WhatsApp Chatbot Automation
Scaling Retail Engagement via WhatsApp | Health & Beauty | Conversational UX | Omnichannel Service Design
Overview
Designed an automated WhatsApp conversational flow integrated with a digital diagnostic microsite to introduce an omnichannel purchasing experience for a major health and beauty retailer.
Background
Designed an automated WhatsApp conversational flow integrated with a digital diagnostic microsite to introduce an omnichannel purchasing experience for a major health and beauty retailer.
Role
UX Strategy
Conversational Design
Interaction Design
Frontend Development
Tools
Miro
SleekFlow
Elementor Pro
Figma
Fillout Form
Year
2025
01 / Discovery & Alignment
The project began with a macro-level audit to explore how physical retail actions could trigger sustained, long-term digital health engagement. In this initial stage, we had to align stakeholders on how a basic chat interface could handle complex, multi-week behavior modification. During the discovery phase, I mapped out a high-level operational framework to group our conversational touchpoints into modular pillars—ranging from initial onboarding triggers to progressive symptom tracking, personalized advice feedback loops, and milestone evaluations. This early alignment tool allowed us to establish clear business rules and lifecycle thresholds before committing to technical development.
Key Activities Undertaken:
Conceptual Pillar Mapping: Grouped unstructured conversational requirements into 5 clean core modules (Greeting, Symptom Tracking, Personalized Feedback, Image Uploads, and Reminders) to evaluate system complexity early.
User Retention Baseline Rules: Outlined the initial lifecycle timeline thresholds, establishing target milestones at days 7, 21, and 30 to track user progress and trigger automated clinic booking fallbacks.
Omnichannel Gap Analysis: Investigated the severe drop-off trends that occur when standard physical retail consumers are asked to adopt ongoing, multi-week digital wellness habits.
The initial high-level conversational architecture designed during discovery to align stakeholders on retention milestones, data requirements (such as day-zero image uploads), and strategic clinical escalation pathways.
02 / Strategic Challenge
The primary strategic challenge centered on creating a balanced ecosystem that could simultaneously handle low-commitment retail engagement and high-complexity medical routing. As shown in the unified layout of the PRE-PURCHASING AND BOOK APPOINTMENT FLOW, we needed to split diverse user intents into dedicated tracks right from a single physical QR code scan. The bottleneck lay in isolating clinical risk: while standard users received quick ingredient profiles, users identifying complex symptoms required a complete shift out of the marketing ecosystem. To solve this, I engineered a parallel, automated clinical diversion path within the same system that handled multi-step clinic locating, live scheduling, calendar synchronization, and automated confirmation touchpoints seamlessly.
Key Goals Established:
Macro Funnel Segmentation: Architect an entry-level classification matrix to cleanly separate incoming physical store traffic into targeted category tracks, as visualized in the Pre-purchasing flow draft component.
Frictionless Clinical Escalation: Build a comprehensive, self-contained interactive booking engine to safely off-ramp complex symptom profiles from automated chats into offline medical consultations, as mapped out in the Complex Symptom logic layer.
A unified strategic overview mapping out the entry-level retail category classification funnel alongside its accompanying complex-symptom medical appointment booking architecture.
03 / System Architecture
To transition users seamlessly from physical retail to a continuous digital habit, I architected a macro-level conversational ecosystem that unified the entire customer lifecycle. As detailed in the comprehensive system map, the architecture divides neatly into two primary environmental states: an entry-level "Pre-Purchasing" track and a highly structured "Post-Purchasing" retention journey. Rather than relying on simple linear messaging, this architecture maps out granular branching dependencies based on product line enrollment (Acne, Sleep, Hair, Menses), defining precise rules for automated data collection, progressive multi-week reminders, and downstream repurchase channels.
Key Structural Frameworks Designed:
Lifecycle State Decoupling: Structured the system to move users seamlessly from pre-purchase category routing into long-term post-purchase engagement loops based on physical store conversions.
