Digitalisation

Scalable AI: Out of the lab and into your insurance business

A lot of businesses still find implementing AI to be a struggle. Even though many companies have already kicked off AI prototype or pilot projects, stabilising algorithms or otherwise supporting scalable AI remains a challenging solution. According to Gartner Research, only 15% of use cases through 2022 leveraging AI techniques such as machine learning, deep neural networks, and Internet of Things environments will be successful. Regardless, AI technology is developing at such a high speed that scalable solutions are beginning to emerge more frequently. 

17.06.2021

Overcoming black box systems

It is clear that even though the speed of AI development is increasing exponentially, businesses run into limits on scalability due to AI’s sensitivity to legacy systems. Not to mention that complex company-wide AI systems often appear as black box systems that are more vulnerable to systematic errors and are more difficult to fix when a failure is detected. This is why it is predicted that, through 2023, at least 50% of IT leaders will struggle to move their AI predictive projects past proof of conception to a production level of maturity.  Needless to say, even with the challenges, AI is disruptive to all industries and is impacting how we do business – insurance industry included. So, as legacy systems present some roadblocks, other avenues are opening up in the form of brand-new ways of doing business. 

Applied AI and new AI-related business models

Take for example the success of Lemonade, the insurance company powered by AI and behavioural economics that recently announced a US$ 34 million B round. Also breaking new ground is the life insurance company, Bestow, that lets AI make the decisions and quotes via an online process which is fast for the user and requires no preliminary doctor’s appointment. To be sure, as more and more insurers begin to launch AI-accelerated underwriting programs (AUW) that take advantage of machine learning capabilities, we can expect to see the traditional application process phase out in return for a more streamlined and simplified underwriting experience. The Munich Re insurance product, aiSure™, is already removing uncertainties by offering an innovative insurance-backed performance guarantee for AI solutions.

Scalable AI as a present-day strategy for the future of insuring

As new ways to construct algorithms continue to take shape and generative modelling evolves, scalability will in turn become more attainable and evaluation metrics will thus improve. This will bring exciting changes to insurtech models in risk management, underwriting, insurance pricing as well as customer relationship management. We can also expect to see traditional business continue to evolve too as more data points are used, providing more in-depth UW, the automation of internal processes such as claims, the automation of external processes, e.g. interfaces to doctors and progressive support for product generation. 

Munich Re regards the broader sector of AI as one of its strategically important topics – one that will create new competitors as well as completely new emerging business opportunities. The feasibility is already evident across all industries and, in particular, insurance companies. Munich Re is dedicated to the belief that AI should serve people while all ethical aspects are respected, and fairness and justice are ensured. Within this principled sphere, we will continue to research AI, while helping our clients maximize the potential of AI for their business. 

Our experts
Hermann Schwarzmeier
IT Architect
Munich Re
Marco Röhrle
Head of AI
ERGO
Dr. Stephan Meyer
Head of AI
Munich Re
Print