Generative AI Hub – Generative Artificial Intelligence in the SAP Environment

In the face of the AI revolution, SAP focuses on comprehensive innovation, treating artificial intelligence as a strategic necessity for every modern organization. The company integrates AI directly into its business solutions through AI Foundation – a multi-layered architecture that forms the foundation of the entire SAP Business AI concept. AI Foundation consolidates proven SAP components and services, making them available through a single centralized entry point – the Unified API Portal.

Key components of SAP AI Foundation

• Joule Studio – a low-code/no-code environment for building custom AI agents,
• SAP Document AI – a solution for automatic processing of data from unstructured business documents,
• Generative AI Hub – and this is the component we will focus on.

Generative AI Hub is not a separate, standalone service, but a key element of SAP AI Core, a service available within the SAP Business Technology Platform. SAP AI Core is a tool designed for efficient AI resource management in a standardized, scalable, and cloud-provider-independent way. It enables control over the full lifecycle of AI scenarios and serves as an engine in which custom solutions can be deployed directly into the SAP environment.

If we want to leverage the potential of generative AI within SAP architecture, Generative AI Hub is the ideal tool. It orchestrates the entire information flow: controls engine execution, optimizes computational resource usage, and manages AI processes, enabling safe and responsible use of LLMs.

Key features of the solution

• Grounding – Generative AI Hub allows LLM models to be embedded in the company’s business context. Thanks to integration with, for example, the SAP HANA Vector Engine, secure access to selected enterprise data can be ensured. As a result, AI responses are not general but precisely tailored to the organizational context.
• Model Agnostic – Using Generative AI Hub, organizations are not dependent on a single technology provider. It provides access to many leading models on the market (e.g., from OpenAI, Google, or Anthropic). They can be freely tested, tuned, and switched between without rebuilding the entire application.
• Security – SAP guarantees full data privacy and isolation. Company data is not used to train public algorithms nor shared with other customers. The service includes built-in content filters, sensitive data masking mechanisms, and protection against cyberattacks (e.g., prompt injection).

Depending on technical and licensing needs, Generative AI Hub offers three ways to use models:

• SAP Managed Models – ready access to tools from external providers (e.g., GPT, Gemini, Claude) managed by SAP. No need for separate licenses, billing, or version updates.
• SAP Hosted Models – solutions hosted directly on SAP infrastructure, ensuring high performance, stability, and full integration with SAP systems.
• Custom Models – the ability to deploy and manage your own or customized open-source models.

Generative AI enables the creation of custom processes and automation. In SAP environments, interactions with Gen AI models can be divided into four main categories:

• Human-to-AI – a user enters a prompt, and the LLM generates a text response or dataset intended for human consumption.
• Software-to-AI – a system automatically sends input data to the model, which then generates summaries or analyses for the user.
• Human-to-Software – a developer provides instructions, and the model generates structured code or output ready for direct system processing.
• Software-to-Software – fully automated communication, where one application sends input to an LLM, and the generated output is immediately processed by another system or software component.

Access to the platform is configured through an SAP BTP account. The SAP AI Core service in the Extended plan (which includes Generative AI Hub functionalities) does not involve any additional fixed fee. When using SAP-managed and hosted tools, you only pay for actual usage-namely, for the number of processed tokens (which consist of the model input and output) and for selected orchestration features (such as data masking, grounding, or content filtering). The lack of a fixed entry fee is a major advantage of this solution. To estimate costs, it is recommended to use the dedicated calculator provided by SAP:

https://ai-core-calculator.cfapps.eu10.hana.ondemand.com/

Model management in practice can be carried out in two ways:

• Through the SAP AI Launchpad graphical interface, a SaaS application available on SAP BTP. It integrates with an SAP AI Core instance and enables intuitive management of AI resources. To use the full functionality of Generative AI Hub in SAP AI Launchpad, the Standard plan is required, which involves a monthly subscription fee on BTP.

• Direct, programmatic management via API requests (available in the documentation and on SAP Business Accelerator Hub). This option is completely free.

From my experience, I can say that a good and cost-effective approach is to combine both methods. You can use SAP AI Launchpad in the free plan for managing basic configuration and deploying models in SAP AI Core, while performing all advanced operations and Generative AI Hub functionalities directly through free APIs.

Generative AI Hub becomes a central AI hub that can be utilized by all systems within an organization. It can be easily integrated with key SAP tools as well as external applications:

• SAP Integration Suite – enables secure and structured connectivity with Generative AI Hub functionalities via a dedicated AI Adapter. Within Cloud Integration (CPI), generative models can be embedded into record flows between any systems across the enterprise.
• SAP Build Process Automation – enhances the efficiency of workflows and RPA bots. Automated processes can include steps that send requests to SAP AI Core, and the model’s responses (e.g., decisions or extracted data) can be used in subsequent stages of the process.
• Custom applications – thanks to open APIs and SDKs, developers can embed Generative AI Hub capabilities into their own applications – both those built on BTP (e.g., SAP Build Apps or SAP Business Application Studio) and fully external systems.

Generative AI Hub is a breakthrough in SAP’s approach to artificial intelligence. Instead of building complex and risky integrations with external public tools, companies receive a ready-to-use, secure, and standardized architecture. The greatest value of this solution lies in combining flexibility (the ability to choose any LLM model) with rigorous security and the unique business context of SAP data. This makes it possible to transform raw information into intelligent actions, optimizing business processes across the organization.

If you want to test these capabilities in practice, SAP provides a free trial that allows you to explore Generative AI Hub directly within the SAP AI Launchpad environment:

https://www.sap.com/products/artificial-intelligence/generative-ai-hub-trial.html