Amazon releases detector for payment fraud

Amazon’s new payment fraud detector is now generally available.

On 3rd December 2019, Amazon Fraud Detector was showcased in preview mode as part of the 2019 re:Invent programme.

This is a fully managed service that helps webmasters to identify online fraud.

However, it is not exclusively limited to payment fraud – the tool also helps to identify the creation of fake accounts.

Various statistics suggest that every year, tens of billions of US dollars are lost to online fraud across every region of the world.

Those who are in the business of running an online store need to remain vigilant about payments that are made with stolen credit cards as well as the creation of fake accounts.

Thanks to machine learning, however, identifying fraudsters by using fraud detection apps is now a possibility.

Amazon Fraud Detector uses sophisticated fraud detection technology that is the result of Amazon’s 20 years’ worth of fraud detection expertise.

By implementing it, the tool automatically detects and alerts you of any potentially fraudulent online activity.

According to Amazon, creating a fraud detection model only takes a couple of clicks and the learning curve is simple even if you don’t have experience of machine learning.

In fact, the whole process can be summarised in five simple steps:

  • Determine the event that you would like to check for fraud.
  • Navigate to Amazon S3, upload your historical event dataset, and then proceed to select a fraud detection model type.
  • Amazon Fraud Detector will build a custom model based on your historical data.
  • Based on what the model predicts, you can then proceed to create rules to review, accept or collect more information.
  • Make an API call from your online application so that you can take action on fraud prevention in real time.

Detailed directions on how to implement it are available on the official AWS blog.