Documents always appearing in Decipher Verification Tab despite training

[nikhiln]

Hi all,

We’re currently working with Blue Prism Decipher and facing an issue: every document we process still gets routed to the Data Verification Tab , even after training.

All the Batches uploaded have default status of Capture performed - ready for Verify

Our setup so far:

• Classification works correctly (document types are being recognized fine).
• We have trained and pushed well over 50+ documents.
• Each time we still need to manually map fields during Verification.
• On the Extracted Data , We are getting Confidence : High , Validity : High

Questions:

1. Roughly how many documents are typically needed before Decipher gains enough confidence to bypass Verification?
2. Are there specific threshold settings or configurations that we may need to adjust?

 

Regards

[Ben.Lyons1]

Hi @nikhiln,

To get high confidence batches to go straight through you can either enable auto-verification in the Batch Type or in the Configuration screen.

Have you tried this yet?

Thanks

[nikhiln]

Hello Ben,

Thank you for your response , based on your screenshot above we have tried enabling the auto-verification in the Batch Type and in the Configuration screen.

But the documents status is still Capture performed - ready for Verify. Also while loading the batch in the Data Verification screen, none of the fields are auto populated and we manually have to map the fields.

[Ben.Lyons1]

Hi @nikhiln ,

Thanks, I understand. It looks like then that there's an issue with the training/capture aspect, but there's a few things that could be contributing to this, such as permissions, DFD configuration and Document Type configuration.

I would recommend checking with you Blue Prism account manager to see if you have any Knowledge Share hours available with our professional services team, as you may be entitled to some product help.

Alternatively you can raise a support ticket and we can take a closer look.

Thanks

[nikhiln]

Hello Ben, 

I have reconfigured the DFD and Document type and now we are able to extract the document directly without going into the verification step.

thanks again for your help

Regards

Nikhil

[kiranb25]

Hi @Ben.Lyons1 ,

I am facing the similar issue, where the exception documents are not separated as new batch and proceed with the correctly captured ones. Any suggestions?

Out of 6 documents in the batch, unable to capture the required fields for one document which stops the batch in "Capture performed - ready for Verify". But our expectation is to process the 5 documents in one batch as " Verify performed - ready for Export" and separate the document as batch in exceptions folder in data verify tab.

Note: In manual verification we can do that manually by going to the specific document and mark as exception which moves to exception folder and submit the remaining but in automatic data verification its not happening.

>Trained the data as required.

>In document type settings:

> In batch type settings:

>marked the same document as exception in manual verification.

settings are enabled in batch as below for auto data verify, even in config tab as well.

appreciate your suggestions.

Thanks,

[Ben.Lyons1]

Hi @kiranb25 ,

Decipher won't automatically identify an exception, they have to be done manually, there are no automatic rules for this. Decipher not capturing a field isn't an exception, it just needs human intervention.

If you want documents with the fields correctly assigned to flow straight through when the batch includes 1 or more documents that are missing fields, you need to select the "Separate documents with high confidence in new batch before data verification" for the Batch Type.

Thanks

[kiranb25]

Got it, Thanks. Appreciate your quick response.

Thanks,