Learn how to extend and personalize SAP applications. Follow the SAP technology blog for insights into SAP BTP, ABAP, SAP Analytics Cloud, SAP HANA, and more.
In a world where large volumes of data need to be processed and managed by organizations, the SAP AI Business Services portfolio comes in handy. By applying machine learning algorithms, these AI-powered services available on SAP Business Technology Platform can (semi-)automate and optimize manual, repetitive and error-prone tasks, and therefore realize more intelligent business processes.
In this blog post, let’s focus on two of these services: Invoice Object Recommendation and Data Attribute Recommendation. You will get to know how both services are becoming one as of December 2021, and why this is beneficial for SAP customers.
What is Data Attribute Recommendation
Data Attribute Recommendation (part of the SAP AI Business Services portfolio), supports with the creation, maintenance and management of structured data by predicting and classifying categorical entities, e.g. product categories, and predict numeric values such as ratings and numbers, using text, numbers, and categories as input. It can be used to enrich missing attributes of data records and classify incoming hierarchical information through state-of-the-art automation workstreams and processes.
It is a generic classification and regression service that applies machine learning to all data records independent from the business problem. Some common use cases are:
Get suggestions of material class and its characteristics when creating new material requests
Get international trade commodity code predictions when adding a new product
Solve master data inconsistencies
Obtain a price estimation for a product based on its description
What is Invoice Object Recommendation
Invoice Object Recommendation follows the same logic, with a stronger focus on procurement and finance use cases. Invoice Object Recommendation proposes G/L (general ledger) accounts, cost centers and profitability analysis dimensions (also known as CO-PA dimensions) for incoming invoices without a purchase order reference.
This is a specific business-related classification service for financial data that applies machine learning to classify invoice line items. Common use cases are:
Prediction of finance related invoice objects where there is no purchase order attached
Prediction of invoice objects in the material management area where no purchase order can be found (and therefore also no invoice objects can be pre-filled)
Merging Invoice Object Recommendation with Data Attribute Recommendation
As SAP continuously aims to enhance the performance of its services offered to its customers and given the fact that both services are enhancing and complementing each other perfectly, SAP has decided to combine these AI capabilities and their business focus within one common SAP AI Business Service going forward. The above explained functionality of Invoice Object Recommendation continues to be available as a business blueprint template* in Data Attribute Recommendation. Customers who want to automate use cases – which have been formerly implemented using Invoice Object Recommendation – can now use the Invoice Object Recommendation business blueprint available within the Data Attribute Recommendation service.
It leverages the Data Attribute Recommendation extended capabilities which will directly benefit SAP customers. Customers can now bundle resources and head for a one stop shop for all their classification and regression tasks. The service offers full model lifecycle and more flexibility in running and training models simultaneously. Further details on benefits will be shared later in this blog.
The service’s functionality will continue as it is, just not as an independent solution. However, there are small but significant changes behind the scenes in data and model management that will greatly benefit SAP customers without reinventing the wheel.
In this article, these changes are explained in more details, including explanations on how the data and model trainings are handled and how the APIs endpoints interact with each other.
Benefits from Using Data Attribute Recommendation Capabilities
As mentioned before, Invoice Object Recommendation users will be able to experience a broad range of benefits by using the Invoice Object Recommendation business blueprint within Data Attribute Recommendation. Please find a short summary of those benefits below:
One single solution for classification and regression tasks
Full model lifecycle
Multiple models at the same time
Models can be deleted at any time
On demand deployment/un-deployment
Flexible number of predictions as output
Flexible input/output structure
Not restricted to G/L account & cost center only
Custom features can be taken into consideration
No currency conversion as mandatory pre-processing step for a customer is needed any longer
Single DAR instance can be used for multiple use cases (when creating multiple dataset schemas)
Train multiple models simultaneously
Using more generic service that would fit multiple use cases
Lower price
Increased file upload size (single upload instead of batch uploads)
Classify and predict a flexible number of invoice objects (such as General Ledger accounts, Cost Centers,…)
Roadmap and Future Releases
We are also continuously enhancing the Data Attribute Recommendation service with additional features and performance optimization. Besides that, the service is also capable of solving regression tasks, there are numerous innovations planned for the upcoming year of 2022 which can be found in theSAP Roadmap Explorer.
Pricing
As the Invoice Object Recommendation business blueprint is part of Data Attribute Recommendation, now the Data Attribute Recommendation pricing applies. We are happy to share that the Data Attribute Recommendation service is cheaper to run than the Invoice Object Recommendation service. The table below shows and compares the earlier pricing of Invoice Object Recommendation with the current one of Data Attribute Recommendation:
Invoice Object Recommendation**
Data Attribute Recommendation**
Metric: Blocks of 1000 records
Up to 14 blocks, 250 €
Up to 5 blocks, 320 €
From 14 blocks, 188€
Up to 100 blocks, 64 €
Up to 250 blocks, 32 €
Up to 500 blocks, 19 €
Up to 750 blocks 13 €
From 750 blocks, 10€
Models/additional deployed models
N/A
0.91/hour (if more than 3. No limit to trained, but undeployed, models)
Feel free to use the Pricing Estimator to simulate the cost associated to your future usage.
Read our technical blog to learn more about the technical implications of this merge.
*) A business blueprint is an entity that encapsulates the business logic of a specific use-case and is responsible for mapping a business problem to a ML task
