The scenario reads your PDFs from storage, recognizes text even in scans, and answers questions about the documents — with a pointer to the source.