Cross-posted. This article’s canonical home is iceberglakehouse.com. The Who, What, and Why of Semantic Layers: The Layer That Decides…
The real bottleneck with enterprise AI isn't retrieval augmentation or model accuracy : it's governance. Policy-as-code (OPA, ABAC, Rego) lets you enforce access rules, column masks, and row filters on any data source at query time.
Ask three teams in your company how they calculate revenue and you'll get three answers. Sales counts bookings. Finance counts recognized revenue. Marketing...
Most teams start building a semantic layer the wrong way: they open their BI tool, create a few calculated fields, and call it done. Six months later, three ...
Both terms appear in every modern data architecture diagram. They're used interchangeably in conference talks, Slack threads, and vendor marketing. And almos...
We already have a data catalog, so we don't need a semantic layer. This is one of the most common misconceptions in modern data architecture.
Your team builds an AI agent. It connects to your data warehouse. A product manager types What was revenue last quarter? and gets a number. The number is wrong.
Most organizations have a data governance policy. It lives in a Confluence page. It defines who owns what data, what terms mean, and who should have access. ...
Every data pipeline you build to move data from one system to another costs you three things: time to build it, money to run it, and freshness you lose while...
Every data team knows documentation is important. And almost every data team has a backlog of undocumented tables, unlabeled columns, and outdated descriptio...
Your organization uses Tableau for executive dashboards, Power BI for operational reports, and Python notebooks for data science. Revenue is defined in Table...
Semantic layers don't fail because the technology is wrong. They fail because of design decisions made in the first two weeks : choices that seem reasonabl...