Tackling tech debt with the GitHub Copilot Cloud Agent

Before the age of agents, tech debt often went straight from the backlog to the graveyard. While it’s important to continually identify ways to improve your services, it can be difficult to prioritize those items over customer-facing enhancements. With the introduction of AI agents, that paradigm has shifted: you can now more easily modernize, refactor, and evolve your codebase, ensuring it never becomes the next legacy. This talk will cover how to embrace the process of modernizing legacy software and how to leverage AI agents, like Copilot coding agent, to keep your codebase up to date. Join the conversation on the Microsoft Foundry Discord: https://aka.ms/AIDevDays/Discord Sign up for an Azure Free Trial: https://aka.ms/AzureFreeTrialYT Get hands-on with today's VS Code lab repo: https://aka.ms/AIDevDays/LabDay2 Catch up on-demand: https://aka.ms/AIDevDays/OnDemand Continue learning next week with AI Apps & Agents' global sessions: https://aka.ms/AIDevDays/Next Register for the January AI Dev Days Hackathon: https://aka.ms/AIDevDays/Hackathon Chapters: 00:00 - Introduction: Meet Brittany Ellich and GitHub Copilot 00:43 - Agenda: Tackling Tech Debt with AI 00:56 - The Tech Debt Graveyard: Why It Happens 02:20 - Consequences of Ignoring Tech Debt 02:47 - Introducing the Tech Debt Elimination Accelerator 03:00 - How GitHub Copilot Coding Agent Works 04:02 - Demo: Automating Unit Test Creation with Copilot 07:03 - Reviewing Copilot’s Pull Request Output 08:12 - Building a Sustainable Tech Debt Strategy 08:54 - W.R.A.P Framework: Write Effective Issues 10:28 - Refining Instructions for Better AI Results 11:41 - Breaking Down Large Tasks into Atomic Chunks 12:50 - Pairing with the Coding Agent: Human + AI Collaboration 14:42 - Time Blocking and Workflow Tips 15:12 - Final Thoughts: Turning Tech Debt Mountains into Managed Terrain