Besides the constant threat of hacking, companies in industries such as banking and e-commerce also need to fend off scammers who are employing increasingly sophisticated methods to avoid detection. Germany’s Fraugster Ltd. argues that artificial intelligence is the answer.

The startup has created a self-learning fraud detection platform that attracted a $5 million investment today from a group of European investors. TechCrunch reports that Berlin-based early stage fund Earlybird led the round while Speedinvest, Seedcamp and an unnamed “large” family office from Switzerland contributed as well.

These investors join a long list of companies that have already been swayed by Fraugster’s pitch. Visa Inc. and several thousand other firms reportedly manage their fraud detection efforts with the help of its platform, which uses deep learning neural network algorithms to find signs of suspicious activity.

At a high level, Fraugster claims that its technology examines transactions in a way that mimics the workflow of human financial analysts. The startup’s software can pull business records from a company’s accounting systems and combine them with related information about user activity patterns to gain an understanding of their habits. From there, every new transaction that a person makes is checked against their history to identify inconsistencies that may indicate malicious intent.

A fraudulent action can be as straightforward as a buyer logging into their online banking portal from a new location and initiating a large transfer to an unknown account, or it may involve much more subtle warning signs. Either way, Fraugster claims that its platform makes it possible to determine whether a transaction is legitimate in as little as 15 milliseconds. The factors that go into each decision are displayed alongside the results to let financial analysts spot potential mistakes.

Fraugster promises to help companies cut fraud by up to 70 percent while reducing false positives that impede legislate clients. Moreover, the startup claims that its platform learns from every transaction, which means that detection accuracy keeps increasing over time. The system, the company says, can anticipate fraudulent transactions before they happen in some cases.

It’s certainly an appealing value proposition, but Fraugster isn’t the only provider that is harnessing artificial intelligence technology to combat fraud. The company is facing off against both fellow startups such as the Accel Partners-backed Shift Technology and Feedzai as well as numerous incumbent providers. Among them is IBM Corp., which last year acquired machine learning provider IRIS Analytics GmbH to strengthen its fraud detection offering.

Image via Pixabay