ServiceNow Podcasts Your data should speak your language with Säde Haveri
Feb 26, 2026
Säde Haveri, entrepreneur and metadata specialist who organizes Helsinki Data Week, argues metadata is the bridge between systems and people. She discusses why data must match organizational language. She covers defining shared meanings, prioritizing critical datasets, embedding definitions early, and using ontologies and knowledge graphs to make semantics machine readable.
AI Snips
Chapters
Transcript
Episode notes
Disconnected Modeling Creates Human Workarounds
- Säde notes companies often model at the physical database level and create separate conceptual models that aren't linked.
- That mismatch forces humans to manually map meanings into catalogs, making catalogs reliant on tacit expert knowledge.
Shift Left Requires Definitions First
- Shift-left requires defining meanings early so producers implement definitions into each system.
- With clear metadata you can automate verification across many systems using AI; without definitions you cannot.
Start With Your Most Important Data
- Start by identifying your most important data that drives decisions and fix those pipelines first.
- Model that data conceptually and make definitions machine readable so semantics can be automated downstream.