Currently i use the document understanding tools to extract invoice and ticket. However, i notice that the validation station will scan through and check for each invoices.
It seem like quite time consuming for me to click everytime especially when i run in 15 invoices at one time.
May i know is it must to validate all invoices? Or is there any way that can save time and not need to go through each invoices?
Would it impact the accuracy if we just scan through certain documents?
You can put a condition to make the action to be validated or proceeded without it.
Usually the confidence can be checked as a deciding condition
generally we create the pilot runs and then remove the validation station if everything seems to be fine.
Thank you for the prompt response. May i know how can i do it?
Could you share the screenshot?
Just not use the validation station. There is a thread for this topic already. Hope it helps
Hi. I have a large quantity of files that want to process with Document Understanding platform. But because of the presence of the Validation Station, it seems an impossible job. Is it possible not to use Validation Station? I have to somehow train the model but is there a way to skip the validation by the human?
Validation Station is not mandatory ,but if the extraction is not 100 % accurate then you can use it.
It depends on the stability ,structure and format of the document and the extracted data accuracy to decide on.
You can as well use it and place a condition to check if the accuracy is greater than 80% or more skip the validation .
This will not open the station for all documents,and will pop up if accuracy is less than 80 % or based on your criteria.
Please refer below link:
Advanced UiPath Document Understanding - Part 4 - Validating Extracted Data | RPA | HyperAutomation - YouTube