Discrepancy in data when pdf is read using Mendix OCR

We have a particular use case in which we need to process the given invoice and Fetch the data and store in our Mendix System. But we have a issue in processing the invoice the OCR isn’t recognizing Invoice details properly. The processing of line items i.e. tabular data is not processed properly, when we upload unstructured pdfs the rows present in the invoice isn’t recognized properly. Is there any way we can come across this issue. Adding to it we are having 2000 unique templates which need to be processed and each one have their template and most of them unstructured one. So any generic solution.  
1 answers

What are you using for OCR?