Skip to main content

11-Scenario Mapping (AMWAY)

Published 2026-05-14Updated 2026-06-304 min read

8 common scenarios (S1–S8) + 3 AMWAY-specific scenarios (S9-A · S10-A · S11-A) = 11 in total. Every scenario specifies a data mix combining in-house data with four external sources (social, weather, economic, regulatory).


1. Overview of the 11 Scenarios

#ScenarioIn-house DataExternal DataURLPrimary Persona
S1Natural-language semantic searchSKU · ABO · in-house reviewsSocial posts/semantic-searchP1, P2
S2Persona chatbot across 5 departmentsAll in-houseAll 4 external/chatAll
S3Category & BU insight cardsOrderTransaction · ABODirectSaleTrends · FX/insightsP3
S4Lifestyle persona matching + clusteringRFM · category affinitySocial personas/personasP2
S5Omnichannel campaign ROAS simulationCampaign · TouchpointSNS · search response/campaign-roasP1
S6External signal fusionIn-house GMV (BU · country)All 4/signalsP2, P3
S7Omnichannel member journeyStore → ABO direct → subscription → re-signup(none)/journeyP4
S8Marketing consent · PII guardrailsConsent · compliance(none)/complianceP4, P5
S9-AABO organization visualization + performance trackingABO tree · PV/BV · level(none)/abo-treeP1
S10-ASubscription lifetime analysisSubscription · event history(none)/subscriptionP4
S11-ADirect-selling law, minors, HFF advertising complianceCompliance · consentRegulatorySignal/regulatoryP5

2. Persona × Scenario Matrix

DepartmentS1S2S3S4S5S6S7S8S9-AS10-AS11-A
P1 ABO Field
P2 CMI
P3 D&A
P4 CX
P5 Compliance

3. Per-scenario Data Mix (Summary)

#Core Data
S1OpenSearch (BM25 + KNN + RRF + Rerank) + Neptune 1-hop
S2Autonomous invocation of 12 tools (semantic_search · neptune_query · campaign_simulator · social_trend_join · weather_join · economic_join · regulation_join · abo_tree_query · subscription_lifecycle · cluster_run · attribution_calc · pii_mask)
S3Snowflake/Neptune aggregation + AgentCore Code Interpreter
S4KMeans 6 + LLM labeling + social keyword enrichment
S5Bayesian (PyMC) + channel · country prior
S6Time series + correlation heatmap + lag analysis
S7Swimlane timeline (store · ABO · subscription · campaign)
S8Bedrock Guardrails + consent matrix
S9-ANeptune ABO Tree DFS + cytoscape.js
S10-ASubscription state machine + churn signature + LSTM (optional)
S11-ARegulatorySignal × ad-text review + minor guard

4. Top-5 Demo Path (30 minutes)

[S1] Natural-language search 4 min
"Vegan protein + new ABO in their 30s + Instagram +20%"
→ SKU (Nutrilite Plant Protein) + ABO + external reviews

[S9-A] ABO organization visualization 5 min ⭐ AMWAY differentiator
"Five-level Downline tree for a Diamond-level ABO"
→ cytoscape tree + PV/BV · level color coding

[S10-A] Subscription lifetime 5 min ⭐
"Top 100 ABOs with 6-month churn signatures + automated win-back campaign"
→ Subscription state · churn z-score · recommendation

[S2] Persona chatbot 5 min (switcher)
"Upline ROI vs subscription retention"
→ Compare answers from P1 (ABO Field) and P4 (CX)

[S11-A] Compliance 3 min ⭐
"Blocking minor ABO sign-ups" + "Reviewing Nutrilite ad language"
→ Block / mask count visualization

5. PoC Timeline (8 weeks)

WeekWork
1–2Data ingestion (in-house N = 1K ABOs + 5K customers + 50K synthetic + 4 external sources)
3–4S1 · S2 · S3
5S9-A ABO tree + S10-A subscriptions
6S5 · S7 · S11-A compliance
7S4 · S6 · S8
8Demo rehearsal + backup