The Problem
Microsoft Copilot needed to be more than an AI chatbot — it needed to meet users in context, surfacing relevant actions at the right moment within Windows. Early entry points weren't optimized for how users actually interacted with the assistant, leading to lower-than-expected engagement and discovery.
The Approach
I focused on optimizing Copilot's entry point to surface more contextually relevant actions based on user behavior and system state. Working cross-functionally with design, engineering, and data science teams, we iterated on placement, triggers, and action recommendations. I also contributed to Copilot's global expansion, ensuring the experience met regulatory requirements (DMA/CTA) and localization standards across the EU and China.
The Outcome
The entry point optimization drove a 15% increase in user engagement. Global expansion efforts extended Copilot's reach by 35% of TAM, bringing the AI assistant to approximately 490M monthly active users across regions. These improvements helped establish Copilot as a central part of the Windows experience.