PDFs are meant to define how a document will look in print. The exact opposite of structured data. So you have to "interpret" it visually. There are software solutions that combine the functionality of an OCR and information location templates and thus extract the right information from the right place in the document and based on the template transform it into structured data. Basically it allows you to define that a document is an invoice and that the invoice number is to be found at coordinates x and y.
Document Intelligence seems to be such a solution and you can give it a try by procuring it from Store.
