Knowledge graph - Scientific news
On October 27, 2026, at its annual TechEd conference in Berlin, SAP SE announced major expansions to the SAP Knowledge Graph and its semantic integration with the Joule enterprise AI ecosystem.
Addressing critical enterprise adoption bottlenecks—where Large Language Models (LLMs) and autonomous AI agents struggle with fragmented data silos, inconsistent business logic, and high token overhead—the expanded SAP Knowledge Graph acts as a centralized semantic foundation. By mapping entities, relationships, business rules, and operational processes across SAP applications and external third--party data environments, the platform provides business context to autonomous agents without requiring manual data migration.
The integration leverages open agent standards, including Anthropic's Model Context Protocol (MCP) and Agent2Agent (A2A) specifications, enabling Joule agents to retrieve verified corporate context dynamically. Within the newly announced Joule Studio developer environment, the knowledge graph automates intent--based discovery of APIs, data products, and tool servers. According to SAP benchmark metrics, grounding LLM agents in this structured enterprise graph accelerates custom agent workflow development from 13 hours to approximately 2.5 hours while lowering operational inference costs by up to 50% through reduced token consumption and eliminated hallucinations.
SAP announced that general availability for the expanded SAP Knowledge Graph across customer--specific enterprise environments is scheduled for the first quarter of 2027.