Data context and governance
Semantic and context layers tell people and AI agents what the data means, MCP connects their tools, and Gateway decides what each of them may query.
Data context and governanceAn AI policy PDF has never stopped a query
An AI governance framework for data access sets out who, and which agents, can query which data, under which policy, and how every query is logged.
ReadAn MCP gateway is how you stop a server being a free-for-all
An MCP gateway is a control point in front of your warehouse. Each person and agent gets a role, a token and a log in place of a shared database password.
ReadCan we stop calling a chatbot on raw tables a company brain?
A company brain is a governed map of your company data that staff and AI agents can ask, built on the warehouse you already run. Here is the build order.
ReadEveryone's banging on about context, but what actually is it?
A context layer is a navigational map of systems, metric definitions and business meanings. The facts on a deal belong in the warehouse, away from memory files.
ReadHorizon and Unity Catalog still won’t tell you which agent ran the query
How Snowflake Horizon, Unity Catalog, BigQuery and Fabric with Purview control what AI agents can query, and what a gateway adds when agents share a login.
ReadHow far do Snowflake Semantic Views actually get you?
Snowflake semantic views keep metrics, dimensions and relationships in the database under normal RBAC. Agents still need a context layer and their own access.
ReadI wouldn't let an agent loose on a Databricks metric view and nothing else
Databricks metric views give you a semantic layer in Unity Catalog. Here is what they cover, and what agents still need: a role, row limits and a log.
ReadI'd stop treating a Power BI model like the company brain
A Power BI semantic model is built for Power BI reports. Keep shared metrics in a warehouse semantic layer so other tools and AI agents use the same definitions.
ReadIf dbt holds the metrics, who decides who can query them?
dbt keeps metrics in your project; Snowflake semantic views and Databricks metric views keep them in the engine. Neither alone decides which agent may query them.
ReadNobody needs the whole enterprise context layer on day one
Build an enterprise context layer one domain at a time, from a table inventory and agreed metrics to ownership, access and a gateway for agents.
ReadPlease stop giving agents the warehouse password
An MCP server exposes tools and data to AI clients. For a warehouse, run each call as a named agent, against certified metrics, with a log of every call.
ReadPointing Genie at raw tables is a fun way to get confident wrong answers
Genie answers well when metrics live in Unity Catalog metric views and each caller sees only their rows. This guide covers the setup, from metric views to access.
ReadSemantics, ontology, context... buzzword bingo, or business value?
Semantic layer, ontology, context layer and knowledge graph are four different things. An agent needs the first three to find the agreed number.
ReadThe metrics should not be hiding in the ingest tool
ELT loads data into the warehouse first and transforms it there, while ETL transforms it before load. When to use each, and where metric definitions belong.
ReadWhere do the metrics live once Fivetran is gone?
Inventory every connector and schedule, stand up the replacement ingest, run both until the row counts match, then cut over. Metrics belong in the warehouse.
ReadWho is Claude logged in as, once you've connected it to Snowflake?
Connect Claude, ChatGPT or Cursor to Snowflake, BigQuery or Databricks over MCP, with certified metrics, a role for each agent and a log of every query.
ReadWTF is a semantic layer anyway?
A semantic layer is a dictionary of definitions for business numbers. Each BI tool used to keep its own, in a format agents could not read. Apache Ossie opens it up.
ReadTalk to Rig
See your stack drawn as one diagram
An architect drafts a one-page diagram of your warehouse, sources and access path, then offers a 30-minute walkthrough.