SAP Certified Partner
SAP Business AI describes the strategic use of artificial intelligence within the SAP application landscape. It is not a separate product, but rather the direct integration of AI into existing SAP business processes and applications such as SAP S/4HANA, SAP SuccessFactors, and SAP Ariba.
The focus is on AI that understands a company’s business context. SAP Business AI does not operate in isolation from the systems but leverages real business data, process logic, and organizational structures. This results in concrete recommendations for action, forecasts, and automations that can be immediately applied in day-to-day operations.
SAP Business AI is an overarching concept that brings together all AI-powered functions within the SAP portfolio. The goal is to apply artificial intelligence directly where business decisions are made—in ERP, HR, finance, or supply chain processes.
Unlike generic AI solutions, the focus is not on general language or knowledge models, but on what is known as the business context. The AI understands how business processes work, which data is relevant, and what impact decisions have.
SAP is guided by three core principles:
The AI is directly integrated into work processes and delivers results exactly where they are needed.
The models are based on consistent, structured enterprise data and deliver business-ready results.
Data protection, transparency, and ethical use are at the heart of the AI strategy.
Distinction from Consumer AI
Unlike consumer AI such as chatbots or public language models, SAP Business AI works exclusively with internal company data within a secure environment. The AI is designed to support business processes and decision-making within the company, not for general conversation.
A key difference lies in what is known as “grounding”: While public models like ChatGPT are based on general internet data, SAP Business AI is connected to the company’s own data model, Customizing settings, and the specific company code context of the respective enterprise. The result is answers and recommendations that are not only technically correct but also valid for the specific company.
The technological foundation of SAP Business AI is the SAP AI Foundation, which is part of the SAP Business Technology Platform (BTP). It provides a central infrastructure for developing and operating AI models and seamlessly integrating them into SAP applications.
Among other things, the SAP AI Foundation enables:
A key advantage is its openness to technology partners. Through the BTP, SAP also integrates powerful models from leading providers such as NVIDIA, Microsoft, Google Cloud, and AWS. This allows companies to benefit from modern AI technology without having to leave their existing SAP landscape.
Generative AI Hub
As part of the AI Foundation, SAP provides the Generative AI Hub—a central interface that enables companies to orchestrate large language models (LLMs) from various providers. These include models from OpenAI (GPT-4), Anthropic (Claude), Google (Gemini), as well as open-source models such as Mistral or Llama. Crucially, SAP acts as a neutral orchestrator, while data protection is ensured through processing within the BTP infrastructure.
For companies that want to develop their own AI applications, SAP AI Core, as the underlying service, offers the ability to run their own machine learning pipelines—from data preparation and training to production deployment.
With Joule, SAP provides a generative AI assistant that serves as a natural language interface for SAP systems. Joule enables users to formulate complex queries and tasks in natural language.
Typical use cases include:
Joule not only provides information but can also prepare or trigger actions, such as creating reports, generating drafts, or initiating processes. In doing so, the assistant always draws on the company-specific context and existing SAP data.
Joule in Practice: Where Is the Assistant Available?
Joule is currently integrated into several SAP applications and is being continuously expanded:
Joule is configured on a role-based basis: What an employee is permitted to view and execute depends on their existing SAP authorizations. The AI cannot display information or trigger actions to which the user would not normally have access.
SAP Business AI can be used in virtually all key business areas. AI supports both operational tasks and strategic decisions and contributes to end-to-end process optimization. The following sections describe the most important areas along with specific use cases:
Finance & Controlling
Supply Chain & Procurement
Human Resources (SAP SuccessFactors)
Customer Service & Sales
IT & Development
Practical Note
Not every use case is immediately available out-of-the-box. SAP distinguishes between embedded AI scenarios (built directly into the application) and optional AI services, which must be activated separately and may be billed on a usage-based basis. A complete overview is available in the SAP AI Services Catalog on the SAP Help Portal.
