Nordic Insurance Company Gjensidige had a paper form for car injury claims that they wanted to be digitised. When we started looking into the process, we found that it had a lot of questions and felt cumbersome for the users. Were really all the questions necessary for all claims?
Digitizing existing processes is not about moving an existing process from paper onto a screen. So when starting a project, we always start asking why. More often than not, the answer when we ask why things are the way they are is that they have always been that way.
When building Gjensidige's new car claims app, we started out looking for any unnecessary questions from the form. We started to review how the answers to the questions in the paper form were used in the organisation - in what cases did a question really need to be answered? What we found surprised everyone - a lot of the data gathered weren't critical anywhere in the organisation! In addition, we found that a lot of cases would always lead to an approval - seems like a candidate for automation!
When listening in to actual customers calling the Gjensidige call centre, we noticed that the call centre representatives did not perform an "interrogation" of the customer. What they did we found was that they responded to the customers' concerns: will I be able to make it to my appointment? Will the insurance cover all the costs? Can I just leave my keys inside the car?
We turned this learning into a design principle in the app we built: help the customer with his or her priorities.
Because of these insights we were able to peel away a lot of the questions and provide a seamless experience for Gjensidige's customers. One unexpected outcome of the app was that the system was able to automatically approve over 80% of the claims using machine learning. This led to a world record for the fastest insurance claims processing with 1.3 seconds processing time. The Gjensidige app is currently used by more than 500.000 customers.