AI case study

PopChillProduct discovery and moderation

Recommendations took days to refresh; staff lost half-days to fraud checks. AI updates feeds in two hours and automates chat screening.

Published|11 months ago

Key results

Refresh Frequency
2 hours
vs every few days
Daily Work Saved
0.5 days
Click Increase
2.5x

Result highlights

Unlock 4 result highlights

The story

Context

One of the most popular pre-owned luxury ecommerce sites in Taiwan and Hong Kong, listing over 100,000 items ranging from designer bags to vintage clothing.

Challenge

A legacy recommendation system took days to refresh and only tracked logged-in users, missing critical engagement signals. Furthermore, staff spent...

Solution
Unlock full story

Scope & timeline

  • Functional image search prototype built in 2 weeks
  • Recommendation system built by 1 engineer in 2 weeks

Quotes

Unlock 5 more quotes

The company

PopChill logo

PopChill

popchill.com

Second-hand luxury fashion marketplace and authentication platform.

IndustryRetail
LocationTaipei, Taiwan
Employees11-50
Founded2021

The AI provider

Cloud computing services, AI infrastructure, and data analytics platforms for enterprises.

IndustryTechnology
LocationMountain View, CA, USA
Employees100K+
Founded1998

Similar Case Studies

Related implementations across industries and use cases

12 AI case studies in Content Moderation

157 AI case studies in Retail

190 AI case studies in Marketing