Structural Metadata vs Semantic Metadata for AI Consumers
AI agents can't fill in the meaning that human analysts guess from context.
Yuki Novak
Contributing Editor
Yuki Novak
Yuki Novak covers semantic layer for ai, ai-ready data layer and agentic data access for AI-ready data.
13 stories
Descriptive Metadata as AI Agent Context
Stale or missing metadata makes AI agents generate confident wrong answers with no warning signal.
Yuki Novak
OLTP vs OLAP for AI Agent Data Sources
Yuki Novak
Revoking AI Agent Data Access at Workload Granularity
Limit AI agent credentials to single tasks instead of standing access to prevent breaches.
Yuki Novak
Federated Data Access for AI Workloads
Federated queries fail silently when schemas drift unless a semantic layer governs what data means.
Yuki Novak
Institutional Memory Gap in AI Agents
Yuki Novak
AI Agent Read vs Write Permission Asymmetry
Yuki Novak
Service Account Risks in Agentic Pipelines
Agents inheriting overpermissioned service accounts compound risk across multi-step workflows.
Yuki Novak
Knowledge Graph Integration for AI Data Retrieval
Knowledge graphs make AI retrieval accurate by encoding what data means, not just how it is stored.
Yuki Novak
Runtime Data Access vs Static Pipeline Delivery
Agents need fresh data at query time, not stale snapshots—a correctness problem, not a speed one.
Yuki Novak
Semantic Layer Implementation on Snowflake
A governed, queryable object that replaces scattered definitions with one source of truth.
Yuki Novak
What a Semantic Layer Is in a Data Warehouse
Yuki Novak
How AI Agents Query Structured Data
Semantic layers, not raw schemas, let AI agents query data reliably without hallucinating.
Yuki Novak