💡 OVERVIEW

Designed an AI-powered recommendation engine that transformed how Account Managers diagnosed and responded to campaign performance drops. Before this, troubleshooting required manually cross-referencing dashboards, metrics, and segment data — a process that could take hours and still produce inconsistent conclusions depending on the analyst's fluency. I led design end to end: problem framing, the recommendation architecture, interaction model, and shipped experience.

✨ WHAT MADE IT MORE THAN A DASHBOARD FEATURE

The easy version of this project would have been a smarter filter or a better alert system. Instead, the core design question was: what does it look like when the system does the analytical work for you? That meant rethinking the entire information architecture — not just surfacing data, but surfacing a prioritized, actionable interpretation of that data. How recommendations were ranked, how root-cause analysis was exposed without overwhelming, how the drawer interaction balanced depth with speed — all of those decisions were in service of one thing: getting an Account Manager from "something's wrong" to "here's what to do about it" without requiring deep analytical fluency.

🎯 IMPACT

Reduced campaign troubleshooting time from hours to minutes, enabling Account Managers to act on performance insights faster and with greater confidence. The recommendation engine became the foundation for a broader AI-powered optimization layer across the platform — establishing the interaction patterns and trust model for how the product would surface intelligent guidance going forward.