When the Brother software does the OCR, what are its settings? For example, is it trashing the original document?
And at a minimum, I would suggest Kind is PDF/PS as the criteria. Even if it would technically work as you have it, you should always strive to be specific in what location you target and what you want to match.