Product Information
What’s New in SAP AI Business Services: Q1 2023 Release Highlights
Dear AI experts,
It’s time to introduce you to the latest & greatest features introduced in our SAP AI Business Services portfolio this quarter, all available on SAP Business Technology Platform:
Document Information Extraction 📲
Automate and scale your business document processing for a wide range of document types using machine learning.
Extract barcode symbology
This new capability enables users to automate processing of barcodes more easily in documents with the ability to get the type of barcode used, such as QRCode.
Read more in this dedicated blog post.
Scene text recognition functionality
Document Information Extraction is able to extract text from images “in the wild”. The service can be now applied for use cases, such as extraction of container IDs in logistics execution.
Read more in this dedicated blog post.
Catalog API for filtering, ordering, and pagination of uploaded documents
This API provides a more efficient and convenient way of working with lists of documents.
Read more in this dedicated blog post.
Embedded support in the user interface
Users can now submit Service Now tickets directly through the Document Information Extraction UI. The UI code automatically collects the information required for the ticket after providing a short description of the ticket.
Enhanced SAP-managed (pretrained) models for standard document types
Users can expect improved accuracy of the header field and line-item extractions for following document types: supplier invoice, purchase order, payment advice. Users can expect an enhancement of both the efficiency (better processing time) and the effectiveness (less errors) of the underlying business processes.
Updated language support for template processing
The Business OCR engine of Document Information Extraction provides a broader support for languages. The catalogue of new languages can be found in this blog post.
Adjustment of the Document Information Extraction UI
We have adjusted the web components of Document Information Extraction UI to simplify the user’s journey when annotating documents. It provides a better user experience and simplification of document-annotation activity. Watch it in action with templates in this dedicated blog post.
Ability to provide fixed data in template extraction fields
The templating functionality has been enhanced to provide fixed values in selected extraction fields for a specific template, for instance, for the supplier or customer. Users can store a field value in a template during annotation, for example, a tax number or a customer number in the supplier’s template, using the more-robust automation provided by Document Information Extraction.
Moreover, the team hosted an SAP Community Call in February in which they introduced a new extraction model approach for purchase orders, payment advice, supplier invoices and more to achieve higher extraction accuracy.
Replay the session below or read this blog post by our product specialist to learn more.
Personalized Recommendation 📲
Get accurate recommendations based on users’ browsing history and/or item description using machine learning.
Content-based model enhancement
We have developed a content-based approach that allows the machine learning model of Personalized Recommendation to learn directly from the item catalog when no clickstream data is available.
This feature helps to:
- Overcome a lack of an AI training data set (clickstream) and support AI model development.
- Enhance the AI recommendation model to deliver higher business value and new revenue streams.
- Reduce privacy risks.
Data Attribute Recommendation 📲
Perform classification and regression tasks using machine learning.
Business blueprint: sales order completion
Data Attribute Recommendation enables you to leverage business blueprints with a dedicated pre-processing pipeline, tailored to specific use cases.
The latest addition is the newly available business blueprint for completing sales order documents.
This new functionality not only provides the possibility to predict missing fields in sales order documents but also qualitative metrics for each field you want to predict.
The business blueprint also recognizes and takes care of imbalanced datasets so that the performance feedback of each label properly reflects the reality.
Learn more how it works with our product manager Daniel Stoerkel 👇
Exposing performance metrics for each label for model training
We have extended the scope of Data Attribute Recommendation from only exposing global model performance metrics after a successful training by exposing performance metrics for single labels.
This feature helps to:
- Provide model performance metrics for each label.
- Gain a better understanding of trained models with multiple labels.
Service Ticket Intelligence: The Path Forward
Service Ticket Intelligence has been set to maintenance mode since February 2023.
All existing Service Ticket Intelligence customer contracts will be honored and supported until their contract expires.
While there will not be a direct successor to Service Ticket Intelligence, the capability continues to be available inside the SAP Customer Experience (CX) solution SAP Service Cloud. Customers should contact their SAP account manager or reference account team for more information.
SAP plans to continue to invest significantly into advancing AI. Learn more on SAP.com.
See you next quarter for exciting product innovations with SAP AI Business Services! 🚀
