“Machine learning” is unquestionably a buzzword. But here at Sift Science, it’s also our bread and butter. So, when talking to potential customers, we often find ourselves demystifying the concept as a starting point.

Here are a few common questions we hear:

So, what is machine learning?

We’re surrounded by data every day – but data on its own can only tell you about what happened in the past. With machine learning, computers use specially created algorithms and mathematical formulas to learn from historical data with the goal of predicting likely future scenarios. Think of machine learning as the equivalent of a human learning from experience.

How is machine learning useful?

Basically, machine learning makes humans’ lives easier and more informed, so we can make smarter decisions more efficiently. The time it takes humans to read, synthesize, categorize, and evaluate data is significant — and machine learning streamlines much of that effort.

What’s a common use for machine learning?

Machine learning is popping up everywhere! One place you may have noticed it is in your inbox. How many spam messages do you notice on a daily basis? Probably very few. That’s because your email account uses machine learning to weed out fake emails, based on data and patterns.

If you didn’t have machine learning working to keep your inbox clear, you’d see lots more Viagra offers and “You’ve won!” messages.

How does machine learning apply to preventing fraud?

Well, we’re glad you asked. Thousands of signals can be red-flags for fraud, from IP address to the time an order is placed. Having a human review every new order or newly created account to scan for these signals – a process known as manual review — is extremely time-consuming. However, a trained machine learning system can learn from historical examples of fraud and adapt to constantly-evolving fraud patterns, predicting future fraudulent activity based on shared attributes or actions.

Curious to learn more? Check out our ebook, The Future of Fraud Fighting, to get a more detailed rundown of the potential of machine learning.