Dear SAP Community,
With SAP AI, you can benefit from built-in AI for your core business processes — connecting finance, supply chain, procurement, sales and marketing, human resources, and IT.
On one hand, SAP offers ready-to-use business AI scenarios built into the SAP applications, you can immediately grow efficiencies across your business processes and gain the foresight needed to outpace your competition.
On the other hand, you can empower your IT teams to extend your SAP solutions with AI capabilities on SAP Business Technology Platform (SAP BTP), and drive faster, better decision-making throughout the organization:
- Build AI into your SAP applications with a library of pre-trained models using our SAP AI Business Services: Document Information Extraction, Personalized Recommendation and Data Attribute Recommendation
- Manage the AI model lifecycle in one central place using SAP AI Launchpad
- Deploy and run AI models at scale without compromising data privacy using SAP AI Core
This quarter, our team is thrilled to deliver new product features for the AI portfolio of SAP Business Technology Platform, helping your IT teams to infuse AI into your apps with scalability and responsibility in mind:
SAP AI Core 📲
Confidently deploy and integrate AI models designed for SAP applications, cost-efficiently at scale while preserving privacy and compliance.
Availability of content package for large language models for SAP AI Core
With the new content package, developers can integrate large language models into SAP solutions using multiple integration workflows. The initial version of the content package supports the deployment of a proxy on SAP AI Core that facilitates interactions with the Microsoft Azure OpenAI Service. Users benefit from easily generating the required assets to connect to large language models and from the standardized API of SAP AI Core that is independent of the model or provider that is consumed.
Read more in this blog post and download the content package here.
Manage files using the Dataset API
You can now upload, download, and delete artifacts from your registered object store using the SAP AI Core Dataset API, instead of using the object store-specific API directly. This is beneficial when direct access to files in the object store is not possible or desirable. For now, AWS S3 is supported by the Dataset API.
More information here: Managing Files Using the Dataset API
SAP AI Launchpad 📲
Transparently manage AI models across the enterprise. Connect to multiple AI runtimes and centralize AI lifecycle management with a user-friendly interface.
UX Improvements
Several improvements have been made to SAP AI Launchpad to improve the user experience:
- The Configurations view now contains metadata information, including the description for parameters and artifacts, as well as the default values associated with the parameters.
- The Create Configuration view now includes a list of accepted enum values as a dropdown and specifies whether a parameter or input artifact is mandatory or optional. Where possible, input artifacts are filtered by type.
The improvements should help users see more at a glance and improve their experience using the guided workflows in SAP AI Launchpad.
More information here: What’s New for SAP AI Launchpad
Document Information Extraction (SAP AI Business Services) 📲
Automate and scale your business document processing for a wide range of document types using machine learning.
Availability of scene text recognition functionality
Extract text from images using the OCR engine for scene text recognition. When you create a schema with the document type Custom, you can choose between two types of OCR engine (Document or Scene Text), depending on whether the text you wish to extract is in an image or not.
See Schema Configuration and Create Schema.
- Reduce manual work and minimize errors by extracting process-relevant texts from scene text objects ("in the wild"), such as seal IDs on freight containers.
Read more in this blog post.
Enhanced payment advice model improving accuracy of the header field and line-item extractions in payment advise notes
The service features now the enhanced (charmer-based) model architecture, improving the overall accuracy for payment advice notes.
- This innovation speeds up processing of purchase advisements, as fewer manual corrections are required.
Read more in this blog post.
Enhanced supplier invoice model for improved accuracy of header field extractions in supplier invoices
The service features now the enhanced (charmer-based) model architecture improving accuracy of header field extractions in supplier invoices.
- This innovation speeds up processing of supplier invoices, as fewer manual corrections are required.
Read more in this blog post.
New procedure for associating documents with templates in the Document Information Extraction UI
We have improved the previous process, where you selected these documents when creating the template or added them later using the “template” feature.
You can now select documents using the new “add to template” function in the “document” feature.
More information here: Add Associated Documents and View and Edit Extraction Results.
- Improves the user experience
- Simplifies document-annotation activities and the “template creation” procedure
Ability to provide fixed data in template extraction fields
You can now include fixed values for selected extraction fields in a template. If you intend to use a template with documents from only one supplier, for example, you can define the supplier’s name as the fixed value for the senderName field.
- Enhance usability
- Simplify and improve the user experience
- Increase the quality of information extraction
New Data Type country/region for Schema Fields
The new data type country/region is now available for schema fields.
Support for Country Code Conversion in Template
The Template feature supports now country code conversion.
Catalog API for filtering, ordering, and pagination of uploaded documents
A new Document API endpoint is available: post catalog.
You can use the following catalog options to get a list of all document processing jobs in a JSON file, allowing for a more efficient and convenient way of working with lists of documents:
- Filtering
- Ordering
- Pagination
The Document Information Extraction UI also supports document filtering, ordering, and pagination.
New Use Case: Reduce Your CO2 Footprint with a Smart Generative AI App on SAP BTP
Generally available on the SAP Discovery Center, this new mission enables users to build an SAP CAP application that interacts with Azure AI through SAP AI Core and a React Native mobile app, implementing a 'Smart CO2 Converter App'.
This app helps users understand how their consumption habits are driving their CO2 footprint by showing the most impactful expenses and best opportunities for emissions reductions.
Start the mission
Missed SAP Sapphire? We got you covered!
At SAP Sapphire, SAP’s flagship event of the year, we announced 15 new business AI capabilities embedded across SAP solutions like SAP S/4HANA, SAP SuccessFactors and others, many powered by generative AI. We have also expanded our partnership with Microsoft on generative AI.
Read the SAP Sapphire News Guide to discover the new capabilities or watch this short recap video instead:
Get started with AI in SAP BTP
See you next quarter for new AI-based product innovations from SAP Business Technology Platform 🚀
In the meantime, follow our AI tags on SAP Community to engage with our experts today! 👩🏻💻👨🏼💻
- Artificial Intelligence / Machine Learning
- SAP AI Core
- SAP AI Launchpad
- SAP AI Business Services
- Document Information Extraction (SAP AI Business Services)
- Data Attribute Recommendation (SAP AI Business Services)
- Personalized Recommendation (SAP AI Business Services)
- SAP Managed Tags:
