Topic starter
23/06/2026 7:00 pm
TL;DR: Governed context improved agent accuracy from 62% to 92% in a KU Leuven test, while ungoverned agents answered confidently wrong on definition-dependent questions and burned more compute, according to Collibra. The deeper issue is assumption collapse: access control and data governance models built for fixed truth sources fail when agents reason across fragmented ontologies at runtime.
Editorial analysis by NHI Mgmt Group, based on content published by Collibra: “Data has gravity. Context does too.”.
Key questions
Q: How should teams govern AI agents that rely on business context from data platforms?
A: They should treat business context as a control input, not a convenience layer.
Q: Why do AI agents produce wrong answers when ontology definitions differ across platforms?
A: Because the agent must infer meaning from incompatible local definitions before it can answer the question.
Q: What are the signs that context governance is missing from an enterprise AI program?
A: Common signs include agents using inconsistent definitions, acting on outdated policy, bypassing semantic standards, or producing different answers from the same data.
Practitioner guidance
- Map the enterprise ontology boundary Identify where core business terms diverge across CRM, ERP, lakehouse, warehouse, and collaboration systems, then assign an authoritative definition source for each high-value term.
- Separate local catalogs from governed context Keep platform-specific catalogs for storage and discovery, but route agent reasoning through a neutral context layer that resolves definitions, lineage, and policy before execution.
- Audit agent dependencies on semantic context Review which agent workflows depend on definitions such as customer, active user, or revenue, and block production use where the governing ontology is ambiguous or duplicated.
Bottom line: Fragmented ontologies make AI agents reason across competing business definitions, which turns semantic inconsistency into a governance risk.
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
25/06/2026 2:45 am
Context gravity is becoming the new governance boundary for AI agents. When an organisation lets each platform define its own business ontology, it creates multiple valid truths that agents must reconcile at runtime. That is not a catalog inconvenience; it is a control failure that changes answer quality, operational decisions, and auditability. The practitioner conclusion is clear: the governed context layer now sits alongside identity as a core control surface for AI.
A question worth separating out:
Q: What is the difference between a data catalog and contextual data intelligence for AI governance?
A: A data catalog helps people find and describe data assets, while contextual data intelligence connects those assets to relationships, entitlements, privacy issues, and governance controls. For AI governance, the difference matters because teams need more than inventory. They need to understand how data moves, who can use it, and what risks emerge when AI systems consume it.
👉 Read our full editorial: Context gravity and AI agent governance across fragmented data estates