The use of SAP Business AI offers companies several key benefits:
Routine tasks such as data entry, reconciliations, and analyses are automated. SAP reports internal time savings of up to 40% for standardized finance processes through automated posting proposals and cash application functions.
Forecasts and recommendations are based on data analysis rather than subjective assessments. Executives receive condensed, context-relevant metrics instead of raw data exports.
The operation of complex ERP systems is simplified through natural language and assistance functions. Employees with less SAP experience, in particular, benefit from Joule’s guided navigation through processes.
New AI scenarios are continuously delivered via cloud updates. SAP S/4HANA Public Cloud customers receive new AI features without separate upgrade projects—they are automatically available in the next quarterly release.
Since AI is embedded directly into SAP processes, there is no need to feed external AI tools with manually exported data. The AI operates on the same data model as the operational applications—no data loss due to CSV exports, no versioning issues.
The licensing model for SAP Business AI is one of the most complex topics surrounding the use of AI in the SAP environment—and at the same time one of the most discussed. This is because SAP Business AI is not a single license that you purchase. Costs arise at multiple levels and depend heavily on which functions are used and how intensively.
Three types of AI functions – three cost models
SAP generally distinguishes between three categories:
Many AI features are directly included in SAP cloud solutions and can be used at no additional cost—provided you have the corresponding cloud license. Examples: automatic account assignment suggestions in SAP S/4HANA Cloud, intelligent text recognition for incoming invoices, and learning recommendations in SAP SuccessFactors.
These features do not require AI Units and are not billed separately. They are an integral part of the respective product subscription.
For more performance-intensive or optional AI services—particularly those based on large language models (LLMs)—SAP uses a usage-based model. The billing unit is called an AI Unit (AIU).
Certain premium features—such as advanced Joule capabilities, access to external LLMs via the Generative AI Hub, or industry-specific AI packages—can be added as add-ons or through separate contractual agreements.
What are AI Units (AIU)?
AI Units are the primary billing unit for AI services on the SAP Business Technology Platform (BTP). They function similarly to prepaid credit: you purchase a quota of AI Units, which are consumed when using AI-powered services.
The number of AI Units a specific operation costs depends on the service in question:
Important for Budget Planning
SAP publishes an overview of individual AI services and their respective usage rates in the AI Services Catalog. This list is updated regularly. For a reliable cost estimate, it is therefore always advisable to consult with SAP or a certified SAP partner such as UNIORG, who can compare the planned use cases with the current catalog.
Purchasing AI Units: CPEA vs. Pay-As-You-Go
AI Units are purchased through SAP BTP—either as part of an existing BTP contract or separately. There are two main models available:
Model | Characteristics |
CPEA (Cloud Platform Enterprise Agreement) | Pre-purchase of a BTP Credit Pool (in euros or dollars). AI units are drawn from this pool. Lower cost per unit, but requires a minimum volume commitment. Suitable for companies with predictable, large-scale AI deployments. |
Pay-As-You-Go (PAYG) | No upfront commitment; billing based on actual usage. Higher per-unit costs, but maximum flexibility. Ideal for pilot projects or occasional AI use. |
BTP Free Tier | Limited quotas for certain BTP services are available free of charge—primarily intended for developers and testing purposes, not for production use. |
What exactly does it cost to get started?
SAP does not publish official list prices for AI Units—actual costs vary depending on the contract model, volume, term, and negotiating position. However, the following guidelines can help provide some context:
Recommendation
Transparent cost calculation requires that the planned AI use cases be clearly defined. UNIORG provides support with the needs analysis: Which AI features are already included? Where do additional costs arise? What is the required annual AIU quota? These questions are best addressed in a structured discovery workshop.
