OCR Engine and PDF

We use some very old software with OCR technology to 'read' invoices from vendors for Accounts Payable. These invoices come to us as pdf attachments via email. But some pdf's from vendors never make it to the software to be read.. they end up in an "error" folder - but I'm not sure why, but I am thinking that it has something to do with embedded images and fonts... etc. i.e the pdf is just too fancy for our antiquated software!  Any thoughts??  These pdf invoices come into our system via an outlook inbox - should I perhaps write a script for  outlook rules to 'convert' all pdfs to a simpler pdf that our OCR software likes? Any input would be appreciated.!