From Underwriter to IT Service Provider – Alexander Stolte’s digitization column
One of the most exciting aspects of digitization in the SME segment of industrial insurance is that of transferring the underwriter’s knowledge to digital systems. Experience shows that underwriters who have assessed risks manually for decades view the idea skeptically. How, they ask, is it to work? They, after all, decide on the basis of years of experience whether or not to insure a risk.
It is clear that the entire treasure trove of experience of an underwriter who deals with individual risks on a daily basis cannot be totally mapped ad hoc in software. But a part of this experience can formalized in a targeted way and incorporated into software in the form of system of rules. Astonishing successes can be achieved in this way for the standard risks in SME business that are the main focus of digitization endeavors.
The starting point is discussions and analyses of standard risks with the underwriters. The focus is how the underwriter would normally react if this or that case were to arise. The aim is to lay down rules that would in most cases – i. e. normally – apply. The answers might be as follows:
- In this particular case I would normally add the following clauses in order to limit our risks.
- I normally add 5 or 10 percent to the premium if this or that circumstance occurs.
One may have the impression that underwriters make “gut” decisions and set premiums arbitrarily, but that is not the case. A good underwriter will follow a fixed schematic in his risk assessment. Structure in risk identification and assessment gives the decision maker certainty. Revealing these structures and implementing them systemically is what the digitization of underwriting is about.
These implicit structures and approaches are the blueprint for an explicit system of rules on the basis of which the software makes decisions. In most cases the system’s decision will tally with what an underwriter would have decided in the case in question. But what about the rest? How is one to assess the cases in which the system decides differently?
What helps here is a look at the reasons that decide an underwriter to depart from his typical risk assessment formula. In some cases he may simply have more information. If the system were adapted accordingly it would arrive at the same decision as the underwriter once more. In other instances a gut feeling on the underwriter’s part – a spark of subjective intuition – is responsible for the divergence. The system is resistant to intuition. It stands stoically by its implemented structures and does not allow itself to be influenced spontaneously by subjective considerations. I see that as a great advantage of algorithmic decision-making even at this stage of digitization. Indeed, once systems of this kind have more and more data at their disposal they can be significantly superior to human risk assessment in the standardized SME segment.
About the author:
Alexander Stolte spent over 20 years as an underwriter assessing risks in transportation insurance. He is now taking the digitization of industrial insurance forward as an IT and consulting service provider. In his column he deals with the experience he has gained from this change of perspective.
Image source: Fotolia – Sergey
