SAP introduced its Generative AI Hub in November 2023 at the SAP TechEd conference, announcing its general availability through SAP AI Core by December 2023. SAP AI Core is a cloud-based service in the SAP Business Technology Platform (SAP BTP). It gives companies a place to run, scale, and manage artificial intelligence and machine learning models. It connects custom AI code directly with enterprise business data.
LLM (Large Language Models (LLMs) in SAP:
• SAP does not build its own foundational LLM from scratch; instead, it provides an orchestration layer that connects enterprise data with leading third-party and open-source models (such as OpenAI, Anthropic, Google, Meta, and IBM Granite).
SAP provides access to and integrates Large Language Models (LLMs) through its Generative AI Hub in SAP AI Core on the SAP Business Technology Platform (BTP) via the following Capabilities:
- SAP Joule Agentic AI - is an autonomous artificial intelligence system built by SAP that evolves Joule from a basic conversational chatbot into an active, multi-step enterprise decision-maker. Instead of just answering questions, it plans, reasons, and executes complex business workflows across ERP, finance, supply chain, and human resources.
- In Non-Technical words - is a smart digital coworker for your company’s software. Instead of just answering questions like a basic chat bot, it can plan and finish multi-step work tasks all by itself.
- SAP RAG - in the SAP ecosystem, RAG stands for Retrieval-Augmented Generation. It is an artificial intelligence technique that connects large language models (LLMs) to a company's private business data. Instead of guessing an answer, the AI searches internal documents or databases first, then uses that real data to write an accurate response.
- In Non-Technical words - is an AI technique that lets an intelligent assistant look up your company's private files, invoices, or inventory data in systems like SAP HANA Cloud Vector Engine before answering a question, ensuring the AI gives accurate, fact-based business answers instead of guessing
- SAP MCP - refers to the use of the open-source Model Context Protocol (MCP) to connect Large Language Models (LLMs) and AI agents directly to SAP enterprise systems, data, and development tools. It provides a standardized plug-and-play architecture so AI assistants can securely query data, run business logic, or assist with development without needing custom point-to-point APIs for every single task. MCP is used to connect the User + AI Agent to the SAP system.
- In Non-Technical words - is like a universal USB-C port that connects AI smart assistants (like ChatGPT or Claude) to an enterprise's massive, complex SAP business system
- SAP Vector Database - The SAP vector database works through the SAP HANA Cloud vector engine, which stores and processes vector embeddings directly inside a standard enterprise database. In other words, a SAP vector database (built into SAP HANA Cloud) is a smart digital filing cabinet. It stores business data not just by matching exact words, but by understanding what the data means. It helps AI chatbots find answers from your company's private files. SAP HANA Cloud Vector Engine maintain this Vendor Database in SAP.
- In Non-Technical words - is like giving an AI assistant the ability to understand the "meaning" or "vibe" of your company's data instead of just searching for exact words. It serves as a secure, smart memory bank built directly into SAP's database system.
- SAP OData service - OData (Open Data Protocol) is a key SAP standard for connecting SAP systems to external applications, cloud services, and third-party tools. It uses standard web tech like HTTP, JSON, and XML so different systems can talk to each other easily.
- In Non-Technical words - is a secure, standard translator that lets an AI system talk to an SAP database. It acts like a smart menu. The AI asks for data in plain terms, and the service fetches only that information without breaking the database.
Click here for link to the Blog Posts on the above topic:
Click here for official links to SAP AI development and applications:
https://discovery-center.cloud.sap/serviceCatalog/sap-ai-core/
https://help.sap.com/docs/sap-ai-core/
https://learning.sap.com/courses/introduction-to-ai-core
(Disclaimer: I am NOT AI Expert but the above is something I summarized based on recent readings)
