Notifications
Clear all
Topic starter
24/06/2026 7:26 pm
TL;DR: As autonomous systems begin making business decisions from enterprise definitions, semantic fragmentation turns into operational risk, according to Collibra and Gartner. The issue is no longer readability for humans, but reliable machine-interpretable context that AI can use without compounding bad assumptions into bad decisions.
NHIMG editorial — based on content published by Collibra: The semantic layer has always been essential. Now it's existential
By the numbers:
- By 2027, organizations that prioritize semantics in AI-ready data will increase their agentic AI accuracy by up to 80% and reduce cost by up to 60%.
- Only 5.7% of organisations have full visibility into their service accounts.
- NHIs outnumber human identities by 25x to 50x in modern enterprises.
Questions worth separating out
Q: How should teams govern semantic layers for agentic AI systems?
A: Teams should govern semantic layers as authoritative decision infrastructure, not as a reporting convenience.
Q: Why does semantic fragmentation create risk for autonomous systems?
A: Semantic fragmentation is risky because autonomous systems do not resolve contradictory definitions the way humans do.
Q: How can security and data teams tell if semantic governance is working?
A: Semantic governance is working when the same business term produces the same outcome across systems and when every AI decision can be traced to one governed definition set.
Practitioner guidance
- Define a single authoritative semantic source Map the business terms that drive high-impact decisions, then assign one governed definition set for each term across BI, analytics, and AI workflows.
- Trace AI decisions back to context provenance Require every agentic workflow to show which definition set, policy rule, and data source informed the decision.
- Test for semantic fragmentation before scaling agents Compare definitions across platforms and flag any term that produces different outcomes in different systems.
What's in the full article
Collibra’s full post covers the operational detail this analysis intentionally leaves for the source:
- How the semantic mapping workflow links physical data to business definitions at scale.
- How model generation incorporates governance rules, metric catalogs, and organisational context.
- How OSI and MCP are used to push governed context into execution platforms and autonomous agents.
- Which data quality, minimisation, and access controls travel with that context in practice.
👉 Read Collibra’s analysis of the semantic layer for agentic AI governance →
Semantic layer governance for agentic AI: are your controls ready?
Explore further
25/06/2026 4:43 am
Semantic consistency is becoming an identity governance dependency, not a data quality preference. Once autonomous systems begin acting on enterprise definitions, inconsistent meaning becomes a governance failure because machines do not reconcile ambiguity the way people do. The old BI assumption that a human can compensate for imperfect context no longer holds. Practitioners should treat the semantic layer as part of the control plane for AI-enabled decisioning.
A few things that frame the scale:
- NHIs outnumber human identities by 25x to 50x in modern enterprises, according to the Ultimate Guide to NHIs.
- 96% of organisations store secrets outside of secrets managers in vulnerable locations including code, config files, and CI/CD tools, according to the Ultimate Guide to NHIs.
A question worth separating out:
Q: What should organisations do before letting AI agents act on business data?
A: Organisations should verify that the agent receives governed context, not just raw data or local metadata. That includes definition provenance, policy inheritance, and usage conditions. If those elements are unclear, the agent should be constrained to assistive use rather than autonomous execution.
👉 Read our full editorial: Semantic layers are becoming identity infrastructure for agentic AI
