I'm not sure how these insurances are sold to end users at small scales. I'm involved in building crop harvest forecast systems to be used by insurers who insure large businesses (think multinationals) who need resources from mostly Asian and African countries for the production of their goods, e.g. palm trees. So this use case is different in scale from your example, our work is mostly in risk modeling on a large scale.
That said, if I were selling insurances, I'd want two things: data to justify or explain my rates to my customers, and data to make predictions of my own risk and/or profit margins. The two might be correlated in a highly competitive and transparent market, but I have a hunch that these insurances are not, and that there isn't a whole lot of pressure to work on a cost plus pricing model, so that these insurers can work on a value-added pricing model.
That said, if I were selling insurances, I'd want two things: data to justify or explain my rates to my customers, and data to make predictions of my own risk and/or profit margins. The two might be correlated in a highly competitive and transparent market, but I have a hunch that these insurances are not, and that there isn't a whole lot of pressure to work on a cost plus pricing model, so that these insurers can work on a value-added pricing model.