How does Application Insights infer the client's device model?

Question

I\'m wondering how Application Insights infers the client\'s device model because the Microsoft docs do not state what method is used. Obviously, they must pull