Hello, I have a simple AP project that uses Smart Fields. The process uses an email archive to bring the invoice into the repo. The email archive setup was to take PDF attachment and generate pages, and do NOT keep the original PDF. That was working.
A weird issue came up with these documents looking blurry in Web Client, so I switched it to not generate pages and keep the PDF.
Now, some invoices are going through SF extraction and returning nothing. It's like SF just doesn't work with some invoices.
I took one of these Amazon invoices that failed to get any data, generated pages, deleted the electronic file, and applied the SF template. It returns the invoice data without a problem.
Why would the file type change the results of the smart fields extraction? Anyone else experiencing this?
Discussion
Discussion
Smart Fields PDF not extracting correctly
posted on March 9
0
0
replied on March 9
Mind sending us a sample document that failed? Or the approximate time when it failed?
0
0
replied on March 9
•
Show version history
Here you go. Just a simple amazon invoice.
In my system, it was created today at 7:45 AM nad SF took a minute.
No invoice number, no PO number, no invoice amount, no invoice date.
As soon as I generated pages and removed the PDF, SF worked
0
0
replied on March 10
Thanks, I downloaded it and removed it from your post because it had bank account info on it.
0
0
replied on March 10
Thank you.
0
0
replied on March 11
So far, it looks like it was because the OCR engine only returned 2 words from the entire document, so there was nothing to work with. We're still looking into whether that's a failure of the OCR engine itself or something happened when the images were generated from the PDF.
0
0
replied on March 12
A fix has been deployed.
0
0
You are not allowed to follow up in this post.