Microsoft patents mobile tool that turns document parts into new files

1. A computer-implemented method comprising: presenting content of an electronic document on a mobile computing device within a mobile version of a computing application; classifying, using a first set of machine learning models, by the mobile computing device, the content into a plurality of components; generating, at the mobile computing device, metadata based on the classifying the content into the plurality of components; transmitting the metadata to a server having a second set of machine learning models, results of the server taking precedence over the results of the mobile computing device in the metadata; after the classifying, highlighting the plurality of components within the mobile version of the computing application; receiving, from the server, a second classifying from the second set of machine learning models of the presented content into a second plurality of components; updating the highlighting based on the second classifying; surfacing, within the mobile version of the computing application, an intelligent-copy element that is selectable by a user; receiving a user selection of the intelligent-copy element; in response to the selection, presenting on the mobile version of the computing application an overlay slide-up interface that includes selectable options to create a presentation document or a word processing document using one or more selected components; receiving a user selection of one of the selectable options; receiving a user input selecting a component of the second plurality of components; automatically generating, on the mobile computing device and without further user input, a new document of the selected presentation document or word processing document that is pre-populated with the selected component as an object having a type corresponding to the component as classified by the second set of machine learning models; and adding, by the mobile computing device, the component to a component data store with a type of the component, the type of the component based on an output of the second set of machine learning models.