Product Build + Automation

POA Automation Engine

WW-IN First Automated Seller POA Evaluation

RoleLead Analyst & Architect
TeamIN CX Andon Cord Analytics
Timeline8 Weeks
StakeholdersAndon Cord Resolvers, Seller Experience, IN Marketplace Quality
Tools
PythonQualtricsData PipelinesSQLSurvey Architecture
73%TAT Reduction
6200DPMO Addressed
1.86MCustomers Impacted
50% of ACsPerson-hours Eliminated

Over 40% of Andon Cords in the IN marketplace are resolved through seller-submitted POAs, but the current manual, email-based evaluation process lacked rigor. With only 24% of WW Auto Andon detections investigated due to resolver constraints, 1.6M detections (6200 DPMO / 1.86M impacted customers) remained unresolved annually.

Working with the AC Team, I created a data pipeline to identify high-impact, eligible offers that were unresolved despite being pulled. I identified key gaps in the current POA framework and built use-case-specific, streamlined evaluation criteria. I onboarded the Qualtrics tool to enable an IN-AC first automated solution for seller communication and response consolidation, and built end-to-end data architecture for delivery of surveys, collection of inputs, review of POAs, and analytical evaluation.

In Feb '25, the team piloted the solution on 100 seller-actionable Andon tickets. Results: 77% tickets flagged for suppression (vs. 90% ideal), 23% POAs accepted (vs. 10% ideal), leading to TAT reduction of 73% from 15 days to 4 days and eliminating person-hour requirements for 50% of ACs.

Why It Mattered

This was a WW-IN first — no other marketplace had automated the POA evaluation process. It proved that combining data architecture with survey tooling could solve a problem that had been considered too complex for automation. The framework is now being evaluated for WW expansion.