Implement import of PDF to RadFlowDocument, using PdfFormatProvider. This would require text recognition (including determining where paragraphs end as opposed to new line inserted because of the layout) and table recognition. This will allow conversion of PDF documents to docx, RTF and HTML.