Introduction to SAP AI Core with Basic Configuration Steps

🌐 SAP AI Foundation – Full Overview

SAP AI Foundation is a comprehensive set of tools and services provided by SAP to help organizations build, run, scale, and manage AI applications securely and responsibly—fully integrated with business data and processes.

💡 Purpose:

To empower enterprises with tools for:

• Building custom AI models.

• Leveraging prebuilt AI services.

• Managing AI lifecycle.

• Embedding AI into business apps (like SAP S/4HANA, SAP SuccessFactors, etc.)

• Connecting with business context and data semantics.

🧩 Key Components of SAP AI Foundation

Let’s break down each building block

1. 🚀 SAP AI Core – The Engine Room

SAP AI Core is the execution layer. It runs AI workflows at scale using container-based infrastructure.

🔹 Key Features:

• Custom model training and deployment.

• Workflow automation with Argo Workflows.

• Integration with any ML framework (TensorFlow, PyTorch, etc.)

• Lifecycle monitoring and retraining.

2. 🎛 SAP AI Launchpad – The Control Tower

A UI-based management layer for SAP AI Core. It allows data scientists, developers, and business users to manage AI lifecycle operations visually.

🔹 What You Can Do:

• View executions, logs, and model performance.

• Monitor retraining workflows.

• Manage artifacts like datasets and models.

• Collaborate across teams.

3. 🤖 SAP AI Services (Prebuilt) – Plug-and-Play Intelligence

List of SAP AI Core Prebuilt AI Services — these are ready-to-use, plug-and-play intelligence services that can be integrated into SAP applications to enhance automation, decision-making, and user experience. These services are provided via SAP AI Business Services, typically through RESTful APIs, and are deployable in SAP BTP (Business Technology Platform).

🧠 SAP AI Business Services (Prebuilt AI Services)

These are ready-to-use, pre-trained AI services that address common business scenarios.

1. Document Information Extraction (DOX)

• Purpose: Extracts structured data from unstructured documents.

• Use Cases: Invoice processing, purchase orders, delivery notes.

• Highlights: OCR, key field extraction, multi-language support.

2. Business Entity Recognition (BER)

• Purpose: Extracts named entities from text.

• Use Cases: Contract parsing, data cleansing, text analytics.

• Highlights: NLP for identifying companies, materials, locations, dates.

3. Data Attribute Recommendation (DAR)

• Purpose: Recommends attribute values based on existing master data.

• Use Cases: Material/Vendor/Customer setup.

• Highlights: Predicts missing fields using historical data.

4. Service Ticket Intelligence (STI)

• Purpose: Classifies service tickets and suggests responses.

• Use Cases: Helpdesk automation, ticket routing.

• Highlights: NLP classification, multi-language support.

5. Product Text Classification (PTC)

• Purpose: Classifies product texts into categories.

• Use Cases: Procurement catalogs, spend analytics.

• Highlights: NLP-based classification using predefined or custom taxonomies.

6. Invoice Object Recommendation (IOR)

• Purpose: Suggests GL accounts, cost centers, and other accounting objects.

• Use Cases: Finance automation, invoice posting.

• Highlights: Trained on financial postings for contextual recommendations.

7. Personalized Recommendation (PRE) ✅

• Purpose: Delivers real-time personalized suggestions.

• Use Cases: Product recommendations, next-best-offers, frequently used materials.

• Highlights: Collaborative + content-based filtering, real-time scoring.

8. SAP Translation Hub

• Purpose: Automates and streamlines the translation of application texts into multiple languages.

• Use Cases: Multilingual Fiori apps, SAP S/4HANA extensions, global rollouts.

• Highlights:

• Integration: Easily orchestrated with SAP AI Core for human-in-the-loop, post-editing, or quality checks.

9. Customer Retention Predictor (CRP) (Legacy / Limited Use)

• Purpose: Predicts the probability of customer churn.

• Use Cases: Loyalty programs, SaaS renewals.

• Note: Often replaced by custom models in SAP AI Core or HANA ML.

⚙️ Integration Details

Summary

4. 🧠 SAP Generative AI Hub – Enterprise-Grade GenAI

The latest addition! A centralized access point to use Generative AI capabilities in a secure, business-context-aware way.

🔹 Features:

• Unified access to large language models (LLMs) from partners (like OpenAI, Aleph Alpha, etc.) via APIs.

• Integration with SAP data and metadata models for better grounding.

• Use cases: Chatbots, document summarization, code generation, contextual insights.

🔹 Key Value:

• Bring generative AI to enterprise workflows with governance and context.

