Personalized premium and services
Foreign experts predict increasing use of AI systems in insurance in the near future, because it is estimated that customers would be happy to modify the data collected by IoT sensors for a lower premium.
The insurance industry has always suffered from a lack of confidence. Inertia of traditional insurance companies in the acceptance of innovations, their attitude to customers that is at the level of statistics and problems in the payment of damages only deepen the gap between insurers and their clients who are difficult to acquire and even more difficult to maintain, because their loyalty depends precisely on trust. Therefore, the key issue for insurers is how to keep a client. This is where the artificial intelligence steps out.
Although it sounds to many people like something that will concern us only in the distant future, artificial intelligence has actually been part of our lives for years now and it is only a matter of time when it will become part of our everyday life. Examples for that are the application Siri (Speech Interpretation and Recognition Interface) on iPhone smartphones that recognizes speech of the owner and communicate with him in natural language giving him recommendations and answers the questions, as well as the development of autonomous vehicles that is already in great progress,
When it comes to insurance, the acceptance of technological innovation in this industry started relatively recently, with the appearance of Insurtech companies that offered easier access to insurance products and more efficient payment of claims, which made traditional insurers reconsider their practice and now they are slowly adapting to new circumstances. However, customer expectations are growing every day – it is no longer just a matter of buying the policies online - customers are now asking from insurers to prove how well they know them in any type of interaction. User experience has to be personalized, but to achieve that, it is necessary to have as much information as possible about every client which can be analyzed and utilized at any time.
Treating clients equally is no longer sustainable - from statistics and averages, the focus shifts toward the individual person. Instead on predictability, insurance premiums are determined on the basis of a particular individual data.
For a person this is an impossible task and therefore the world's largest insurance companies like Allstate, Liberty Mutual, Progressive, State Farm and Geico have already turned to the use of artificial intelligence, Internet of Things, sensors and machine learning in the collection and processing of customer information, as well as platforms and applications based on AI in the actual contact with the insured person in order to collect and analyze customer data, detect certain patterns and gain a new perspective on each client individually, with the aim of deepening customer relationships and creating mutual trust.
Experts predict increasing use of artificial intelligence in insurance in the near future, because it is estimated that customers would be happy to modify the data collected by IoT sensors for a lower insurance premium. As reported by Bloomberg, according to a survey by multinational IT service companies Tata Consultancy Services, by 2020 insurers will invest about $ 90 million in technology of artificial intelligence.
According to Milutin Cvetković, Business Development Manager at Saga system integrator company, new client’s expectations are primarily reflected in their autonomy and the possibility to do more things independently from the service provider.
”Self-service culture” is the one that retains the existing or attracts new clients, but of course, we should not forget that the users’ trust is based on technical characteristics of services - stability, security and speed, then the availability of the service (anytime, anywhere), as well as on its personalization (only and exactly for me, as I want it)”, explains Cvetković.
He explains that the so-called “Internet of Things” in the first step in this. The term represents the networking of multiple devices, systems and services, which, according to Cvetković, gives the ability to gather as much data as possible from as many places as possible and then to connect them to the individual client:
"Only treated like this, they become information. Data is collected in different ways - via sensors, manual input, donation by the client (in the form of filling up the form for input or response to a survey), via monitoring their behavior and habits in the existing in-house dedicated applications (mobile, desktop or web), to social networks or their correspondence with the service provider. "
By 2020 insurers will invest about $ 90 million in technology of artificial intelligence
Using the example of frauds which pose a major problem for insurance companies (in the United States,about 40 billion dollars is spent every year on covering frauds, while in the UK 350 frauds is covered on daily basis), Cvetković explains how artificial intelligence – AI, can help insurers, but also their clients.
“Technology-enabled automation ejects wise human action wherever possible and therefore the possibility of errors or malversation. In particular, if we are talking about damage application, the time stamp intersected with authority of photography is very important for determining the size and connectivity of damage, or damages. Currently, insurance companies are mostly concerned with the prevention of fraud on the client’s side, but at the client’s expense. The effect of this ‘at a client’s expense’ is further enhanced when it comes to honest customers. When it comes to car insurance, customers have to be present at the damage assessment and undergo the procedure (which costs both time and money), and the fact is that the client’s final impression of the insurance company will be based on his interaction with employees (appraisers, civil servants) and the procedures themselves.”
Such and similar processes within the same insurance company, says Cvetković, can be significantly improved by monitoring and analysis of the history and behavior of the client, taking and processing of customer data on the basis of validation of sources from which they are sent (authorized user, geolocation, time stamp and etc.), and with the openness of services and end-user autonomy.
Latest trend – Insurance for smokers via selfie
One of the most interesting novelties in the world when it comes to artificial intelligence is that insurance can now be purchased through selfies. Since last year, startup insurance company Lapetus has offered the opportunity for the client to take a selfie, based on which the application SMILe (smoker indication and lifestyle estimation), though facial analysis, accurately determines the health risks associated with smoking and life expectancy of smokers without the need for complicated medical examinations.
