Intelligent Invoice Scanning – Review UI Enhancements with release 2202

[stefan_resag]

While SAP Business ByDesign Intelligent Invoice Scanning provides a high degree of process automation, still most of the scans should be reviewed by a user before creating an invoice. With SAP Business ByDesign Release 2202 several enhancements have been added to further improve user productivity. In this blog I will introduce those:

1. Bounding boxes for scanned values

Before release 2202 the PDF panel on the right side of the review UI was used to simply display the scanned PDF document. With release 2202 this PDF panned has been rebuilt completely. First and most obviously, bounding boxes indicate that this section of the PDF document was source of the scanned value. Blue bounding boxed show header fields, whereas orange bounding boxes are used for item data. When hovering over with the mouse, the corresponding scanned value and field name are highlighted as a tooltip.

2. Highlighting and positioning scanned values

Although bounding boxed and tooltips already simplify the review of the PDF document significantly a user might still find it difficult to locate the scanned values in a document, especially for large and crowded invoices. To address this a user can now easily locate and highlight a scanned field by simply clicking on one of the reviewed values in the left side of the UI.

As a result, the selected value will be highlighted with a tool tip and the PDF document is automatically scrolled to the corresponding value.

 

3. Annotation of fields in the PDF panel

Finally, in case a value was either not scanned correctly or not scanned at all, users can select the corresponding section in the PDF panel with the left mouse button pressed. As a result, a popup will be launched where users can correct the scanned value and map it to a field in the document. This can be done both for header and item data.

In my example the user has selected the order ID in the PDF panel and maps it to the ‘Order ID’ field from the drop-down list. Of course, the annotation information is saved in the scanned document for later reference, in case needed.

Hope you find this new PDF panel helpful when reviewing scanned invoices or credit memos.

Please also have a look at my video.

[r_ray]

Hi Stefan, some nice updates in the 2202 release.

I have a few questions regarding the solution:

1. Can you default the currency from the supplier master data?


2. How long does the re-training take? In a previous video there is mention of 3 months plus


3. Can the PDF control allow continuous scrolling as opposed to having to click page up/down?


4. If linked to a PO, do you ignore the scanned line details when creating the invoice?

I've not seen anything in the 2205 release notes that covers these points.

Thanks,

Richard

[stefan_resag]

Hi Ray,

thanks a lot for your comment. Please find below my answers:

1. The currency is determined by the Document Information Extraction (DOX) service. So far we have not planned any defaulting of the currency based on ByDesign master data, since we think that the scanning result is more accurate.


2. Before re-training happens the invoice documents undergo a personal data check, that will cause 90+% of the invoices to be excluded from retraining (GDPR). Currently we are working on a solution to also use these invoices and also to reduce retraining duration, but currently it is still 3+ months.


3. May I ask you to create an improvement request in our customer influencing portal: Influence Opportunity Homepage - Customer Influence (sap.com)


4. There is a business configuration question that controls this (see screen shot below):
   
   
   
   • If the highlighted question is not selected, then the system will simply add the PO items in case the PO ID was scanned
   
   
   • If the highlighted question is selected, the system compares the scanned items with the items of the purchase order and only takes over the values that are not scanned from the purchase order for those items, like account assignment. The scanned quantities and values are however used.
   
   
   
   

Best regards

Stefan

 

 

[LlucH]

Hello Stefan,

 

Nice blog and nice improvement!

But, regarding point number 3, what about supplier and buyer company?

If you try to select the information directly in the pdf document, system gets text data but does not match any business partner, so in this case it does not work for those fields.

Example:

In your pdf you can select "IT-Profs Inc.", select that this is the Supplier data, and this text will be transferred to Supplier field. But then ID SRSUP01 will not be matched so, system will not consider this field as properly filled, and you will have to type it directly in the Supplier field, as before the improvement.

Same for buyer company.

 

Could you review if this expected behaviour?

 

Thanks,

Lluc

[stefan_resag]

Hi Lluc,

right, the system only copies over the text, but it should behave similarly as if you were typing in the text. Please try the following: select the text in the PDF, so that it is copied over to the reviewed value. Click into the reviewed value and add a space. System should now trigger the value help, from which you could select the correct supplier or buying company. Because of this behavior we found it is still reasonable to support that feature also for supplier and buying company.

Best regards,

Stefan

[former_member79992]

Hi Stephan,

thanks for improving - nices ones!

I´ve got two questions:

1. although we have put the vouchers into training, the "status of re-training" is always "not started". What are we doing wrong? (we are testing in the UAT-environment)


2. is there a possibility to read out the manual adjustments of the users in the review or to call up the documents that were given into the training in tabular form (logging would be advisable for GDPR reasons alone - but would also result with the solution to question 1)?

Thanks and best regards,
Thomas

[stefan_resag]

Hello Thomas,

thanks for your comment.

1. Have you selected the check-box that confirms that there are no personal data on the UI? If so the status should change. Please be aware that another automatic check in the DOX service once more checks against personal data and this might lead to a rejection for retraining. Please raise an incident if in your case the status does not change.


2. The review object is PSM released and you should be able to access it via a partner add-on.

Best regards,

Stefan

[former_member79992]

Hi Stefan,

thanks a lot for your answer - we struggle a bit with the review object. Unfortunately "ScannedInvoiceStagingObject" is not PSM released. Do you know the object-name?

Thanks and best regards,
Thomas

[stefan_resag]

Hi Thomas,

according to our development the object is available in the 2208 version of the SAP Cloud Studio. I have attached a screen shot of the same.

Best regards,

Stefan

[former_member79992]

Thank you very much, Stefan!

[cgrosselin]

Hello Stefan,

Thank you for the detailed blog.

Is this supporting the creation of custom extension fields ?(Through the studio or even KUT)
I can see that the BO's root node itself is extensible, but I imagine there is an underlying logic to map fields ?

Best regards,

Clement

[stefan_resag]

Hi Clément,

creation of custom extension fields is not possible as per my knowledge. When you talk about mapping logic, I assume you mean the mapping between the PDF document and the scanned fields, or do you refer to the mapping between scanned and reviewed values? The recognition of the PDF fields is based on a global model, that cannot be influenced via KUT or add-on.

Hope this clarifies your question.

Best regards,

Stefan

[cgrosselin]

Hello Stefan,

I was indeed talking about the mapping between the PDF and a scanned field, with the possiblity of adding a custom field there. (The idea would be to map a custom serial number taken from the PDF).

Thank you for your reply, this does clarify my question.


Best Regards,

Clement