AI coding assistants are becoming part of the infrastructure engineer's toolkit, but how do we use them effectively when working with real infrastructure?

In this practical session, we'll explore GitHub Copilot with Azure infrastructure and infrastructure as code (IaC), using both Terraform and Bicep. We'll go beyond code completion to see how Copilot can help us understand existing code, investigate Azure issues and work through infrastructure changes.

Through practical demonstrations, we'll explore how context and instructions influence the results, how connected tools and Model Context Protocol (MCP) servers can help investigate an Azure environment, and how to guide an AI assistant through a troubleshooting workflow.

Generating code or identifying a possible cause is only part of the job. We'll look at how to challenge assumptions, review proposed changes, validate Terraform and Bicep, and keep control of what ultimately gets deployed.

This session is aimed at infrastructure engineers who want to use GitHub Copilot as a practical tool in their day-to-day work, while understanding its limitations and knowing when to trust, question or verify its output.