How do you build and test a live NBA experience when there's no live game to test against — and how much faster can you go once AI agents join in? At Sky's Prague team we ship Peacock Courtside Live, driven by a feed we don't control: score, clock and game state changing second by second, from backends we don't own, flawless across markets, devices and networks.

This session starts with two habits that make that manageable — if you repeat it, automate it; if you wait for it, mock it — then goes further with AI. You'll learn to automate repetitive API work with an API client (variables, environments, scripting), mock and reshape what you can't control with a proxy tool (map local, modify live data, simulate errors), connect these tools to AI agents via MCP, use AI to speed up problem analysis by pointing you at the failing call, and have AI generate the rules and scripts for you from a plain-English description. You'll leave with a concrete, modern API workflow.