Multi-Week Dynamic Retention Logic: Mapped the specific daily tracking feedback loops for individual cohorts, tracking parameters like product consumption timing, symptom severity scales, and optional progress photo captures.
Full-Circle Conversion Triggers: Engineered final-stage repurchase pathways that cleanly redirect users to e-commerce storefronts or localized brick-and-mortar retail locations at the conclusion of their evaluation periods.
The initial high-level conversational architecture designed during discovery to align stakeholders on retention milestones, data requirements (such as day-zero image uploads), and strategic clinical escalation pathways.
Mid-level logic architecture detailing photo-upload exceptions, automated question intervals, and the recurring 7-to-28 day informational tip distribution matrix.
04 / Core Execution
Translating our structural system architecture into a production-ready chatbot environment required a multi-stage execution framework that bridged collaborative UX writing with rigorous system validation. The process initiated within Miro, where the copywriter established the foundational dialogue frameworks. My role focused on taking those revised copy blocks, staging them within the layout grids, and meticulously preparing all accompanying product and promotional image assets to ensure they met strict platform formatting specifications.
Once the assets and strings were finalized, I ported the updated content into the live engine. While our Tech Lead engineered the underlying backend webhook triggers and custom API integration layers, I took ownership of configuring the conversational nodes, staging the edited copy strings, and running end-to-end execution tests on the entire conversational flow to eliminate dead ends. To complete the ecosystem loop, I validated the final user touchpoints, verifying the interactive receipt-upload module, mobile message display parameters, and web-view digital voucher delivery.
Core Interface Solutions Delivered:
Conversational Staging & Design Ops: Staged approved copywriter revisions into comprehensive conversation matrices within Miro, aligning image assets, dynamic fields, and menu parameters across split system versions.
Visual Asset Engineering: Prepared and optimized all system-wide product promotional graphics to match platform-enforced display aspect ratios and compression limits.
Production QA & Validation: Managed node updates and conducted comprehensive end-to-end interaction testing inside the live flow builder, validating user fallback behaviors while collaborating with the Tech Lead on API hook boundaries.
Conversational interaction blueprints and localized copy string layouts updated and staged during the asset preparation and content optimization phase.
Production configuration interface where I staged revised copy and conducted end-to-end path testing alongside custom backend API layers engineered by the Tech Lead.
Final end-user mobile interface screens demonstrating the validated receipt-upload loop, interactive chat feedback strings, and branded digital voucher fulfillment.
05 / Strategic Outcomes
The pilot provided critical data on a highly niche UX challenge: building a hybrid "phygital" loop where purchases happen physically in-store, but product exploration and education happen digitally over chat. While this model successfully drove point-of-sale validation, scaling it introduced systemic hurdles.
Initially, the verification flow was too tedious, causing user drop-off. We successfully streamlined the security logic, but the project hit broader resource limits: a tight marketing budget capped initial traffic, and limited customer support manpower made it difficult to manage live queries when automated handoffs failed.
Additionally, the third-party automation tool had limited native features. Stitching multiple external platforms together to bridge these functional gaps increased technical complexity and pushed tool subscription costs over budget. Ultimately, we made the strategic decision to pause the service. Despite being sunsetted, the project served as an invaluable testing ground for our team’s omnichannel retail strategy and conversational design capabilities.
◆
Key Retrospective Insights:
UX Needs Support Manpower: Conversational design relies heavily on customer operations. Streamlining the digital interface often shifts the troubleshooting burden onto human agents; without dedicated support staff, the channel cannot scale.
Hidden Cost of Feature Gaps: Relying on basic third-party automation tools causes software fragmentation. Renting extra platforms to patch missing features causes subscription fees to scale exponentially and exceed experimental budgets.
The Friction Optimization Curve: Security steps require a strict balance between data protection and transactional speed. Every extra verification loop drastically reduces completion rates on mobile messaging screens.