Business-Blog | adesso insurance solutions

Fighting insurance fraud with AI

Written by Melanie Hoppen | 19.08.2020

Machine learning, pattern recognition, and a treasure trove of data help insurers uncover fraudulent claim notifications and irregularities. AI against criminals: A topic that is becoming increasingly important for apps as well.

Insurance fraud is as old as the insurance industry itself. Thomas Mann had already reported irregularities in his novel "Buddenbrooks" as early as 1901. A lot has changed since then, but fraud is still around. State-of-the-art technology helps companies quickly make the right decisions in retail.

The customer is in a hurry, and the insurer has little time

In the event of a claim, the insured persons are understandably in a hurry. An insurance company sees a small dent on a car as a trivial matter, whereas it means agitation and uncertainty for the customer. The customer wishes to know that their claim is being settled as fast as possible.

Insurers in the property, accident, third party, and auto insurance sector are also in a hurry. Ultimately, the business in question is a retailer that is also characterized by its low margins. This is why the level of automation in this division is also accordingly advanced in most companies.

However, retail and automation also run the risk of having fraudulent schemes fall through the cracks. In this case, systems with "artificial intelligence" assist the clerks.

AI detects fraudulent patterns and uncovers anomalies

With the German Insurance Association's reference and information system (HIS) plus the GDV's credit agency, insurers have two important data sources at their disposal, which provide information about policyholders that have become suspicious. As such, querying this data pool is an important instrument for clerks when processing claims. But that alone is not enough.

This is because viewing the history does not help with discovering "new tricks", i.e. patterns or scam networks. And this is precisely where the AI systems come in, as already successfully used by the DEVK in automotive insurance, for instance. At DEVK, the clerks who handle the entire process from the initial claim report to settlement are assisted by AI. And it does so for 160,000 claims per month.

Protection in (more than just) apps - Behavioral Biometrics

However, dishonest policyholders are not the only ones with fraudulent intentions. In the age of digitization, credit institutes and insurance companies are making their services increasingly available online and via an app. The data and information saved in a user account are a worthwhile prize for criminals. AI also makes use of its advantages when it comes to fraud protection. To fight against account takeovers or the opening of falsified customer accounts, the market offers protection systems that use an unmistakable customer characteristic: their behavior.

Behavioral biometrics compile a profile of the authorized user from various parameters and can monitor the use of apps or online accesses in real time. The AI, operating in the background, compares public behavior to differences with previously gathered data. Screen typing and swipe gestures alone already set individuals apart, such as the pressure or the angle with which the finger touches the display. AI systems can also uncover anomalies here with greater reliability and speed than any human, and they can enact countermeasures.

 

Would you like to learn more about our AI solutions for fraud identification? Then contact our expert Karsten Schmitt directly.