Technical writing on AI implementation, Microsoft Fabric, and enterprise analytics.
What it actually takes to expose a SQL Server data warehouse to Claude via the Model Context Protocol — authentication, query scoping, and keeping responses grounded in schema.
Key decisions in bronze/silver/gold layer design that affect downstream analytics performance and maintainability — and the ones that are easy to get wrong early.
A practical overview of the deployment process, managed identity authentication, and the integration points that differ from direct Anthropic API usage.
Dynamic RLS using username() and USERPRINCIPALNAME() works well at small scale. Here is what changes when you have hundreds of roles, large semantic models, and DirectQuery sources.