• Avoid hallucinations by connecting with structured SAP data.

5. 🧾 Business Data and Context – The Foundation of Enterprise AI

💡 Think of this as the “knowledge graph” of your enterprise.

SAP systems like S/4HANA, SuccessFactors, Ariba, etc., generate rich, structured business data—invoices, customers, materials, POs, employees, etc.—and SAP provides metadata that explains the relationships between them.

🔹 What It Includes:

• Master Data (e.g., materials, vendors)

• Transactional Data (e.g., sales orders, invoices)

• Metadata & Semantics (e.g., relationships, hierarchies)

• Business Process Context (e.g., workflow steps, rules)

🔹 Why It Matters:

• Makes AI outputs context-aware and explainable

• Reduces hallucinations in generative AI

• Enables compliance, traceability, and trust

🧩📊 Integrated View – How All These Pieces Fit Together

Let’s club everything into a layered architecture for SAP AI Foundation:

🔷 1. Data & Business Context Layer

• SAP Data: SAP S/4HANA, SuccessFactors, Ariba, etc.

• Semantics: Business objects, data models, workflows.

➡️ Feeds into AI for contextual intelligence.

🔷 2. AI Capabilities Layer

🔷 3. Integration & Deployment Layer

• REST APIs and SDKs.

• Integration with SAP BTP and apps.

• Embed AI into SAP Fiori apps or side-by-side extensions.

🔷 4. Governance, Security, and Monitoring

• Role-based access control (RBAC).

• Lifecycle tracking and audit logs.

• Retraining triggers and performance monitoring.

🧠 Imagine This in Practice:

Let’s say you’re in procurement, and you want to automate invoice processing and improve supplier predictions.

You might use:

1. ✅ Prebuilt AI Service – Document Information Extraction for reading invoices.

2. ✅ Custom Model in AI Core – Predict late supplier delivery based on history.

3. ✅ GenAI Hub – Summarize supplier performance for decision-making.

4. ✅ Business Data Layer – Pull supplier master data and PO history from S/4HANA.

5. ✅ Launchpad – Manage and monitor all of this centrally.

🎯 Summary Table

📊 Final Summary Table: SAP AI Foundation Components

SAP AI Foundation Overview

Configure SAP AI Core on SAP Business Technology Platform (BTP)

Here are the steps to set up SAP AI Core on SAP Business Technology Platform (BTP):

Prerequisites:

Before setting up SAP AI Core, ensure you have:

1. SAP BTP Account – A valid SAP BTP account with the necessary entitlements.

2. SAP AI Core Service – The service must be enabled in your SAP BTP account.\

Step-by-Step Configuration

Step 1: Create a New Subaccount

1. Log in to the SAP BTP cockpit.

2. Navigate to the Subaccounts tab.

3. Click Create Subaccount.

4. Enter the required information, such as subaccount name and description.

5. Click Create.

Step 2: Enable the SAP AI Core Service

1. Navigate to the Services tab.

2. Search for SAP AI Core.

3. Click on the SAP AI Core tile.

4. Click Enable.

Step 3: Create a New Instance of SAP AI Core

1. Navigate to the Instances and Subscriptions tab.

2. Click Create Instance.

3. Select SAP AI Core as the service.

4. Choose the desired plan and click Create.

Step 4: Configure the SAP AI Core Instance

1. Navigate to the Instances and Subscriptions tab.

2. Find the newly created SAP AI Core instance.

3. Click on the three dots at the end of the row.

4. Select Go to Application.

5. Configure the instance as desired (e.g., create models, upload data).

Step 5: Assign Roles and Permissions

1. Navigate to the Security tab.

2. Click Role Collections.

3. Create a new role collection or assign an existing one.

4. Add the necessary roles and permissions for SAP AI Core.

Step 6: Test the SAP AI Core Instance

1. Navigate to the Instances and Subscriptions tab.

2. Find the SAP AI Core instance.

3. Click on the three dots at the end of the row.

4. Select Go to Application.

5. Test the instance by creating a new model or uploading data.

The SAP AI Core instance is now set up and ready to use. You can use it to build, train, and deploy machine learning models.

To dive deeper into this subject, please refer to the LinkedIn articles linked below.

Exploring SAP BTP: Key Components and Benefits

A Beginner's Guide to SAP BTP Configuration and SAP CPI Integration Steps

SAP CPI Front-End: A Deep Dive with an IDoc Integration Example

Integrating SAP S/4HANA with a Third-Party Logistics Provider

From Python to SAP AI Core: A Simple Path for Predictive Analytics in Banking (Explained for Non-Techies)