How would you structure the architecture of a prototyping app using MVI to handle complex user inputs and AI model responses?
