Onesurance supports brokers and insurers with AI technology that can predict customer behavior. Dennie van den Biggelaar and Jack Vos started the data science company in the Netherlands in 2022. Since the beginning of this year, Jan Andel has been responsible for the Belgian market.

Where did the idea for Onesurance come from? Jack Vos: "As a financial and insurance advisor, I have advised a lot of private and corporate clients. When the online carrier Lemonade entered the Dutch market in 2019, I thought that digital underwriting and claims settlement might become a big be- threat. I thought it was a very interesting technology, but I also noticed that it is difficult for Lemonade to acquire and retain customers. After all, we need human advisors for that as well, but data- & AI-driven solutions can make customer retention strategies much more effective."

In the Financieel Dagblad I read, "Dennie van den Biggelaar emphasizes that in insurance there is an evolution of specific AI rather than a revolution of generative AI. What do you mean by that?

Dennie van den Biggelaar: "Generative AI, such as ChatGPT, has been a huge hype since November 2022. Generative AI allows us to mimic human skills, but the outcome is not reliable enough for many applications in the insurance industry. AI solutions that are specially designed for such a specific task and can perform it with high reliability are."

For example,
Dennie van den Biggelaar: "With our AI solutions, we can predict, for example, whether a customer will be planning to switch to another provider within a week, a month, a year. The broker who knows this can take timely action and prevent the customer from leaving. After all, insurance is all about long-term customer relationships. Losing an insurance customer has a much greater financial impact, than losing a customer in the retail business."

Jack Vos: "We can predict with 90 percent certainty which customers will want to cancel a policy (or policies) within 12 months. But suppose an insurance broker has 100,000 customers and 20,000 of them are in the danger zone for policy cancellation, it does become difficult to keep those 20,000 on board by calling them one by one. That's why we also predict the Customer Lifetime Value (CLV). This is the future net customer value. That way, the insurance office can set up a targeted action to approach the customers with a high churn probability as well as a high CLV on a priority basis in order to retain them."

Dennie van den Biggelaar: "We have found that churn can drop dramatically, up to 50 percent. We know that with certainty because we work with control groups. If churn is good and efficiency is improved, we can deploy AI modules to grow sales. One example is the Next Best Policy module. That predicts which policy a customer will need in the near future. After all, the best way to grow is to sell more policies to existing customers. We can also use data and our AI solution to search for the most interesting new customers in the region.

What services are most often requested?

Dennie van den Biggelaar: "Our help is most often called upon to reduce churn or to stop terminations as quickly as possible. Once that's done, we often go after that to help grow a portfolio."

Who are your main audiences?

Jan Andel: "Our AI technology can be deployed by insurers and brokers. But, of course, the number of policies must be sufficiently high."

Jack Vos: "We can't help a firm with 500 clients, because it makes no sense to set up an AI process for such a small number. Those brokers already know their clients and therefore don't need us at all. Those can estimate for themselves who the runaways will be."

Jan Andel: "But for brokerage firms working together in groups, our services can of course be interesting. In the Netherlands, we offer a group module through ANVA, the Dutch variant of BRIO. So that could also be possible in Belgium. Or through the broker organizations FVF, BZB-Fedafin and Feprabel. We offer modules for groups, who those group members are doesn't matter."

How do you determine the price?

Jack Vos: "We look at the number of clients. The more clients, the higher the rate, but because we have scalable our technology, it is now also very affordable for brokers."

You create the predictive analytics based on data from the back office of brokers and insurers. How do you ensure the ethical use of the data and the protection of privacy?

Jack Vos: "In two ways. First, we follow all the ethical rules imposed by the EU, the Netherlands and Belgium, and we go one step further. An ethical cycle is provided for each offering. We always ask ourselves which stakeholders play a role in the process. We consider whether the customers involved, brokers, insurers and perhaps other parties, could be adversely affected by the predictive algorithm we want to integrate. If that is the case, if certain groups are discriminated against e.g., we try to mitigate that disadvantage. So we think about that before the process is drawn out. Secondly, all data is brought in anonymously. So we never work with data or combinations of data that can be traced back to a person."

Jan Andel: "And our practices are constantly screened by the Cronos Group. We are part of this innovative technology group that participates in more than 570 companies, working with Belgian insurers and banks. " and feed customer predictions back to the relevant system. Sooner or later, even the less innovative players will have to work more with data- & AI-driven solutions. If you wait too long, in the near future you will shortchange your customers in service quality and cost. That's why I recommend exploring now how you can link your existing IT landscape to innovative solutions that use AI. Even though that can be challenging with older systems.

Jan Andel: "Change always involves investment, but companies with legacy systems can already make their basic processes to customers (and brokers) more customer-friendly without too much investment."

Will AI applications further widen the gap between forerunners and followers?

Jack Vos: "Yes, the forerunners are taking a spectacular lead. According to the latest studies, they are often twice as profitable. And I think that lead will get even bigger, ge- living because AI applications offer exponential possibilities. That expo- nential aspect is still underestimated."

Jan Andel: "Our solutions can also be of interest to smaller brokerage firms"

Dennie van den Biggelaar: "Moreover, all our algorithms are transparent and can be explained. In addition, we are in the process of registering our algorithms, even though this is not currently mandatory."

Consultant Capgemini writes in its reports that insurers must become "data masters. But our insurers are still stuck with legacy systems that predate the Internet. Do you have a solution for that?

Jack Vos: "We at least have ways to extract data from legacy systems in a standardized way. That's not easy, but technically it can be done."

Dennie van den Biggelaar: "We extract the data from legacy systems with APIs. With this we train the algorithms

Jan Andel: In Belgium, Onesurance works closely with Jan Andel. He is an author and transition coach with 35 years of experience in commercial and underwriting roles for Belgian and Dutch intermediaries and insurers. Jan is also fluent in French and is therefore our expert for the Belgian and French markets.

Dennie van den Biggelaar: Econometrician and data scientist with 15+ years of experience as an AI strategist in the retail and services sector.

Jack Vos: With 30 years of experience in the insurance industry, he also owned a Dutch brokerage firm.

Read the original article published in the VVP here.