After hurricanes Harvey, Irma and Maria hit Florida and Texas in 2017, 50% of those insured were still not compensated six months after the disaster. This example illustrates one of the black spots that the insurance industry is facing today: the extremely long delays before payment of benefits.
The Parisian startup Descartes Underwriting, specializing in insurance against natural disasters, intends to respond to this problem by adopting a still undeveloped approach, that of parametric insurance whose model is based on data (mainly meteorological data but not only ). Once an anomaly is observed (like a pixel burned on a satellite image in the case of a fire) the compensation is automatically triggered.
"This approach is opposed to the indemnity model where after a disaster experts follow one another to determine the amount of compensation, resulting in long months of waiting", exposes Tanguy Touffut, founder and leader of the young company and former director of Axa Global Parametrics (Axa subsidiary dedicated to parametric insurance).
Data to better assess risks
Assurtech, which has just joined the incubator of La Défense Swave, promises to reduce response times to a few days or even hours. It is not addressed to individuals but to large multinationals. "Descartes Underwriting is a subscription agency. Our job is not to distribute insurance products, nor to bear the risk but to model it to give a price to this risk and to create innovative products to cover it. We take out risks on behalf of insurers and reinsurers »says the startup leader.
To succeed, the young company is banking on the use of new data sources and the use of home algorithms. "Having enough data and using external data allows us to be more relevant about estimating the risk and thus to have a fairer price for the best risks while some insurers have the habit of add, as a precautionary measure, an additional margin linked to uncertainty, explains Tanguy Touffut. In addition, the automated approach to risk management contributes to lowering the insurer's treatment costs and thus to returning a greater portion of the premium to the insured. Finally, the use of these parametric data helps to fight against insurance fraud.
For example, if the fleet of a manufacturer is damaged as a result of a hail event (a shower of hail of a few minutes can quickly represent 200 million dollars of damage), the damage will not be noticed by experts. They will be directly evaluated from data related to the weight and size of hailstones.
Raised 2 million euros
Formally established in January 2019, the young company is already working for more than five major insurers, reinsurers and funds..
"We have signed contracts with major industrial groups but also with major hotel parks whose activity is particularly sensitive to weather conditions", indicated TanguyTouffut.
The startup also appeals to investors. It has just completed a fundraising of two million euros, raised by the specialized site Mind Fintech, with the BlackFin Tech fund. "These funds will allow us to buy new data from major suppliers, satellite players and the world of IoT [Internet des objets, ndlr]. They will also allow us to recruit engineers and data specialists. Whoever develops the best machine learning and artificial intelligence techniques will have a competitive advantage over others. So it's a skill we want to internalize, " details the contractor. The startup, which currently employs four people, should quickly count ten. She hopes to collect 100 million bonuses in four to five years.