Today PW is unable to extract automaticaly the file properties from JPEG files, which is troublesome if you want to insert photos into the system and filter them by date (when they have been taken). You should improve PW to make these properties extraction automatically
IPTC and EXIF data could be add to the attribut exchange as a default one and prepopulate.
We would then just need to map the correct attribute in PW.
This idea could also be extend to PDF.
PDF having the same layout could be OCR using region mask. Each OCR region could then be mapp to an attribut.