Joule: Note the separate licensing requirements
Joule is already included in certain SAP Cloud solutions (e.g., in SAP S/4HANA Cloud Public Edition starting from a specific edition), whereas in other products it is a paid add-on. The key factor here is not only access to Joule, but also the scope of Joule’s capabilities—that is, which process steps the assistant is actually permitted to perform. When reviewing licenses, companies should distinguish between:
A central component of SAP Business AI is the responsible handling of data. SAP ensures that company data is not used to train public or third-party AI models. Processing takes place exclusively within the respective customer instance.
Furthermore, SAP adheres to regulatory requirements such as the EU AI Act. Transparency, traceability, and data protection are integral parts of the AI architecture and governance.
EU AI Act and SAP
With the phased implementation of the EU AI Act starting in 2024/2025, the classification of AI systems by risk level is coming into focus. SAP has announced that it will classify all Business AI scenarios accordingly and document them transparently. For user companies, this means: SAP provides compliance documentation at the system level—but corporate governance (who uses AI for what purpose, with what data) remains the responsibility of each company itself.
Data Isolation and Tenant Segregation
A common concern regarding AI in the cloud is whether a customer’s own data might be used to train other customers’ AI models. SAP has clearly ruled this out: Each customer instance is isolated; data remains within the respective tenant and is not used for cross-tenant model training. The same principle applies when using external LLMs via the Generative AI Hub—requests go through SAP as an intermediary, not directly to the LLM provider, and no logging occurs for training purposes.
SAP Business AI is the next logical step for SAP users who want to digitize their processes and establish data-driven decision-making. The tight integration of AI into business processes creates measurable added value that goes beyond traditional automation.
Rather than replacing people, SAP Business AI supports employees with intelligent analytics, forecasts, and assistance functions. This makes artificial intelligence an integral part of modern enterprise software.
The key to project success lies in proper classification: Which AI features are already included in existing licenses? Where do real additional costs arise? And which use cases deliver the fastest return on investment? These questions can only be answered on a case-by-case basis—based on a structured analysis of your own SAP landscape, process maturity, and strategic digitalization goals. UNIORG supports this process from the initial assessment to productive implementation.
No. It is typically used via a consumption-based model (AI Units) on the SAP Business Technology Platform. You are billed only for the AI capacity you actually use. However, many basic AI functions are already included in existing SAP Cloud licenses without requiring additional AI Units.
AI Units (AIU) are the consumption units for AI services on the SAP BTP. The number required depends on the planned use cases and their volume. Simple classifications consume few units; generative AI queries via Joule consume significantly more. An estimate of annual requirements can be prepared with an SAP partner based on the AI Services Catalog.
Joule is already included in certain SAP Cloud editions. In other products, it is a paid add-on. Additionally, depending on usage intensity, AI unit costs apply for generative queries. A reliable calculation requires reviewing the existing license agreement as well as the planned Joule use cases.
The deepest integration occurs in SAP S/4HANA Cloud. However, other SAP Cloud solutions can also use AI functions via SAP BTP.
SAP Business AI focuses on the specific business context, is seamlessly integrated into SAP processes, and processes company data within a secure environment, rather than providing general, context-free answers.
No. SAP contractually ensures that customer data is not used to train models for other customers or public models. Processing is performed separately for each client within the respective SAP instance.
These include, among others, automatic posting suggestions (Cash Application), intelligent invoice recognition, demand forecasting in procurement, and basic Joule functions. The scope of features varies depending on the edition (Essentials vs. Advanced) and is expanded with each quarterly release.
The Generative AI Hub is a central component of the SAP AI Foundation on the BTP. It enables access to and orchestration of large language models from various providers (OpenAI, Anthropic, Google, Mistral, etc.)—within the secure SAP infrastructure.
CPEA (Cloud Platform Enterprise Agreement) is a prepaid credit model for BTP services, including AI Units—it offers a lower cost per unit but requires a minimum purchase. Pay-As-You-Go provides maximum flexibility without any upfront commitment, but is more expensive per unit. For continuous, production-level operations, CPEA is generally more cost-effective.
