Munich New technological possibilities will significantly change the business of insurers in many areas in the coming years. Through artificial intelligence (AI) and the use of new forms of communication, many things will be easier, faster and more individualized in the future, from the conclusion of the contract to the processing of claims.
“Like at the airport or in the supermarket, where employees support customers to digitally check in themselves or to scan goods, insurers should help to complete simple products online and show how easy it is,” says Jan Meessen, Industry Leader for insurers at Google in Hamburg However, the industry also comes up against natural limits.
For many years, insurers followed new trends with a delay compared to other sectors. The main reason was the solid profits, which only made changes necessary to a small extent. That is now changing massively.
Everywhere, innovative teams are busy with the possibilities that arise from the insurers’ huge treasure trove of data. After all, providers have known for some time how their customers drive their cars, how their homes are equipped and what their health is like. By processing huge amounts of data – called Big Data – the industry wants to be able to make personalized product offers and identify dangers at an early stage. Above all, however, the systems should develop rapidly through continuous learning.
Markus Lange, Tribe Lead Innovation and Digital at Signal Iduna, started working on data science in mid-2018. The first models developed in-house have been in use since the end of 2021. Now he is talking about a breakthrough in artificial intelligence on a broad scale, because freely available and pre-trained models such as the GPT-3 language processing model from the American non-profit organization OpenAI can be used. By this he means applications in voice communication, of which the chatbot ChatGPT is currently the best known.
“For us as an insurer, it is now a matter of understanding the new models for our purposes and applying them,” says Lange. Insurers have to follow two strands: On the one hand, the technology has to be mapped in their own systems. This is usually done using cloud solutions, as these offer the necessary computing power. On the other hand, you have to explain the direct benefit for employees and customers.
pitfalls in transition
The challenges the industry is facing can be seen in detail. At the moment, chatbots are dominated by the deterministic process in which a predetermined dialogue is conducted. Put simply, the follow-up question builds on the previous answer. The model looks like a tree, with the trunk going up into smaller and smaller branches.
However, the future is stochastic dialogue, where customers can ask their questions openly. But there are pitfalls lurking here. “A lot can go wrong with a small database, as ChatGPT showed at the beginning,” warns Frank Sommerfeld, head of Allianz Versicherungs AG. That’s why Allianz still thinks it’s too early to introduce large customer groups to the stochastic process.
When developing digital insurance products, these new products should first be tested early on with a small and affinity target group, and then more and more segments should be added to the test. Feedback is constantly being obtained and implemented so that the product keeps getting better. Even after the final launch, testing will continue and the product will be optimized. “The user experience of a digital product has to be like the experience of visiting an insurance branch – we’re still a long way from that at the moment,” says Google manager Meessen.
In many cases, this also involves the question of the extent to which technological innovation can support customers and when the technology has to hand over contact to a human being. “An omnichannel strategy is only intelligent when it recognizes when a channel change is necessary – for example, a change from the chatbot to a human as soon as the chatbot can no longer help,” Meessen continues. Many providers are currently working intensively on pushing this transition as far back as possible.
>>Read here: How ChatGPT creator Sam Altman will use AI to change your life
There are concrete applications of chatbots and the possibilities of artificial intelligence for insurers in three places where they usually come into contact with their customers: when concluding a new contract, when communicating with existing customers and in the event of a claim. The latter is always a special challenge: For many customers, the image that they have of their insurer is decisive here.
At Allianz, many damage reports are now processed completely online, with the customer being guided through a digital process. As a rule, these are comparable standard damages.
Intelligent claims processing
Artificial intelligence has also been integrated into claims processing at Signal-Iduna for around a year and a half. “In the future, it will help processes to run faster and more easily and tend to result in more correct decisions,” expects Thomas Jacobi, Head of Composite Claims at Signal-Iduna.
In the second half of the year, the insurer intends to provide its customers with a simple solution for taking photos and filming damage to motor vehicles. These then go to a service provider for a quick estimate of the repair costs using AI.
>>Read here: AI stocks – With these four ETFs you can now get the trend in your portfolio
In the future, artificial intelligence will also be increasingly used by insurance fraudsters. It is feared that images could be manipulated much better in the future than they are today. Or using ChatGPT, the exact sequence of events in the accident could be described as realistically as possible.
The insurers themselves are aware of these increasing dangers and are trying to counter them. “We also use artificial intelligence to detect cases of fraud, to find out whether images have been edited, for example,” says Allianz board member Sommerfeld. Nevertheless, the providers remain realistic: On the one hand, insurance fraud has always existed, and on the other hand, Signal Iduna expert Jacobi believes that there will never be 100% security against attempted fraud.
Looking to the future, Jan Meessen expects insurers to make further technological and digital progress in all areas – especially given that artificial intelligence is gaining in influence and importance. One approach in the “Service” area, for example, are explanatory videos that convey the topic of insurance, which tends to be cumbersome, in much simpler terms than has been the case so far in the general terms and conditions. “The demand for explanatory videos on YouTube has grown significantly since the first corona lockdown and offers insurers the opportunity to increase customer loyalty,” observes Jan Meessen.
At Allianz they are already putting this into practice. There, customers receive personalized videos in which they are explained, for example, that accessories such as the bicycle lock are also insured in the event of a bicycle being stolen.
More: These six industries could benefit enormously from AI