A system and method provides automated prediction of filenames and storage locations for scan files generated from a user's scan of their financial documents. A prediction model is generated for the user based on weighted values derived from preexisting filenames and preexisting storage locations for their stored electronic files. Text or images from each scan document is analyzed and weighted and compared to the user's prediction model and a suggested name and storage location is automatically generated. The suggestions are confirmed or corrected by the user prior to storage, and any corrections are used to update the user's prediction model.