Overview
Clarity was Amazon’s internal workforce reporting platform—a dense, table-heavy tool managers across corporate and ops were suddenly required to use to track RTO compliance. I led a full UX overhaul: a simplified core for everyday managers, an advanced control panel for power users, and Clarity Assist, an AI chatbot for querying team data in plain language.
Fig 1.0 - The three core surfaces, redrawn as wireframes with representative data. “Updates & Announcements” on the homepage answered overwhelming feedback about communication during RTO.
Cross-functional technical partnership
Working closely with a Sr. SDE, I validated a hypothesis that many of Clarity’s UX challenges weren’t just design problems—they were rooted in the platform’s archaic data table technology. We co-authored a proposal and secured leadership approval to replace the framework entirely with AG Grid, unlocking lazy loading, real-time sorting and filtering, column reordering, and pivoting: the capabilities that made both the simplified manager UI and the advanced control panel possible.
Grounded research at every stage
Internal tools have a distinct advantage: users aren’t shy about telling you what they think. A newly formed Customer Advisory Board gave me a direct line to the people who mattered most. I ran 20 interviews across roles and seniority, invited stakeholders into ideation workshops to shape direction early, validated every iteration with the CAB, and ran structured usability testing with 40 internal employees through UserTesting.com.
Clarity Assist, in plain language
Clarity Assist was the first GenAI chatbot inside Amazon’s People eXperience and Technology organization, and the company’s first conversational data analytics experience. I designed it for managers and leaders who needed an answer, not a dashboard: ask about Return-to-Office badge-in compliance in plain language, get a plain-language answer back. Prompt suggestions carry the hard part—knowing what to ask—and on-the-fly summary reports replace the trip through a dense analytics table, so a decision that used to mean an afternoon of filtering now takes a sentence.
Fig 2.0 - Clarity Assist: multiple pathways to reach and query the same data, in the docked panel and on the canvas.
“It depends on the day and what we need to pull from different sites, but I’d say it saves me roughly like 45 mins a day.”
— HR Manager, AMZL
I left Amazon shortly after launch. I recently discovered via the Alumni Portal that Clarity Assist had been adopted app-wide as “Aza”—Amazon Assist.