US Patent 12111869 Identifying an implementation of a user-desired interaction using machine learning

A system for identification of an implementation of a user-desired interaction. The system receives a user-provided data item associated with the user-desired interaction type. The system determines a user group of a user. The system searches a plurality of applications to determine a plurality of interaction types supported for the determined user group. The system determines a plurality of suggested interaction types from the supported interaction types based on the user-provided data item. The suggested interaction types are ranked according to relevancy. The system retrieves, from the plurality of applications, first metadata and first NLP tags for the most relevant suggested interaction type and second metadata and second NLP tags for the next most relevant suggested interaction type. The system formats and communicates to the user the first metadata, the second metadata, the first NLP tags, and the second NLP tags according to rules of the determined user group.