The concept of automation is inextricably associated with robots, and they are associated with the future. Meanwhile, today many processes, both in production and services, are robotic, which consumers do not even know. Automating insurance sales processes becomes all the more necessary the more goods people want to insure. The more wealthy a society becomes, the more things we can protect and the more we can afford to buy policies.

Companies don’t make it with the number of customers?

Every year, the insurance market records records for both product sales and paid claims. As you can easily guess – it is not about a radical increase only in the number of claims, but also about an increase in the total number of, for example, cars in Poland, and therefore the proportion of motor claims. According to the Polish Insurance Association (PIU), the number of collected premiums has doubled over the last 10 years. The value of policies sold grows by an average of PLN 1 billion year on year, while the value of claims paid grows by PLN 500 million (1). This data shows the scale of the market, but is it possible to handle all requests for offers and compensation manually? The owners of insurance companies are looking for solutions that will support sales and will partially help people in activities that they cannot cope with due to their volume. So how do insurance companies deal with the efficient and effective sale of their products in the conditions of continuous increase in demand?

RPA and chatbots for customer service

The development of automation in the sales service is becoming a key trend in a world where everything is more and more and faster and faster. Therefore, in order to be able to carry out more and more sales, processes that a human cannot handle, must be automated. Robo-consultants, i.e. chatbots – programs that independently conduct a dialogue with the client, register and support the handling of simple claims, are becoming more and more popular. They are also used to educate and raise awareness among customers, when processing complaints, assisting in the payment process, giving opinions by customers, selling policies or collecting information. They even have a name and a face so that the client feels comfortable in contact with them. Thanks to their application, a significant amount of time and resources of insurers can be directed towards more profitable activities.[2]

Another technology that is being implemented more and more often is Robotic Process Automation (RPA), which supports insurance companies in automating the processes of paying policies (they may include even several hundred thousand payments per month), where an employee is replaced by dedicated software (software robot) to support repetitive, multi-series activities. This is a scale that even the largest sales teams cannot cope with.

Artificial Intelligence on hotlines

Artificial intelligence algorithms have made their way into the insurance industry for good. Apart from chatbots, some insurers (but also banks and telecoms) introduce a voice menu in the hotline. The technology allows you to avoid situations that are frustrating for the customer, in which one gets lost in a confusing IVR menu or repeatedly transferring the call to subsequent consultants. It also has an economic advantage for the client, because gaining quick access to reliable information does not have to pay high costs for long calls to the hotline.

Image recognition improves the life of the insurer and the client

One of the very frequently used AIs is automatic image recognition. The application is so diverse that it serves both to improve customer service, through facial recognition to verify identity, and to guarantee and process claims faster. Optical character recognition helps to digitize documents and makes them easier to save and search. Thanks to satellite images, it is possible to accurately estimate agricultural insurance and assess its risk, based on available information about crops, weather and landscape. Home insurance prices can also be refined using pictures of the insured property that are traditionally difficult to capture, such as the roof of a building. Extreme risk events such as floods, hurricanes, tornadoes and fires can be monitored and their impact on insurance claims updated in real time thanks to automatic image recognition. In summary, such solutions improve the work of insurance companies in each of their departments.

Fighting frauds

Insurance fraud and extortion is a scourge that the insurance industry grapples with around the world. Only in Poland, the amount of losses resulting from the payment of compensation for damage that was not present remains at the level of 10% of revenues. The problem is all the more so because crimes related to insurance fraud are practically impossible to fully identify. Thanks to the use of artificial intelligence, insurers can catch at least some frauds, and thus save money.

Our Data Science team at Promity has created a solution that contributes to reducing the number of frauds. The fraud detection process takes place in three stages: First, individual damage is segmented on the basis of photos sent by the client. Then, from the photos, essential features are extracted from the damage images, and the algorithms compare the features with the set of compensation cases. Thanks to the use of Deep Learning, we can compare the characteristics of the entire damage with other similar ones and find similar-looking damage in the database. Due to the high sensitivity of technology, the algorithms, with the help of engineers, “learn” the distribution of data on which they operate, so in the initial phase of technology implementation it is required to overtrain the models (neural networks) and tune the algorithms for a given distribution. The implementation of automatic comparison of a given case with tens of thousands of other cases in a very short time and with high precision is tempting, bearing in mind the possibility of stopping financial losses.

What will be in the future?

In fact, it is difficult to predict the future, even if there are premises to predict it. One thing is known – artificial intelligence will develop in the insurance industry in the same way as the industry itself. An increasing number of people, an increasing number of goods will increase the demand for insurance, and thus the sales process will have to be basically automatic, and a person will appear in critical moments of the sales process. Will the industry be immune to the crisis? Another challenge for insurance companies is the so-called insurance gap, consisting in the fact that less than half of the assets worldwide affected by a natural disaster were insured (according to forsal.pl).

Products are created thanks to which it is possible to insure household appliances and audio / video devices against the effects of surges or to protect property against strong winds. They are not popular yet, but they may become a standard in the future, because according to the PIU survey, even 80% of respondents are not able to cover the costs of repairs with their own funds. The climate crisis may increase the number of policies sold, and thus force insurance companies to further automate sales and claims handling.

[1] https://marutitech.com/insurance-chatbots/

[2] https://towardsdatascience.com/how-are-insurance-companies-implementing-artificial-intelligence-ai-aaf845fce6a7