Appreciate the question jndsn. My background is in statistical learning itself, though I have a long experience in P&C insurance. In short I'd say "using machine learning" is not one thing and varies significantly in approach, quality of application, etc.; there's a variety of things we're adding that explain why our software provides a 3-4 point loss ratio improvement for our first customers that would be too long to explain via post here. If you're interested in getting involved and/or learning more I'd encourage you to apply via the instructions listed at the bottom of the links above.
No comments yet.