. Feedzai can be deployed in-house, versus the cloud, based on the preferences of the retailer.
  • Riskified provides end-to-end fraud detection for ecommerce merchants. It uses machine-learning models to protect against fraud and avoid costly chargebacks. Multiple global brands use this complete solution.
  • Signifyd is a popular fraud platform that offers a 100 percent guarantee against fraud. It utilizes machine learning and works with leading commerce platforms, such as Shopify, Magento, Demandware, and BigCommerce. Signifyd claims over 5,000 sites use its service.
  • Unfraud is a startup that offers an artificially intelligent fraud solution on the Shopify platform with the key features of self-learning and real-time detection. Merchants not on the Shopify platform can still use Unfraud via its APIs.

How to Choose

To select the right solution for your business, consider the following factors.

  • Ease of integration with your commerce platform.
  • Uses the fraud guidelines for your business and is learning and optimizing continuously.
  • Fast fraud detection, ideally under 100 milliseconds.
  • Extensive analytics.
  • Works in all the regions where your company operates.
  • Complies with local laws and regulations.
  • Handles peak loads and scales with your business.
  • Always available.
  • Highly secure.
  • Affordable.

Does your company use an artificially-intelligent fraud detection platform? Please share your experiences in the comments, below.