How Did the CEE Get Its First Insurtech Unicorn?

Last week, two-year old Ominimo has officially reached a unicorn status. Over roughly 15 months, Ominimo expanded its business from a valuation of €200 million ($230 million) during its Series A round led by Zurich Insurance Group, to a $1.6 billion valuation for Series B (EBRD-backed).

How did this happen in such a short span of time?

Understanding the industry inside-out

It all started with Dusan Komar (CEO) and Dennis Weinbender meeting at McKinsey. In their work, they were mostly concerned with two things: applying AI and machine learning in pricing, distribution, and claims; and the so-called digital business building, where they worked on “digital attackers” and digital B2C solutions for the major insurance players.

Over their time at McKinsey, a few recurring problems in the industry turned up.

Systems based on technology from the 80s and 90s that were extremely rigid. They made modern pricing and introducing new variables impossible, recalled Dusan. “I remember asking one client how many man-days it took them to add a new variable to a linear model, and the answer was devastating: 700 man-days. They had to manually adjust it in a thousand different places!”

Traditional insurers use about a dozen variables and linear models, which is not enough. For example, in the Netherlands, out of 10 million drivers, an average insurance player might distinguish maybe 10,000 risk clusters. That means thousand of drivers end up in the exact same premium bucket even though their underlying risk is different, Dusan explains. As a result, good drivers subsidize bad ones because everyone in that bucket pays the same premium. “We said: if we can use 150+ variables and non-linear machine learning models (like XGBoost) to differentiate 10 million different risk profiles, we can offer good drivers a lower premium.”