Home AI SAP Business AI & AI Units

SAP Business AI & AI Units

Visual Woman and robot look at each other at a desk, the robot is sitting

When it comes to AI in the SAP universe, there are basically two different approaches:

  • SAP Business AI: This includes the ready-made AI features and functions provided by SAP.
  • Proprietary AI applications on BTP: Here, customers develop their own applications using the SAP Business Technology Platform (BTP) and services such as SAP AI Core and SAP AI Launchpad.

This wiki briefly and concisely examines both approaches.

Table of contents

SAP Business AI is not a standalone product, but rather a collective term for AI functions that are deeply integrated into the SAP cloud portfolio. These functions support users in various business areas by:

  • Text summaries
  • content generation
  • Smart recommendations
  • Making predictions

Available features: Base vs. Premium

There are currently 351 Business AI features available, which are divided as follows:

  • 185 Base Features: According to SAP, these are already included in every cloud subscription at no additional cost and can be used without restriction.
  • 126 Premium features: There is a charge for using these advanced features.

The complete feature catalog can be viewed in the Discovery Center: Business AI Catalog

Payment for premium features: SAP AI Units

Premium features are billed via SAP AI Units. These function as a virtual currency that enables flexible access to AI functions.

The most important features of AI Units:

  • Cross-product use: They enable services to be purchased and used across products (e.g., SAP S/4HANA Cloud, SAP SuccessFactors, SAP Ariba) with a single currency.
  • Transparent consumption: Consumption is measured using clear business metrics—e.g., per request, user, or data record. The official overview per function can be found here: Price List
  • Increased predictability: Companies can purchase packages of AI units and budget for their use in full.
  • Annual validity: AI Units are purchased annually and expire after 12 months if not used.

Important distinction

Billing via SAP AI Units applies exclusively to standardized Business AI functions. The use of SAP AI Core or SAP AI Launchpad for developing your own AI applications will continue to be billed via Capacity Units (CUs) or monthly fixed prices.

2. Use case: Why do I need SAP AI Core and the SAP AI Launchpad?

A typical use case for these tools is the development of an internal chatbot that answers HR inquiries, for example. Such inquiries may relate to the number of remaining vacation days or requests for employment references.

The aim is to reduce the workload on employees, shorten response times, and simplify access to knowledge. The basis for answering such questions can often be found in company documents such as HR guidelines, FAQs, and contract documents.

The solution: The interplay between SAP AI Core and SAP AI Launchpad

Both components are needed to implement such an AI project, and they perform different tasks:

  • SAP AI Core serves as the technical runtime environment for the AI model.
  • SAP AI Launchpad is the central management and control interface for the entire AI lifecycle.
  1. SAP AI Core: Technical Implementation

The development of the chatbot with SAP AI Core involves the following steps:

  1. Model development:
    Data scientists develop a retrieval-augmented generation (RAG) application based on LLMs. The texts and HR documents are vectorized to enable context-based responses.
  2. Containerization:
    The model—including data and vectorization logic—is bundled in a Docker container. This means that all dependencies (e.g., Python libraries, embedding models) are clearly defined and reproducible.
  3. Deployment & scaling:
    SAP AI Core is used to roll out the container on the BTP infrastructure.
    The platform ensures that the model scales automatically, for example, when many employees use the chatbot at the same time.
  4. Automation & Monitoring:
    Argo Workflows automate recurring tasks (e.g., retraining, index updates for new documents). AI Core monitoring functions capture response times, error rates, and model quality in real time.
  5. Integration into business processes:
    The model is provided as a web service (API endpoint) and can be integrated directly into SAP-based systems such as SuccessFactors or Workzone—for seamless access by employees.
  1. SAP AI Launchpad: The central control level

While AI Core ensures technical execution, the SAP AI Launchpad provides the necessary interface for management and control.

