Skip to main content
kirang100133
Known Participant
Jun 19, 2023 7:14 AM UTC
Question

AI SKILL-Document Extraction

Hi all,

I am trying to extract a data from PDF using the new AI SKILL feature.

Attaching the screenshot of my PDF and Process Model.

The Output is a stored in a CDT with the same field and name.

This service does not give any data after execution and it takes 5 minutes of time for a simple process with 1 node.

Kindly advice how to proceed.

8 replies

stefanhelzle0001
stefanhelzle0001
Brainy
Jun 19, 2023 7:22 AM UTC

This is the initial training phase. You will have to add a validation step after the extraction and route to reconciliation.

kirang100133Author
Known Participant
Jun 20, 2023 3:59 PM UTC

can you please elaborate more.

stefanhelzle0001
stefanhelzle0001
Brainy
Jun 20, 2023 4:46 PM UTC

What exactly do you want to know?

Maybe my blog post helps you to get started.

https://appian.rocks/2023/05/29/appian-records-evolution/

janakiramk4892
New Participant
Aug 10, 2023 10:26 PM UTC

That is correct Stefan. I did struggle as well initially. But once it's thru initial training with Reconciliation, we will not need the additional reconciliation. Also, I noticed doc extraction is taking very long even for very simple document with just 1 page and 4 fields. 

1) Add "Reconcile Doc Extraction" smartservice after "Extract from Document".

2) once the extraction is done, "Reconcile" will create a task which can be accesses in Temp. Complete the task, with instructions in the task which is straightforward.

3) Once the above task is complete, you should be good to go and see the extracted data going forward.

JR

Terms & ConditionsAccessibility statement