Insurance Fraud Detection: When You Shouldn’t Plug and Play

In my experience, the Plug-and-Play fraud detection approach doesn’t always work. Yes, insuring can be viewed similarly on a global scale. The principal is the same. You make an agreement with your customer, the insured. The customer pays a premium and you, as an insurance company, take care of the risk and help your customer whenever something goes wrong - providing it meets the policy terms and it’s not a fraudulent claim.
But you can’t take the same approach in the detection of fraud in Europe compared to Latin America. You can’t use the same ruleset in Japan as you would in Switzerland. The local law and legislation, the cultural differences and the approach of an insurance company towards its customers are just too different.
The differing approaches come to light when discussing and examining claims handling and fraud prevention processes. In my role as a Senior Counter Fraud Expert, I make it a priority to familiarize myself with these differences. You can design, develop and build the greatest tools, but if you are not able to translate them to the local situation, they will not work. Therefore, as a solution provider, creating a team with a wide range of expertise among staff members is very important. People with an IT background will obviously be required due to our digitized world, but practitioners from the insurance business are also vital to a successful provider team. Together this is a good mix to push development and results forward.
One Size Does Not Always Fit All
The local situation in each country is not something you can learn from the Internet. You really need to go out there: listen, ask questions, observe, make mistakes and retry. I’ve been in many countries to do just that: I sit together with our insurance customers and their SIU’s and discuss the local situation. Only then is it possible to learn that a certain approach, perfectly fine in country A, just does not work in country B.
I’ve also learned that every company has its challenges in converting fraud detection hits into savings. This is partly due to the local situation. To give an example: In some countries, witness statements are not taken after traffic accidents as this approach does not fit into the culture. This difference must be reflected in the fraud prevention system. It doesn’t make sense to use rules and patterns that decide on witness information in a traffic accident. It would generate false positives and create results the insurance company can’t work with.
In another country, the insurance company does not contact the insured directly when the customer is insured through a broker. This impacts the way the insurance company will investigate potential fraudulent cases. This characteristic also must be translated into the fraud prevention system to determine which cases should be forwarded to the SIU.
The Use of External Data is Not Always Necessary
Another topic is the use of data from external databases, which can prove to be very helpful. But I have seen situations where this external data is used with every claim. The data can come from several external sources. Chances are high the insurance companies must pay for this external data per use. This should and could be more efficiently implemented. Depending on the claim, insurance companies potentially have enough data inhouse to assess the risk. Systems can be pre-configured to decide when it makes sense to consult an external database and when not. In many claims it becomes clear, based on the information available within the insurance company, whether you are dealing with a good customer that can be fast-tracked. Do you really want to spend the time and money on additional sources? Furthermore, due to privacy laws and the availability of external data, accessing these databases may become an issue, so it is therefore important to understand data regulations in the country you are operating in.
Customization is Key
Nobody wants to start implementing a fraud prevention solution from scratch. This approach would be way too time consuming and expensive. Therefore, we deliver a standard fraud prevention solution, coming with best practice rules and patterns.
However, for a successful fraud detection strategy, it is vital to understand the customer, the local law and legislation as well as the local cultural aspects. Once we know and understand, we can adapt our rulesets and pattern recognition and provide advice on how to work with these results.
Therefore, I take a hands-on approach with my customers. We compare the technical results with the common investigation knowledge and experience of each insurance company and adjust the decision logic in our fraud prevention solution. Adjusting the solution is easy and fast as our solution uses a white box approach which is transparent in the decision making and setup. We even teach our customers to make adjustments and additions themselves if they prefer to work alone.
We bring our 95% ready solution and give you access to it. Together with you we will tailor the solution to address your local situation and meet your specific needs. We will be your partner. We will be there at go-live and work on the results together.
This all comes down to teamwork. Not a vendor-customer-relationship. Teamwork. Partners. We achieve the best results together.