There are lost of ways to extract text from a PDF file. I done it, I used VB.net
The challenge was to identify each Invoice Line, then isolate them from the comments in between.
The programming cost is considerable, because you need a sub routine for each Vendor. Each has its on Invoice format.
As per scanning or OCR a mailed invoice will be difficult, in some cases impossible.
It depends of the quality of the printer used, paper, etc, etc. I don't think it a feasible solution.
EDI still the best way to interchange data between companies. It has evolved since the 70s, and current.
Most decent ERP systems have it.
Most ERP system have a way to "export" documents to text files.