What if your home could reason about its own energy usage? Not just follow rules, but actually think?

In this session, I'll walk you through how I built a fully agentic home automation system using the GitHub Copilot SDK, turning a real-world power optimization challenge into a hands-on exploration of what agentic AI looks like beyond the chat window.

We'll cover how the SDK's agent model maps to a home automation context: defining tools and skills for reading sensor data, controlling devices, and reacting to dynamic conditions like energy tariffs. You'll see how session lifecycle, agent orchestration, and tool invocation come together in a system that genuinely reasons about when to run your washing machine or whether to charge your EV, rather than just executing a fixed schedule.

Expect real code, honest trade-offs, and a few surprises from living with an AI that controls your electricity bill.