I would like to see programming as a school subject
Sofie Quidenus-Wahlforss is one of Germany’s youngest AI entrepreneurs. The 37-year-old is the founder and CEO of the Berlin AI provider omni:us and has the vision of a digital information society accessible to all. She can be found on the list of the “Top 50 Women in Tech 2018” issued by Forbes and is currently digitalizing the insurance industry, a sector that still operates with highly traditional methods and in which some companies are finding it very difficult to change course. But she is convinced that the technology of her company will benefit not only the insurance companies, but above all their clients. As part of our series “Europe, Your Digitalization”, we spoke to her about the challenges of digitalization in the insurance industry and why it is so difficult to recruit AI experts.
Sofie, you work in one of the promising as well as, simultaneously, one of the most traditional industries. AI meets insurance. How did you happen to focus your attention and that of omni:us on the insurance industry and to decide you wanted to digitalize this sector?
When we started out, we initially had one thing above all: a technology. So our primary aim was to understand as precisely as possible where our solution would find problems it could solve best. We wanted to identify the industry in which our AI would offer the greatest benefits. We found this need in the insurance industry. The companies in this business have to deal with a large number of widely different documents every day, above all in claims management. Recording and assessing this material is extremely costly in terms of work and time today. Digitalization offers huge potential in this regard, and this is exactly the point where our technology comes in.
What are the greatest challenges during collaboration with, or the digitalization of, insurance companies?
Insurance companies are often both very old and very large establishments. On the one hand, tradition plays an important role for them while, on the other hand, structures have developed and grown over long periods of time. IT systems stand out especially as tangible examples of this as they can vary from one department to the next. The same is true of decision-making channels and processes. For instance, we see over and over that attitudes toward new technologies are very open at the upper management level, but the departments are dominated by reservations. This extends decision-making processes and requires “educational work” at many points to allay the concerns of employees.
What do you foresee happening in the future development of AI, especially in the insurance sector?
The phase of AI implementation is just now beginning. A lot of the previous discussion has been no more than hype, but in recent years more and more applications that can generate true added value have appeared on the market. So we have barely begun with the implementation, and we frequently find it almost impossible to imagine all the things that AI will make possible. But we must also remain realistic; AI is not a magic wand and will not solve our problems in miraculous fashion. What it can do, however, is handle repetitive and memory-intensive tasks much faster and more reliably than people. In short: wherever there are data, AI will be around as well in the future. Large quantities of data are being produced in all possible areas of our lives today. This is especially true for insurance companies. Their many documents contain hidden data treasures that are not being utilized at all today because the effort using the means available at this time would be too great. AI will change this. Not only does it offer to companies tremendous potential to save resources and to achieve a much higher level of automation – it also gives them back the freedom to take very good care of their clients, to “get to know them,” and to put together offers for them that actually meet their desires and needs. The implementation of these technical opportunities will also determine who is successful on the market in the future. Before long, nothing will function in this industry without AI.
Are other European countries already operating on the market at a higher level of digitalization? Can Germany possibly learn from others?
There are undoubtedly some aspects in which Germany is lagging a little behind. The infrastructure is an ongoing topic of discussion. For example, we would certainly benefit if the expansion of broadband services were accelerated. I repeatedly notice things in everyday life as well. In Austria, my home country, for instance, digitalization in the offices of public authorities has already reached a far higher level. When I have to take care of official matters, I can usually do it online simply. In Germany, I must still go to the office personally, take a number, and print out my documents in duplicate so I can bring them with me.
Is the AI master plan of the Germany government sufficient to catch up with the world’s leaders, or is it no more than a drop in the bucket?
I am an entrepreneur, not a politician, so I want to be careful about handing out advice. But it is not necessarily the case that Germany has fallen hopelessly behind the best performers in the world. In top research, for instance, German institutes are right at the forefront. The problem has more to do with the frequent sluggishness in turning research results into commercially viable applications. Naturally, this is also a matter for the market, but politicians must improve the general conditions, e.g., by simplifying the provision of venture capital and making this step more attractive for investors. This would be extremely important for startups so that they can develop their ideas to market maturity. And startups are often the bridges between research and the market because they are agile and can quickly seize upon and implement the continuous advances in development arising from research. This insight has perhaps not yet been anchored sufficiently in people’ consciousness; the AI strategy still concentrates strongly on the traditional player industry and midsize businesses.
How do you recruit your employees?
Recruiting international AI experts is a challenge. The recruiting processes are significantly more intensive and take longer than in other sectors. We invest plenty of discussions and maintain a complex assessment center with the aim of finding suitable candidates wherever they are in the world. Last year, we established an “AI Hub” for data scientists from all over the world in Berlin. We determined at that time that the motivation “to crack a really hard nut” is enormously important. The appeal of becoming a part of a team that tackles difficult challenges for which there are currently no solutions anywhere in the world is enormous. Especially when it concerns an innovation that will noticeably change our everyday lives. Moreover, we offer to our potential AI experts the freedom to be creative, a working environment with data science at its heart, data records and computer performance for experimentation, colleagues with international expertise from respected universities, and the freedom to publish the results of their work during international conferences and to participate in top international events. Thirty-seven of our 60 employees from 31 countries work on the tech team.
And what must happen so that more and better experts are trained in Germany?
If I could have my wish, programming would be a school subject like a second foreign language so that young people can discover and find out what coding really is and whether they enjoy doing it. Degree programs and advanced training would certainly help as well. But this can certainly not be realized in short order. We assume that we will have to look for talented people on the international market in particular in the future as well. However, Germany – especially Berlin – is truly attractive to foreign experts.
How digital are your insurance companies? Do you have any practical examples?
The industry is still in the very early stages of digitalization. The potential is huge, however, even though there are clearly differences among the various companies. Still, what I have determined is that there is plenty of curiosity everywhere. Most of the companies have understood that they have to do something. Frequently, though, they do not know exactly what and how. Often enough, their thinking is restricted too narrowly to improving current processes. Yet AI offers the potential to realize truly data-driven work and to replace old and tedious processes. This means that the data, not pre-defined procedures, determine what the next steps will be. This makes companies faster, more agile, and more efficient. And they regain time and breathing room to respond empathically rather than bureaucratically to the needs of their clients.
But when we see that investments in insurtechs are clearly on the rise and that some large insurance companies can be found in the ranks of investors, I believe that it will not be long until we see some changes. We want to be the leaders here.
Sofie, thanks for the interview!