What is Semantic MDM?

Question

Accepted Answer

Semantic MDM (Semantic Master Data Management) applies the disciplines of enterprise MDM — golden records, entity survivorship, schema stewardship, and change governance — to the outward-facing semantic layer of a website or digital platform. Where traditional MDM governs master data inside the enterprise (customers, products, suppliers, employees), Semantic MDM governs the machine-readable entity definitions that AI agents, search engines, and agentic commerce systems encounter at the AI-facing boundary. The process runs in three stages. First, Audit: every entity definition across the domain is crawled, validated against Schema.org, and assessed for duplicate @id references, conflicting definitions, orphaned nodes, and semantic misalignment between declared meaning and operational intent. This is the diagnostic equivalent of a data quality assessment before MDM implementation. Second, Enrich: entity conflicts are resolved, canonical @id references are established, and survivorship rules are applied to produce a single authoritative definition per entity — the Golden Entity Record. Third, Publish: the validated entity layer is compiled into a Semantic MDM Data Catalog — a machine-readable, AI-traversable index of all golden entity records, continuously updated via change data capture to maintain alignment between operational reality and declared semantics in real time. The VISEON platform is powered by Qlik, using Qlik's Associative Engine for semantic relationship resolution and Qlik Replicate (CDC) for real-time replication from source systems — the same infrastructure that governs master data inside enterprise boundaries, now extended to the AI-facing semantic boundary. The parallel to traditional MDM is direct: just as MDM eliminated duplicate customer records across ERP and CRM systems, Semantic MDM eliminates duplicate entity definitions across the knowledge graph, ensuring that every AI agent querying your domain retrieves the same, authoritative, semantically aligned answer.

Frequently Asked Questions