Skip to main content

Momo eCommerce & Live Innovation PoC

Published 2026-05-14Updated 2026-07-024 min read

"An eCommerce innovation PoC that ties massive SKU, member, live-broadcast, campaign, and in-house logistics data together with external trend, weather, economic, and competitor signals — solved through Ontology + Agentic AI."

Momo (富邦媒 / momo.com)'s millions of SKUs, TV home-shopping and live commerce, mobile app, and in-house logistics (24-hour delivery) consolidated into a single KG. Public information only.


🚀 Interactive Demo

👉 Open the demo page (new window)

Hands-on with 11 scenarios plus the core UIs for live-commerce attribution, delivery SLA, and recommendation diversity.


1. PoC One-Line Summary

AxisContent
For whom5 departments: Marketing · Category · Search/Recommendation · CS · Logistics
WhatMassive SKU + Member + Live broadcast + In-house logistics + 4 external sources
HowOntology (KG 25 classes) + Agentic AI
DifferentiatorLive commerce attribution + Delivery SLA monitoring + Recommendation diversity

2. Business Structure (Public Information)

AxisContent
ChannelsMobile app (#1 traffic in TW) · Web · TV home-shopping · Live broadcasts · Line/KakaoTalk
SKUsMillions (long-tail strength)
Logisticsmomo Fast (24-hour delivery), in-house fulfillment centers
ParentFubon Financial subsidiary

3. Five-Department Personas

CodeDepartmentKPI
P1Marketing (App · Campaigns)App ROAS · Live concurrent viewers and conversion
P2Category · TrendNew-product trend SOV · Category LTV
P3Search · Recommendation · MarTechRecommendation accuracy · Diversity · Search latency
P4Customer Service · CRMNPS · Refund rate · Delivery satisfaction
P5Logistics · MDDelivery SLA · Per-region fulfillment load

4. 11 Scenarios Summary

#ScenarioData Mix
S1Natural-language semantic searchSKU + Member + Live review + External
S25-department persona chatbotAll tools autonomously invoked
S3Category · BU insight cardsIn-house GMV + Trend + Competitor
S4Persona matching + ClusteringRFM + Live viewing + Social
S5Omnichannel campaign ROAS simulatorCampaign · Live · App · Messenger
S6External signal fusion4 sources (Social · Weather · Economy · Competitor)
S7Omnichannel member journeyApp → Live → TV → Re-purchase
S8Marketing consent · PII guardrailsBedrock Guardrails + Ad-copy review
S9-MoLive-commerce broadcast attributionViewer → Cart → Purchase funnel
S10-MoIn-house logistics delivery SLA + Order optimization24-hour promise vs measured
S11-MoMobile-app recommendation · Search diversityDiversity · New-product exposure · Long-tail

5. AWS Stack (Common)

Same as LG H&H + live-stream attribution (Kinesis), delivery SLA (TimeStream), ~800K edges (massive SKU + live events).


6. Top-5 Demo (30 minutes)

OrderScenarioTime
1S1 Semantic search4 min
2S9-Mo ⭐ Live commerce attribution5 min
3S10-Mo ⭐ Delivery SLA + ordering4 min
4S11-Mo ⭐ Recommendation diversity5 min
5S2 Persona chatbot4 min

7. Deliverables