Skip to Content
Product Information
Author's profile photo Maria Goepfert

Document Information Extraction: Unlocking Greater Value with Improved Payment Advice Extraction

We are thrilled to present a new payment advice extraction model, following the path of our successful invoice and purchase order models. This update enhances accuracy, consistency, and efficiency for payment advice documents, streamlining the accounts payable process and minimizing manual intervention.

The need for better payment advice extraction

The processing of payment advice documents is a crucial step for the later clearing of invoices in the accounts payable process. While payment advice documents usually feature only a smaller set of fields on the header level compared to invoices or purchase orders, the payment advice documents often contain long tables that often span across multiple pages. Subtotals and carry-over values should not be confused with line-items and columns must also be correctly identified on subsequent pages.

As the previous extraction model was not always living up to our expectations and the expectations of our users, we are happy to announce a new model that significantly enhances the extraction capabilities for this document type.

The new extraction model, based on the transformer-based charmer architecture, demonstrates superior accuracy in data extraction from complex documents. For invoices and purchase orders we could already demonstrate that the new architecture allows for significantly improved results for line items, date fields, amounts, and auxiliary fields, as well as better handling of table layouts and nested information. With the improved payment advice extraction model, we build on top of these successful models and offer a better handling of long tables with subtotals and carryovers.

The picture below compares the extraction of the old vs. the new model. Unlike before, the new extraction model now extracts all line items are properly and carry-overs do not get confused.

 

The better extraction capabilities of the new models not only give better results but also yield more trustworthy confidences as can be seen in the following example:

 

In a nutshell: benefits of the new model

The improved Payment Advice extraction model offers several benefits to users:

  1. Greater accuracy: The new model reduces extraction errors across all fields (line items, date fields, amounts, and auxiliary fields) and by improving the extraction of essential fields leads to a more reliable and efficient accounts payable process.
  2. Improved consistency: The model provides more consistent results, even when dealing with complex tables that span multiple pages or contain nested information.
  3. Reduced resource footprint: The new model requires fewer computational resources, making it more efficient and sustainable for businesses to use.
  4. Seamless integration: As with our other improved models, customers using SAP solutions built on top of the Document Information Extraction service will automatically benefit from the higher accuracy of the new Payment Advice model.

 

The enhanced payment advice extraction model offers significant advancements in data extraction capabilities. By leveraging our improved invoice and purchase order models, we provide a much better extraction accuracy on each field and a significantly higher consistency for long tables, increasing the efficiency for payment advice data extraction and streamlining the accounts payable process to save even more time and resources.

We are curious to learn about your experience with the new model, so do not hesitate to share your experience in the comments.

 


Learn more

Read more about the news of Document Information Extraction on the help portal!

What is Document Information Extraction?

Document Information Extraction is one of the SAP AI Business Services on the SAP Business Technology Platform (SAP BTP). This ML-enabled service is available through the Cloud Platform Enterprise Agreement (CPEA) and also in the Pay-As-You-Go (PAYGO) model.

Tutorials & Learnings

Blog posts:

SAP Community Page:

Assigned Tags

      Be the first to leave a comment
      You must be Logged on to comment or reply to a post.