Specifically, the Launchpad offers the following advantages:

  • Transparency & control:
    The launchpad can be used to manage the entire lifecycle of the chatbot—from models and deployments to runtimes, metrics, and versions. This allows decision-makers and project managers to keep track of performance, costs, and usage.
  • Easy monitoring:
    Performance indicators such as response times, token consumption, or error rates can be viewed directly in the UI—without having to delve deep into technical logs.
  • Low-code/no-code prototyping:
    Departments (e.g., HR or IT business partners) can use the launchpad to create prototypes, customize prompts, or upload documents for RAG scenarios without any programming knowledge.
  • Integration with Generative AI Hub:
    Various foundation models (e.g., GPT-5, Claude, Gemini) can be tested via the model library. This allows you to evaluate which model provides the best answers for HR-specific contexts.
  • Orchestration & workflow design:
    In the "Orchestration" section, complex processes such as anonymization of sensitive personal data or translation can be configured, tested, and integrated directly into the chatbot workflow with a single click.

Sample cost calculation

To illustrate the cost structure, the following scenario is assumed for the HR chatbot: 1,000 inquiries per month.

The costs for such a project consist of two main components: the usage-based costs for SAP AI Core and the fixed price for the SAP AI Launchpad.

  1. Variable costs: SAP AI Core (billed in capacity units)

The costs for operating on SAP AI Core are consumption-based and are measured in capacity units (CUs). They consist of the following three factors:

component

Description

Example consumption per month

Conversion to CUs

GenAI Tokens

Use of the model (input + output)

Approximately 2.94 million GenAI tokens

6.6 CUs

Storage

Storage and access to HR documents

1 GB per day (over 30 days)

 6.2 CUs

Retrieval (RAG)

Document search & context extraction for the answers

5 text blocks per request × 1,000 requests

6 CUs

Total costs (SAP AI Core)

approx. 18.8 CUs/month

(Note: The conversion values are based on SAP Note 3437766 and current AI Corecost rates in the Discovery Center.)

  1. Fixed costs: SAP AI Launchpad

Regardless of the actual use of the chatbot, a monthly flat fee is charged for the SAP AI Launchpad. This fee covers the provision of the central management and monitoring interface.

  • Monthly fixed price: approx. €580

Summary of costs

The total monthly cost for this AI project is therefore made up of the variable, usage-based costs for SAP AI Core (billed in CUs) and the monthly fixed price for the SAP AI Launchpad.

3. SAP AI Core: The technical backbone for AI applications

SAP AI Core is a runtime environment for AI/ML applications that is offered as a BTP service. It supports the entire lifecycle management with the goal of transforming executable code from AI/ML applications into a robust, scalable, and maintainable application. Its main goal is to transform code into a robust, scalable, and maintainable application. Technically speaking, AI Core orchestrates a variety of proven open-source technologies such as Kubernetes, Docker, and Argo Workflows.

Interaction with the AI Core takes place programmatically via a set of REST APIs. These APIs are used to perform all operations, such as creating and managing model deployments, resource groups, metrics, and much more. Details of these APIs can be found here: https://api.sap.com/api/AI_CORE_API/overview

In addition, developers with Java/JavaScript expertise will have the opportunity to develop their AI applications using the SAP Cloud SDK for AI, and SAP is making the SAP Cloud SDK for AI (Python) – generative available for the Generative AI Hub.

Key advantages:

  • Framework-agnostic: Independence in the choice of ML framework (e.g., TensorFlow, PyTorch)
  • Automation & scalability: End-to-end management of training, deployment, and monitoring—including auto-scaling.
  • Governance & Security: Adherence to corporate guidelines and compliance requirements.
  • Seamless integration: Direct integration into SAP applications on the BTP.
  • Generative AI Hub: Access to leading foundation models (OpenAI, Google, Mistral, Azure, etc.) via a unified interface.

Billing: SAP AI Core is billed using capacity units (CUs).

