Model Context Protocol (MCP) is one of the most important new standards in AI engineering. It defines how AI agents can safely interact with real systems—files, APIs, and databases—using structured tools instead of ad-hoc prompts.
In this session, we dive into how MCP works and demonstrate how it can be used to query, maintain, and even help build Azure SQL databases.

We explore the official MSSQL MCP Server, how it exposes schema, metadata, and controlled SQL execution, and how AI agents like ChatGPT or Azure AI Foundry can use it to:

- inspect tables and structures

- run guarded read/write SQL operations

- help automate schema creation and migrations

- generate or validate T-SQL

- assist with troubleshooting and performance tuning

Through live demos, you’ll see how MCP brings safe, auditable, enterprise-grade AI automation directly into the SQL workflow—without giving an LLM full database access or embedding secrets in prompts.

If you’re working with Azure SQL, Copilot, or want to build AI-powered data tooling, this session gives you a clear picture of how MCP will reshape the future of database operations.