After IoT system has done its work, there comes automated processing of customer information - machine learning, and with it the information about the product provided in the simple and quick manner, in which, when it comes to artificial intelligence, the most important role is played by the so-called front end and associated services that actually publish this information. Cvetković explains that the combination of these technologies (from IoT to machine learning) represent chatbot solutions and adds that the example for that is the intelligent chatbot Rea which has been recently implemented by Saga into the system of Raiffeisen Bank.
"We are talking about the software that emulates the human communication to facilitate the daily work processes and it has already started its first conversations with beneficiaries of this bank. In order to get in touch with it, it is enough for the beneficiary of the bank to open an account on the social network Facebook and search the official site of the bank, while for Viber communication, one needs to be the follower of Raiffeisen public account.
With the help of AI, insurance premiums are assessed on the basis of an individual instead of on the basis of predictability.
Service itself is multilingual and provides a variety of information, from the simplest service ones, such as the location of the nearest ATM-s or branches, one’s credit status, or giving support in case of theft. The uniqueness of this solution is that, by using NLP modules (Natural Language Processing), it allows users to communicate with customers in a way similar to inter-personal communication. Processing of entered phrases and requirements consists of finding meanings on the basis of which the right answers are provided, such as, how to become a client of the bank, or what documentation is required for different services, "Cvetković said.
Better service, plus savings
A dialogist to the Insurance Council notes that the use of chatbot system is one of the ways to increase the chance of customer retention.
“Customers want their banking or any other type of financial services to operate in a similar way like Google or Facebook, and Weaver platform on which Rea is based, with chatbot solution on the front end side, was designed to favour and facilitate interaction with the end customer, and therefore is the perfect solution for insurance companies. Based on data collected on clients, insurers can detect what the customers need and what would suit their needs at the right time, and based on this, they can create an individual offer. Also, this solution makes it possible for the client to initiate a session on his own or to be automatically recognized (whether online or in the app), and to be preventively offered assistance of any kind,” explains Cvetković.
He notes that, depending on the amount of data and sources that carrier Weaver platform is linked to, the degree of autonomy of self-employment is growing, and it is also possible to report the damage remotely, from home, from the site of an accident or damage, via Viber for example (automobile liability insurance, full coverage, etc.).
“Also, we are able to convey all the relevant pieces of information and keep them for later and further processing in the system, and with the use of AI and machine learning, existing processes and algorithms of decision-making are additionally optimized and the usefulness of reporting is improved.”
In this way, it is possible to make huge savings in terms of the necessary internal resources in insurance company,” says Cvetković.
The impact drones have on the industry - the flying over in service of insurance
The use of drones in risk and real estate damage assessment is increasingly spreading around the world. It is estimated that in the US alone their impact on the insurance industry in the future will be worth 6.5 billion dollars, and here's the list of benefits drones are already bringing to insurers and their customers:
Easier real estate assessment
Drones allow up to 50 percent more efficient evaluation of real estate, thanks to automation, the use of sensors and ability to take pictures from the air. They also save on personnel, which, thanks to drones, can be better deployed within the company, and also reduce the risk of injury of employees who no longer have to "climb" the roofs to assess the damage.
More accurate risk detection
The use of drones for risk assessment prior to the conclusion of insurance contracts can reduce the cost of the policy, if, for example, real estate is unlikely to be affected by a natural disaster, and this allows the personalization of the premium.
Less room for frauds
After a major disaster, such as floods or earthquakes, when a large number of insured persons make out their compensation claims, the use of drones in assessing damage of real estate can reduce the number of frauds and false reports of damage.
Drones and AI – a winning combination
Drones and artificial intelligence go “hand in hand”. Their integration into AI systems will reduce dependence of drones and enable the reduction of costs and time needed for the payment of damages, thus increasing the satisfaction of insurers and customers. According to Milutin Cvetković, drones will be increasingly used for real estate insurance, and now they can be implemented to existing systems of AI and machine learning.
Cameras in vehicles
Listing additional advantages of using artificial intelligence, our interviewee singles out the possibility of a better recognition of the client in comparison with traditional reliance on encryption and password.
“Artificial intelligence can recognize the pace of a person's life, the places where he stops, his physical appearance, and is even able to distinguish between a photo and a real person, with a distinctive eye movement. Successful transmission and image processing allows authorization of users in a way pleasant to them, so the need for questioning customers about data such as customer’s name, father's name and so on in order to determine the identity ceases, whether it is online shopping or just conversation with the call center. On the other hand, some insurance companies are already using cameras installed in the vehiclefor the purpose of automobile liability and damage assessment, and we all know the importance of the video recording or images as the evidence. With the help of AI, it is possible to transfer and store the image immediately, and the image processing is important in the prevention of adverse events, so in the future autonomous vehicles will receive lower insurance premiums,” says Cvetković, and adds that with the help of artificial intelligence it is posible to identify and access limited areas of business, and even to discover the processes that precede a robbery.
According to him, the technology has matured in terms of infrastructure (in terms of speed of data) and processes (in terms of power, or computing power), and all above mentioned technological solutions based on the IoT approach and AI mechanisms will certainly provide significant savings to the companies that are providing service, and on the other hand, it will facilitate the use of the existing services or provide customers with the new ones that they have not been able to use until now, be it for safety or performance reasons.
Source: WORLD OF INSURANCE | No. 3 | Year VIII | Belgrade, March 2018.