Microsoft Dynamics 365 | Integration, Dataverse...

Invoice capture - OCR

(1) Share

Report

Posted on by vix7

Hello everyone,

I've tried Invoice capture, a new solution from Microsoft (which was released in September). I tried to learn OCR on 50 invoices which are from similar vendors, but it didn't learn anything. Sometimes for some fields (e.g. vendor name) it is nearly 100% sure of itself, but sometimes the percentage for the same vendor is like 50%. How can I make it better, and learn OCR this things? Do I need maybe a better license? Does even OCR learn right now or is it different for every invoice that I upload no matter the same vendors or other fields? Also, when he recognizes the wrong field I usually link him to the right field, but for some fields like vendor name I can't even put the link on the right place because he didn't gave me option to learn him that.

Just note that I have read the Microsoft Learning for Invoice capture, and I don't need links like this: https://learn.microsoft.com/en-us/dynamics365/finance/accounts-payable/invoice-capture-overview

I need answers.

Thank you!

I have the same question (0)

All responses (2)

Answers (0)

MuthukumaranAX 2,903 on at

Like (0)

Report
Copy link

Link copied!

Hi,

We are trying to adopt an OCR solution, and we noticed that OCR does not read certain fields correctly and fails the invoices. I think your approach to training the model with sample invoices is correct. However, you might face some issues in reading the file due to the poor quality of the document or the tool being unable to read it correctly.

One possible solution could be, you can create and train your model as needed which could result in better improvements. This feature of attaching your model in Invoice capture has not been released yet.

Thanks

Was this reply helpful? Yes No
Suggested answer

Aghajanyan Ruben 2 on at

Like (0)

Report
Copy link

Link copied!

Hello,

You might consider using 3-rd party control to integrate OCR functionality directly within Dynamics 365 CE forms.

It enables you to design models for various document or image layouts and link them to entities—all without requiring any coding.

Additionally, it includes an Action (API) that allows you to leverage OCR capabilities through Power Automate, workflows, or other front-end or back-end implementations.

For further details, please refer to the links below.

https://www.youtube.com/watch?v=0QyMir25L2Q

https://marketplace.bevercrm.com/pcf-controls/11

Was this reply helpful? Yes No