Giving AI agents visibility into SQL Server with MCP | Data Exposed MVP Edition

Anthony Nocentino and Anna Hoffman show how to give GitHub Copilot controlled, real-time visibility into SQL Server using MCP (Model Context Protocol) servers. They demo a custom DMV-based MCP server for DBA workflows and a zero-code Data API Builder approach for app data access, with guardrails and permissions defined by you.

Overview

This episode of Data Exposed (MVP Edition) demonstrates two MCP servers that let GitHub Copilot interact with SQL Server through structured tools rather than direct database access.

Key idea: the agent calls tools exposed by an MCP server; the MCP server runs the queries you allow and returns only the results you explicitly expose.

What they build and demo

1) A custom SQL MCP server for DBA workflows

• A custom-built MCP server that exposes a set of tools (described as ~30 tools) backed by SQL Server DMVs.
• Intended for DBA-style tasks across multiple SQL Server instances.
• Demonstrations include Copilot-driven troubleshooting and health checks, such as:
  • Diagnosing blocking
  • Checking Always On Availability Group (AG) health
  • Snapshot-style health checks

2) Data API Builder (DAB) as an MCP-backed data access layer

• Uses Data API Builder (DAB) to provide natural-language access to application data.
• Emphasizes zero code required for the data access layer.
• Permissions and what data is exposed are defined in configuration, acting as guardrails.
• Demo includes an application-style interaction (placing a product order) without writing SQL directly.

Architecture and control mechanisms (high level)

• Copilot (agent) does not connect to SQL Server directly.
• Copilot calls MCP tools.
• The MCP server:
  • Implements the tool surface area (what actions/queries are allowed)
  • Executes the database queries under controlled conditions
  • Returns only the allowed outputs

Video chapters

• 0:30 Intro to Anthony Nocentino
• 1:00 Why SQL MCP Server, for developers and DBAs
• 2:13 What is an MCP Server
• 2:58 Solution architecture
• 6:30 Control mechanisms
• 7:30 See the agent in action
• 8:45 SQL Server snapshot health example
• 14:20 Using Data API Builder as an application developer
• 19:00 Natural language interactions
• 20:30 Anthony's #1 piece of advice!

Resources

• SQL MCP server repo: https://github.com/nocentino/sql-mcp-server
• Companion blog post: https://www.nocentino.com/posts/2026-05-24-giving-copilot-visibility-into-sql-server-with-mcp/