Question in Document understanding invoice processing

Help Activities
, , , , ,
1

Hi All, i have a scenario in Document understanding. I’m using document understanding to extract my invoice details. I have 4-5 specific invoice types to extract data. If i get any 6th invoice format i need to reject that without processing. Can anyone help me how can i achieve this.

1 Like
2

Hi @Learner007 ,

Could you let us know How are you identifying and accepting the 4-5 invoice templates as the valid ones to Process further ? Is it already done ? Also maybe provide us with the Details of the Current Implementation.

3

@Learner007

If you’re using For Each to loop the files, through index, you can stop the loop.

image

4

Hi @supermanPunch we are not identifying any yet i need to setup that now. Before i’m getting all correct invoices. It’s getting all correct. Now getting extra files i need to setup files identifying. Is it possible to do it with taxonomy

5

Hi

Actually you can use a confidence threshold to filter out any invoices that do not match your specific types in UiPath Document Understanding
The “Confidence Threshold” is set to a high value to filter out any invoices that are not likely to be of your specific types. Any invoices that do not meet the threshold will be rejected and the workflow will continue processing the accepted invoices.

Check this document for more details

https://docs.uipath.com/document-understanding/standalone/2021.10/user-guide/data-extraction-overview

Cheers @Learner007

6

Hi @Manju_Reddy_Kanughula i can’t do like that because i don’t know which file is wrong one

7

Hi @Palaniyappan . Is there any option where i can reject my document by checking the taxonomy headings. Because if few of my “ENTITY” doesn’t match with my invoice means we can reject right. Can we setup that kind of option, is it possible to do that.

8

How are you validating the correct and wrong on what basis?

9

@Manju_Reddy_Kanughula till now i haven’t setup that. I want to do that now. I don’t know how to do that.

10

Extracted data will be considered only if the threshold set is met
If it doesn’t meet the confidence level then it will be rejected

@Learner007

11

@Learner007 ,

You would require to understand the Difference between the template files, between the 3-4 accepted invoices and the other Invoices or other Templates that maybe present.

For Example :

  1. If you have 3-4 Invoices template and if these invoices are from a fixed set of Vendors, they have their Name in the Document. We can include those Vendor Names in the Document for the classification part. So it only recognises those documents.

This is to be done by setting up the Classification Part in the workflow.

12

@supermanPunch invoice structure is same and vendor also same. Only values will be changed from invoice to invoice

13

Ok @Palaniyappan i will try this also

14

@Learner007 ,

Would suggest you to start the implementation maybe by referring to the Document Understanding Process Template which should be accessible from the Studio. You should then be able to understand the different stages of the Document Processing that is done.

There are documents already present in this template, You would just require to configure the values in the Orchestrator related to your Tenant.

Then try to also pass the documents that you have and check if the Extraction works for Other Invoices or other Templates as well.

Maybe some parts would need to re-created or re-configured.

image
image922×384 13.3 KB

15

Sure @supermanPunch i will try this once