SAP BTP - Document Information Extraction

[Abhi105]

Intro

Hello everyone !

In this blog I am going to talk about Document Information Extraction Service. This is an AI/ML based service to capture data from specific document formats ( PDF, TIFF, PNG, JPEG ) thus reducing effort and errors.

Use Case

Capture the Sales order number and Sales Order Texts from an incoming email attachment, capture the result and update the Sales orders in S/4HANA. This solution entails multiple BTP & S/4HANA components: -

• BTP SAP Business Process Automation


• Business Application Studio - CAP


• S/4HANA OData


• BTP Connectivity and Destination


• BTP Document Information Extraction

Over a period I will try to share the other pieces of the bigger picture via blogs but for now I will focus on the last piece where Document Extraction Service reads data from a PDF file

Pre-requisite

Subscribe to “Document Information Extraction UI” in the BTP subaccount. Please note the trial variation “Document Information Extraction Trial UI” might provide only standard Schemas and Templates. Hence to have an option to create custom ones, a non-trial option might be required.

 

Set Up

Once subscribed, you can see the following in your subaccount: -

Subscription in BTP Subaccount

Click it!   You would see the below menu options :-

Menu

Settings

1. Click on Schema configuration and Create as highlighted below


2. A popup appears. Enter the details as shown below :-

Create New Schema

      3. On Clicking Create you should see the Schema created with the possibility to add Header and Line-Item fields

Schema Creation Continued...

4. Add Header fields


5. Add Line-item fields


6. Click on Activate

Templates

1. Add the values as shown

Create Templates

2. Activate

Document

1. Add the values by clicking + button

Create Document

 

2. Fill in the parameters as shown below

              

Fill parameters

3. Drag Drop the file format which will come over in the email attachment and from where the service is suppose to read the data.


4. The file format I have kept is very plain & simple to showcase that any layout is understandable by the algorithm provided the file is in the correct format.


5. Click on Steps2, Steps 3 and Review button to check if the document is uploaded successfully and confirm.


6. Post Confirmation, the status might take a few seconds to become READY. Once done, Click on Extraction Results

 

Extraction Results

 

7. Click on Edit and Enter the values for Header and Line items

Header

Line Items

        Save the values.

 

8. Add to Templates

Add to Templates

9. Add Another document with the same layout following step 1-6 and check the extraction results. It should now pick up the values from the document automatically

 

 

This service could be used in SAP Business Process Automation to fetch the records and carry out further processing with the relevant data. Stay tuned !

 

More Info

Please refer to SAP Help