If the source PDF is a non-selectable scan, the feature automatically runs Google’s OCR (Optical Character Recognition) during the export process, making the new personal copy searchable.