Smart Doc: AI-Powered Contextual Knowledge Assistant

The increasing amount of unstructured data has posed problems with respect to their proper processing and retrieving information from them. Various attempts have been made by researchers to find answers through AI and NLP approaches such as document summarization, speech-to-text conversion, and semantic search techniques. Yet, the existing approaches are limited to single-source approaches, meaning that none of them is capable of dealing with multiple sources. In this paper, we will use Smart Doc: AI-Powered Contextual Knowledge Assistant, which can be considered as an integrated approach for the processing of multiple sources of data, such as documents, video transcripts of YouTube videos, and comments. We leverage the potential of semantic embedding and the RAG technique to process multiple sources of data. Smart Doc demonstrates high accuracy and effectiveness compared to traditional methods based on keyword matching.