A small survival guide for insurance companies regarding the use of AI

Brussels, 11/09/2019

This is not a drill

Remember when your phone’s sole purpose was to text or call people? Let’s speed up to 2019: your phone has replaced your computer, your iPod and your camera. It can do pretty much everything thanks to the thousands of apps you can download. You can also pay with it while asking Siri to help you find the best place to have a great cup of coffee in the neighborhood. We now live among self-driving cars and smart technologies. Business as usual used to be good enough. It’s not anymore: while disruptive companies are becoming the new norm, artificial intelligence is revolutionizing the way we live and work, and the insurance sector is no exception.

While most of European citizens are studying the plausibility for it to happen in real life, other countries are clearly ahead of us regarding the use of AI in business. For instance, American and Chinese insurance companies give us a glimpse of what the insurance sector could be made of a few years from now in Europe.

Within the giant Chinese insurance company Ping An, the Xiang Hu Bao healthcare-coverage product covers 65 million clients, that are managed by only 50 employees. Apart from compensating them in real-time, the company has additionally developed facial expression analysis software, that studies the claims that policyholders have provided by video, in order to detect whether they are telling the truth.

In the United States, the start-up Lapetus is trying to personalize life insurance policies simply by taking a picture of your face, making it possible for the application to estimate your lifespan and offer you a custom-made life insurance product accordingly. This should remind us that ethical questions should be placed at the center of the debate on AI. As for the start-up Lemonade, it requires 90 seconds to subscribe to a personalized insurance policy and only 3 seconds to reimburse in the event of claims.

“The head count associated with claims management will be reduced by 70 to 90% by 2030.”


Data Science: all the cool kids want to do it

So, what’s next for us? No major waves of job cuts are to be expected, even though some professions will disappear, such as administrative task managers: according to McKinsey, the head count associated with claims management will be reduced by 70 to 90% by 2030. However, skills that cannot be automatized or digitized such as problem-solving under uncertainty, collaboration, and emotional intelligence will become increasingly valuable on the job market. Artificial intelligence should not be seen as a threat, but as a way of optimizing our work by suppressing boring and repetitive tasks from our lives- if the shift is done in time.

In the insurance sector, artificial intelligence will allow us to better assess risks (especially regarding risks with limited history such as cybersecurity), to carry out a better prevention through prediction, especially regarding auto, health and natural catastrophes. We will also be able to perform a faster, finer and smarter underwriting with a reduced threshold above which the risk is insurable.

For instance, artificial intelligence is already used by the start-up Cytora which indexes data from websites, press articles and public documents, analyzes it thanks to Machine Learning algorithms and then defines new risks that are not insured yet. It will also allow personalized pricing and faster claims management along with a better customer service, better fraud prevention and a reinforced rentability.

A shift is already happening in the market, and the most sought-after profiles are data engineers, data scientists, technologists, cloud computing specialists and experience designers. Actuaries will still be central for insurance companies; however, their set of skills will have to evolve accordingly. As a matter of fact, Belgian universities begin to offer double master’s in actuarial and data science to meet this need.

“43% of the operations conducted in the insurance sector could be carried out by the AI.”


Drive the change or be driven by it

This shift towards a digitized business model should not be driven by the IT department only but sponsored by members of the management committee. On one hand, the Chief Data Officer should rely on a solid team to carry out projects, but on the other, employees from each department should be aware of the importance of data issues as well. An Agile team organization will also be the key to success for many projects.

Although data science technologies are currently used to a certain extent (mostly regarding automation and fraud detection), companies should adopt a broad-based strategy and rethink the insurance company’s own model in line with the expected macro-level changes in order to become a truly data-driven company. The strategy adopted by top management must include the four key elements of any AI strategy according to McKinsey: data capabilities, models and tools, change management and organization and talent. This last element is decisive as hiring the right talent will help companies succeed in their transformation.

Insurance companies should keep in mind that the assumption of being “too big to fail” is never a good idea. Neglecting emerging digital technologies leads to bad business decisions: ask Kodak or Toys’R’Us how the business is doing. Ultimately, according to a McKinsey Global Institute study, 43% of the operations conducted in the insurance sector could be carried out by the AI. We will move from a “detect and repair” scheme to a “predict and prevent” approach, which will remodel every aspect of the industry.

While the short-term challenge is to reduce costs, the long-term challenge will clearly be to make customers love their insurance company, a crucial aspect for the sector in order to avoid the extinction of traditional actors.

To catch up on all our latest articles, make sure to subscribe to our monthly newsletter here !


Written by Leonie Correard

Analyst at Asquare Partners