Arbol, the AI-native insurer, has launched an Innovation Unit to be led by Dr. Clement Atzberger, who has been appointed as Chief Scientist.
The new unit will develop modelling, data and underwriting research that will act as a foundational intelligence layer supporting Arbol’s specialty risk products, expanding the range of risks the reinsurer writes.
Arbol builds and develops catastrophe models for 14 perils in-house that it prices, and the newly established unit aims to connect its models, the data pipelines that maintain model accuracy, and the underwriting research that connects model output to pricing and contract terms.
The unit will also support underwriting Arbol’s specialty lines, including agriculture, energy, wildfire and flood, aimed at closing protection gaps for “difficult-to-price risks.”
Dr. Atzberger will initially focus on agriculture and drive advanced modelling to support increased weather and yield protection products that cover gaps standard insurance policies leave open. He is a remote sensing scientist with over three decades of experience leading research initiatives.
Dr. Atzberger was a part of the team of scientists instrumental in producing TESSERA, a data model that converts raw radar and optical satellite time series into AI-ready representations of the status of a land surface.
Sid Jha, Founder and Chief Executive Officer, Arbol, commented on the appointment, “Foundation models are becoming widely available, and we were glad to help make TESSERA one of them. With our team of leading quants, scientists and engineers, led by Dr. Atzberger, we are able to calibrate models against real underwriting outcomes, closing critical coverage gaps and pricing climate risk with greater accuracy and efficiency than before.”
Atzberger added, “In academia, you are usually done with a model once it is accurate enough to publish. Insurance demands more, because someone has to be willing to pay a claim on what the data says, and that is a big part of why I came here.
“Many of the risks Arbol writes are ones other carriers decline because the data they have was never good enough to price on. The work of our unit is to keep improving the models and the data behind underwriting decisions, so that more of what gets declined today becomes something we can write.”