  • Predictive AI: Costs are based on compute (node/hour) and storage usage (GB/hour).
  • Generative AI: Costs are based on the consumption of GenAI tokens, which are converted into CUs (see SAP Note 3437766).
  • Free accessis available via the Free Tier (however, this does not include the Generative AI Hub).
 

4. SAP AI Launchpad: The central control center of the AI project

The AI Launchpad is the user interface and control center for AI Core and Generative AI Hub. It enables the creation, monitoring, and management of AI projects in an intuitive interface—even for non-developers.

Key advantages:

  • Overview & control: Centralized management of models, deployments, and metrics.
  • Low-code/no-code: Enables rapid prototyping without in-depth programming knowledge.
  • Access to the Generative AI Hub: Easy selection and evaluation of the right language model for each use case. Configurable workflows in the Generative AI Hub: Prompt engineering, RAG (retrieval-augmented generation), anonymization/pseudonymization, translation.

The AI Launchpad is basically the recommended way to interact with the AI Core and the Generative AI Hub. It makes work easier and enables the rapid development of a working prototype. However, programming and the use of SDKs are essential for integration into existing applications or more complex workflows.

5. Conclusion: Tools are only part of the equation

The rise of artificial intelligence opens up enormous opportunities for companies. In this dynamic environment, tools such as SAP AI Core and the AI Launchpad are crucial enablers for not only recognizing AI potential, but also actively shaping it.

Strategic classification is a key success factor in this regard:

  • For companies that are already deeply embedded in SAP BTP, these tools represent a logical and seamless extension to drive their own AI innovations.
  • At the same time, it is crucial for every company to choose the AI portfolio that best fits its own cloud strategy. By 2025, all major providers will have powerful AI offerings that need to be evaluated.

But the true power of technology only unfolds in the right hands. The best tools are a catalyst for what really matters: the right application and the necessary expertise.

The starting point for any successful AI project is therefore not the tool alone, but a clearly defined use case that creates tangible added value and solves a real bottleneck.

Success is ultimately the result of powerful synergy: a clear business case, the right technological expertise, and the right tools to make this vision a reality.

Want to learn more about AI Units & Co. and find out how to successfully launch your AI project?

 Then book an appointment with our SAP AI expert Robert Kehrli here.

 

 

Published by:

Jannis Schulz

Professional SAP Analytics consultant
author

How did you like the article?

How helpful was this post?

Click on a star to rate!

Average rating 5 / 5.
Number of ratings: 4

No votes so far! Be the first person to rate this post!

INFORMATION

More information

Cover_Photo_SAC_AI_ML_Features_at_a_glance

SAC AI features explained: Joule, Just Ask, and Smart Predict

This wiki explains how to use Smart Predict to create automated forecasting models...
Visual Databricks and BDC Wiki

What is Databricks? What is the BDC? The ultimate guide to the perfect combination!

In today's data-driven business world, the ability to efficiently analyze and use large amounts of data is crucial for...
Your guide to successful SAC migration

SAC Migration by Q2 2026: The Guide to Transitioning to the Optimized Story Experience

The time for the conversion of SAP Analytics Cloud (SAC)...
Hands with three stars representing the different technologies: SAP Analytics Cloud, SAP Business Data Cloud, SAP Datasphere.

Feature update for SAP Business Data Cloud, Analytics Cloud, and Datasphere

This wiki article summarizes the most important content of the webinar on the topic:...
Lord of the Rings association with connection to SQL and dbt as fighters.

SQL and dbt: The future of modern data transformation

The article describes data processing in companies. Both...
9.1 Differences between SAP Databricks and native Databricks

SAP Databricks vs. Native Databricks: The detailed comparison for your company

In today's data-driven business world, the ability to efficiently analyze and use large amounts of data is crucial for...
Wiki What is Artificial Intelligence (AI) (2)

What is a semantic layer? Definition, benefits and role in modern data architectures

This Wiki article explains what a semantic layer is and why...