SAP AI Business Services

Overview

SAP AI Business Services provide pre-built, reusable artificial intelligence (AI) functionalities aimed at automating and optimizing common enterprise business processes. These services are delivered as cloud-based solutions on the SAP Business Technology Platform (BTP), enabling integration into existing SAP applications (e.g., SAP S/4HANA, SAP SuccessFactors) and custom-built applications. The primary objective is to enhance operational efficiency by embedding AI capabilities directly into workflows, rather than requiring organizations to develop AI models from scratch.

The suite includes a range of services designed to address specific business challenges. For instance, Document Information Extraction and Document Classification assist in processing unstructured data from business documents, such as invoices, purchase orders, and resumes. Business Entity Recognition identifies and extracts key entities like names, addresses, and dates from text. For customer service scenarios, Service Ticket Intelligence automates the classification and routing of support tickets, while Personalized Recommendation generates suggestions based on user behavior and preferences in e-commerce or internal applications.

SAP AI Business Services target technical buyers and developers within organizations that heavily utilize SAP solutions. Developers can integrate these services via REST APIs, with SDKs available for Python and Java to streamline the development process. The services are pre-trained on business-relevant data, reducing the need for extensive custom model training in many common use cases. This approach allows enterprises to adopt AI faster, focusing on process improvement and automation rather than the complexities of foundational AI model development and deployment. The services are designed to work within the SAP ecosystem, ensuring compatibility and data governance standards relevant to enterprise environments.

According to Gartner's definition of ERP, these AI services extend the core capabilities of enterprise resource planning systems by adding intelligent automation layers to functions like finance, procurement, and human resources. This integration aims to transform traditionally manual or rule-based processes into more agile and data-driven operations, supporting digital transformation initiatives within large organizations.

Key features

• Document Information Extraction: Automates the extraction of structured information from unstructured documents like invoices, purchase orders, and receipts using machine learning.
• Document Classification: Automatically categorizes incoming documents based on their content, facilitating routing and processing.
• Business Entity Recognition: Identifies and extracts specific entities (e.g., names, locations, dates, product numbers) from free-text fields.
• Data Attribute Recommendation: Suggests and automatically fills in missing data attributes for various business objects, such as product categories or material groups.
• Service Ticket Intelligence: Classifies and routes customer service tickets to the appropriate department or agent, and suggests solutions based on historical data.
• Invoice Object Recommendation: Recommends general ledger accounts and cost objects for incoming invoices to streamline financial processing.
• Personalized Recommendation: Provides tailored recommendations for products, services, or content based on user behavior and preferences.
• Cash Application: Automates the matching of incoming bank payments to open invoices, improving accounts receivable processes.

Pricing

SAP AI Business Services pricing is based on custom enterprise agreements. Customers typically engage with SAP sales for specific quotes tailored to their usage requirements and chosen services. Some services may offer a free tier through the SAP Business Technology Platform (BTP) to allow for development and testing.

For detailed pricing information, refer to the SAP AI Business Services pricing page.

Common integrations

• SAP S/4HANA: Direct integration to enhance ERP processes like invoice processing, cash application, and master data management.
• SAP SuccessFactors: Used for processing HR documents or recommending suitable job postings.
• SAP Customer Experience Suite (e.g., SAP Commerce Cloud): Powering personalized product recommendations and customer service automation.
• SAP Business Technology Platform (BTP): The foundational platform for deploying and managing AI Business Services, enabling integration with other BTP services and custom applications.
• Custom Applications: Integration via REST APIs, allowing non-SAP systems to consume the AI capabilities.
• Python Applications: Use of the Python SDK for integrating services into Python-based development environments.
• Java Applications: Integration into Java-based projects using the provided Java SDK.

Alternatives

• Microsoft Dynamics 365 AI: AI capabilities embedded within Microsoft's suite of enterprise business applications for CRM and ERP.
• Oracle AI Apps: AI-driven features integrated into Oracle's cloud applications for finance, HR, supply chain, and customer experience.
• IBM Watson Discovery: An AI-powered search and text analytics engine that can extract insights from complex documents, serving similar document intelligence needs.

Getting started

To begin using SAP AI Business Services, developers typically provision the desired service on the SAP Business Technology Platform (BTP) and then interact with it via its REST API or a provided SDK. The following Python example demonstrates a basic interaction with a hypothetical document processing service, assuming authentication and service endpoint setup are complete.

import requests
import json

# Replace with your actual service URL and API key/token
SERVICE_URL = "https://api.sap.com/aiservices/document-information-extraction/v2"
API_KEY = "YOUR_SERVICE_KEY"

def extract_document_info(document_path):
    headers = {
        "Authorization": f"Bearer {API_KEY}",
        "Content-Type": "application/json"
    }
    
    # In a real scenario, you'd upload the document data (e.g., base64 encoded PDF)
    # This example simulates sending a document ID or reference.
    payload = {
        "client_id": "my_client_id",
        "document_type": "invoice",
        "document": {
            "id": "doc_12345",
            "uri": "file:///path/to/invoice.pdf" # Placeholder for actual document data upload
        }
    }
    
    try:
        response = requests.post(f"{SERVICE_URL}/document/extraction", headers=headers, data=json.dumps(payload))
        response.raise_for_status() # Raise an exception for HTTP errors
        return response.json()
    except requests.exceptions.HTTPError as http_err:
        print(f"HTTP error occurred: {http_err}")
        print(f"Response: {response.text}")
    except Exception as err:
        print(f"An error occurred: {err}")
    return None

if __name__ == "__main__":
    # Example usage
    document_to_process = "./my_invoice.pdf" # Placeholder path
    print(f"Attempting to extract information from {document_to_process}...")
    extracted_data = extract_document_info(document_to_process)
    
    if extracted_data:
        print("Successfully extracted information:")
        print(json.dumps(extracted_data, indent=2))
    else:
        print("Failed to extract document information.")

This Python code snippet illustrates how to make a POST request to a hypothetical Document Information Extraction service endpoint. Developers would replace the placeholder URL and API key with their specific service instance details obtained from the SAP AI Business Services documentation, and implement actual document upload mechanisms if required by the specific service API.