Skip to main content

S8. Marketing Consent & PII Guardrails (AMWAY)

Published 2026-05-14Updated 2026-06-301 min read

1. URL & Personas

  • /compliance · P4 (CX), P5 (Compliance)

2. Active Policies

PolicyAction
Violence · hate · sexual · misinformation (Bedrock 4 topics)Block
Suspected medical / legal adviceBlock + guidance
Sending campaigns without marketing consentBlock tool
PII (name, contact, card)Auto-mask
Suspected minor-ABO sign-upBlock context
Health functional food advertising language (efficacy / treatment claims)LLM-labeled guard
Direct-selling level guaranteesAuto-block

3. Processing Pipeline

  1. Chatbot inputs and outputs → Guardrails
  2. On block: reason + policy ID
  3. Auto-mask PII
  4. All blocks and masks → Audit DynamoDB

4. Output UI

  • 7 policy cards
  • Real-time trace
  • Daily block counts

5. Demo Scenarios

  1. "ABO cust_001234 contact" → PII masking
  2. "Minor ABO sign-up simulation" → block + reason
  3. "Ads emphasizing Nutrilite efficacy" → LLM guard auto-rewrites