VSM03 Workshop: Team Development for the Future with Agentic AI & GitHub

Most developers use GitHub Copilot daily — but when it produces the wrong output, hallucinates, ignores your instructions, or just stops being useful, most reach for trial and error. This full-day workshop changes that. Join GitHub forward-deployed engineers Mickey Gousset and Ken Muse as they take you deep inside how AI coding tools actually work — and how to get the most out of them.

Whether you’re using Copilot or a different agentic solution, this workshop gives you the principles and foundation to make sense of it. You’ll start with a solid grounding in how large language models function — including a clear explanation of what tokens actually are, why they matter, and how to use that knowledge to your advantage. You’ll see how context windows and token limits shape what the model sees and decides, and how to choose the right model for a given task.

From there, you’ll explore the agentic loop in depth: how Copilot processes your prompts, invokes tools, manages state, and decides what to do next. That understanding unlocks real debugging skills — so you can diagnose why the model is behaving unexpectedly, fix bad outputs, and build more reliable AI interactions from the start. You will discover how to use instructions, prompts, skills, agents, and hooks the way they were intended, shaping how Copilot behaves for your team and your codebase. Along the way you’ll understand how other tools — like Agent SDKs and Agentic workflows — support deeper, customized AI solutions.

Leave the day with a clear understanding of how to get consistent, high-quality results from AI tools — and the ability to debug and tune them to achieve that.

You will learn:

• How the agentic loop works under the covers — including how models process context, manage tokens, and make decisions — so you can debug unexpected behavior and reason confidently about any AI tool.
• How features like instructions, prompt files, skills, agents, and hooks fit together to let you build agentic practices that are consistent and tailored to your development needs.
• How token usage, context design, and model selection directly affect output quality — and gain the techniques to improve results without relying on guesswork.