Every engineering organization is adopting AI coding tools, but everyone isn't seeing the same transformative results. Amazon Stores set out to find a way to accelerate teams - increasing output by three times - through AI-native development and achieved a 4.5x median gain. The difference wasn't the tools alone; it was pairing them with new practices that changed how engineers work. In this Masterclass, Amazon Stores leadership will unpack the playbook behind those results, with a focus on spec-driven development - where structured specifications become the connective tissue between human intent and AI execution, enabling agents to reason across complex codebases instead of hallucinating. You'll explore how teams reskilled engineers to become orchestrators of AI agents rather than executors of every detail, restructured codebases and documentation for AI consumption, and drew a sharp line between disciplined AI-native development and ad-hoc "vibe coding." Underpinning this work, AWS services - including tools like Kiro - provide the development infrastructure that teams build on. We'll also candidly discuss what bottlenecks we encountered at 3x speed and how to navigate the cultural change needed to drive this transformation. Whether you're leading a startup engineering team or an enterprise organization, you'll leave with a practical framework for turning AI adoption into sustained, measurable development acceleration.