“A man with new ideas is crazy until he is successful.”

Who would have thought 20 years ago that a bot will call us and it will be & nbsp; able to answer our questions? Today it is a fact. In every area of life, also in the work of an insurance agent, artificial intelligence can significantly improve the comfort of work and facilitate the performance of many duties.

We have many tools and possibilities at our disposal that used to be solely in the hands of man. Here are some of them:

Hello, says the bot – i.e. virtual consultants or chat bots, which can relieve the agent from the initial data collection from the client. In addition to talking to the client, they are also able to use geographic or social data.
Price comparison websites – here the customer enters basic data on his own, reads the offers, and then has the opportunity to contact the agent in order to purchase insurance or obtain more information. There are also comparison websites that enable direct purchase of the policy online.
When they see you, that’s how they write you, i.e. video verification. We do not need to meet the client in person to be able to confirm his identity. Artificial intelligence will do it for us.
ChatGPT- can facilitate the agent’s work on many levels. From searching for potential interested in insurance, through advertising and marketing, to adapting the offer to specific recipients.
Supporting bots – can support, for example, the emotional intelligence of an agent or a broker, through the transcription of a conversation. They look for sentiments and direct the agent to the right style and way of speaking.
Big Data, i.e. huge data sets, giving us more information about customers, and thus the possibility of offering a personalized offer, e.g. based on the history of websites viewed on the Internet, medical history, information about penal points. We can also build an offer based on the customer’s previous experience or his activity in social media.
Bots, chats and data sets also support agents’ marketing activities. They allow, inter alia, For a better selection of marketing activities, they will suggest how to communicate with customers. Thanks to artificial intelligence, we are able to estimate the value of our sales funnel, which will allow us to better predict future revenues and we will be able to take steps in advance to adapt the sales strategy to the current situation.
Artificial intelligence is able to predict transactions or recommend a product. If someone spends their money on specific purposes and in a specific way, they will definitely take advantage of the insurance. AI finds here regularities that a person might not even think about.
There is advertising, there is a customer, there is a product. It’s time for a contract. Blockchain allows you to conclude contracts and settlements. It also helps to automate and speed up many processes such as creating and monitoring insurance history or initial risk assessment.

Artificial intelligence It gives us many opportunities to improve our work. a chance to reduce financial outlays, expand the scope of our services, better personalize offers, services or communication, and increase the number of customers. However, in addition to the benefits of AI, it carries some risks. It can skip the real needs of the client if they go beyond the range of definitions they know. Not all customers’ needs are zero-one, so not everything AI is able to help us. It works on the basis of learned models. If it is to deal with something, even prosaically simple, that has not been learned before, it will simply not solve a given problem. Connection difficulties, possible errors or deliberate actions of malicious programs may interfere with artificial intelligence. Since it works on huge databases, it is important to keep them properly secured. The use of modern amenities also carries the risk of unclear rules of liability. If our bot makes a mistake, who will answer it? It does not matter to what extent we use AI support, its correct functioning is possible only in the correlation with human work. AI checks a person, and a person has to control him.

We must learn everything that is new, and so is artificial intelligence. If we want to use its resources correctly, we must devote funds not only to the purchase of software, but also to training in its operation. As with all trainings, it is an investment.

Knowledge of the solution is not the only barrier that may limit the use of AI, there are also legal issues.

The insurance industry in Poland is one of the most legally regulated sectors of the economy (KNF, Financial Ombudsman, Office for Personal Data Protection, UOKiK, PIP). Consequently, we cannot automate it just like that. in 2021 The European Insurance and Employment Retirement Programs (EIOPA) has published a report entitled Artificial Intelligence Governance In the European Insurance Sector ”. It contains the recommended rules of using artificial intelligence in the insurance industry. 6 basic rules were distinguished[1]:

The principle of proportionality

According to it, the insurance company should introduce a management system that is proportional to the nature, scale and complexity of its activities. This step should be preceded by an individual assessment of the impact of the AI application, made to adapt the appropriate management measures to a specific case of implementing an artificial intelligence solution. Ultimately, the use of AI and the measures taken for this purpose should be proportional to the impact of their use on the situation of consumers and the insurance company. Then the insurance company should verify the above-mentioned resources in the context of the target fulfillment of the ethical effect and trustworthy use of artificial intelligence.

The principle of honesty and non-discrimination

The use of AI by insurance companies should be characterized by integrity and a complete lack of discrimination against all stakeholders. As a result, they should avoid behaviors causing social inequalities, especially in the context of customer ratings and offering price-optimized products. The principle of fair use of data manifests itself primarily in respect for the consumer’s autonomy and the development of AI systems in order to provide him with full support in decision-making processes. It is worth mentioning the reliable and transparent management of AI processes, fair operation and rational undertaking of activities that take into account the results of AI systems, as well as minimizing the use of personal data.

The principle of clarification and transparency

The use of AI is associated with the need to prepare a comprehensive and understandable explanation of how the artificial intelligence model works. The effect of applying the above measures should be the consumer’s support in the proper understanding of the algorithm’s operation and the conscious decision making. The whole process must take place with the full awareness of the consumer that he is dealing with an artificial intelligence algorithm, not a living person.

The principle of human supervision

The essential issue is ensuring that insurance companies provide an appropriate level of human supervision over AI systems throughout the entire cycle of their operation. The task of the societies is to assign and divide appropriate functions and tasks to people involved in the implementation and use of AI systems. The choice of a higher or lower level of human surveillance should be proportionate to the nature, scale and complexity of the risks associated with the specific application of AI and take into account the combination of management measures applied to the given case of its application.

Data management principle

In order to ensure appropriate standards of personal data protection, insurance companies using AI should comply with the relevant national and EU regulations throughout the system’s lifecycle.

The principle of resistance and stability / quality of solutions

Insurance market entities should use proven and reliable AI systems. Each time they should take into account their intended use and the potential for causing damage. The operation of the system should be assessed and monitored on an ongoing basis, while developing appropriate measures of their effectiveness. Additionally, they should be implemented within the framework of properly secured IT infrastructures.

The creation of the report and its content suggest that the issue of using artificial intelligence in the insurance sector is slowly beginning to be noticed by supervisors that approach it with proper caution and understanding. Of course, we are not dealing here with “hard regulations”, but it should be assumed that we will also see them. It also seems that the creation of similar guidelines from the hands of the Polish KNF is also only a matter of time.

[1] https://gu.com.pl/sztuczna-intelligence-w-ubbezpieczenie-wedlug-eiopa/