We have an customer that is already using DocuWare but it is not happy with the manual work required once the tray OCR result don't bring supplier name and ABN(Australia Business Number) from document header.
I was checking the PDF on adobe reader and looks like the header of the PDF is a image and the rest of the document is a normal PDF structure.
After some tests we figure out that converting the PDF to JPEG before importing to tray, improved the OCR result bringing the supplier name and the ABN.
So we are thinking to provide an solution that is going to convert the PDFs to JPEGs before importing into the tray.
This issue is only happening for just some suppliers documents. But they receive a lot of documents from these suppliers.
The customer is receiving all documents via email so one solution could be:
- Create an app the is going to read the account payable email inbox folder
- For each email
- Get the attached PDFs and convert to JPEG.
- Create a new email with the same subject and email body
- Attached the JPEG images
- Move the original email to /pdf folder
- Send email to the final email that DocuWare is reading
Please feel free to give any advice or ask any question.
Quick note. (I imported the same document into Amazon Textract and the OCR result was much better.)