Concept card

SAP Knowledge Graph

As of 2026-07-24T14:00:00Z

Edited by Cédric-Olivier Mâry, sole editor and publication director — 27 years in the SAP ecosystem. How these pages are produced, and where automation is used.

What is SAP Knowledge Graph?

The Knowledge Graph stores meaning, not facts — it augments Datasphere rather than replacing it, because graphs excel at multi-hop entity traversal while warehouses excel at aggregation and time-series.

What it is

The SAP Knowledge Graph is the semantic substrate that lets AI reason about how a business is actually wired, rather than just what happened in it. It is not a separate database sitting off to the side; it is a graph-shaped view over the business entities — customer, material, order, cost centre, employee, vendor, plant, sales organization — and the relationships between them, drawn from data that already lives across S/4HANA, SuccessFactors, Ariba, Concur, and Datasphere. Its job is to hand AI surfaces, Joule first and partner-built agents next, a map of meaning: not just "here is a table of sales orders" but "this entity is a customer, this relationship is a reports-to link, this attribute rolls a cost centre up into its parent."

Why it matters

A query like 'open POs with vendors that are also customers in segment X' needs custom multi-table SQL joins in a warehouse but is a native traversal in a knowledge graph.
The five entity classes (business partners, materials, documents, org structure, master data) each project from S/4 or another SAP app — the graph is a view, not a separate database to maintain.
Positioning the Knowledge Graph as a Datasphere replacement misreads the architecture — it's the semantic layer that gives Joule and partner agents a queryable map, aggregation still lives in Datasphere.

Key points

Semantic-AI substrate connecting business entities + relationships across the SAP estate.
Five entity classes: Business Partners · Materials · Documents · Org Structure · Master Data.
GA Q1 2026 integration with Datasphere + BDC + Joule (per the Analytics Legends research ledger).
Joule uses graph traversal to identify the right Analytic Model for a question.
Master per entity class: S/4 for customer/vendor/material; SuccessFactors for employee; Ariba for procurement.
Hierarchies modeled as native graph relationships, not separate dim tables (10× faster).
DAC propagates from Datasphere onto graph queries — single source of truth.
Coverage phasing: P1 (BP + Materials + Documents) ~ 4-6 months; full 5-class ~ 12 months.
SAP Knowledge Graph is mastered only when it changes a named buyer decision.
Start with the semantic contract and control model before demonstrating the tool.

Terms used on this page

Knowledge Graph: SAP's semantic-AI substrate — graph of business entities + relationships across the SAP estate.
Entity class: Category of business object: Business Partner, Material, Document, Organisational Structure, Master Data.
Graph traversal: Multi-hop query pattern that follows entity-relationship edges. Faster than SQL joins for semantic discovery.
Master per entity class: Pattern of designating one application as the authoritative source for each entity class (S/4 for customer, etc.).
AM mapping (graph): Metadata declaring which Datasphere Analytic Model exposes which Knowledge Graph entity. Drives Joule's AM discovery.
Industry Cloud verticals: Domain-specific entity extensions (utilities, life-sciences, retail, automotive, public sector). Phased GA 2026-2027.
Coverage phasing: Sequenced rollout — phase 1 BP + Materials + Documents (4-6 months), phase 2 OrgStruct + MD (full 12 months).
DAC propagation onto graph: Pattern where graph queries inherit row/column security from Datasphere catalog DAC; no duplicate ACL state.

Sources

Full card available to members. What the full card adds: the full decision framework · the SAP vs Snowflake / Databricks / Fabric comparison · the common pitfalls and their fix · the cheat sheet · the architecture schemas · the code blocks · the facts worth quoting.

Open in the app →