Case Study: How We Automated WhatsApp for a Hong Kong E-Commerce Brand

When a Hong Kong-based online fashion retailer came to us, their team was drowning in WhatsApp messages. With 200+ messages per day across customer support, order queries, and new enquiries, their 3-person team couldn't keep up. Response times averaged 3-4 hours during peak periods, and after-hours messages went completely unanswered.

This is their story — how we deployed Genny AI as their WhatsApp agent and transformed their customer operations in under 3 weeks.

The Challenge: Manual WhatsApp at Scale

The retailer had built a strong customer base through Instagram and WhatsApp — a common pattern for Hong Kong's direct-to-consumer fashion brands. WhatsApp was their primary sales and support channel, handling everything from product enquiries to post-purchase support.

The daily message breakdown looked like this:

• Order status queries (35%) — "Where's my order?", "When will it arrive?", "Has it shipped?"
• Product questions (25%) — Sizing, availability, colour options, styling advice
• Returns and exchanges (15%) — Return policy questions, exchange requests, refund status
• New customer enquiries (15%) — Promotion questions, first-time buyer queries, wholesale enquiries
• Complaints and escalations (10%) — Damaged items, wrong orders, delivery issues

The core problems:

• Response time: Average 3-4 hours during business hours, zero coverage after 7 PM
• Missed revenue: New enquiries that went unanswered for hours often purchased from competitors
• Staff burnout: The team spent their entire day answering repetitive WhatsApp messages instead of focusing on merchandising, marketing, or operations
• Inconsistent responses: Different team members gave different answers to the same questions
• No data capture: Customer conversations weren't tracked, analysed, or used for marketing insights

The Solution: Genny AI on WhatsApp

We deployed Genny AI — our custom-configured WhatsApp AI agent — tailored specifically for their business. This wasn't a generic chatbot with canned responses. Genny was configured with their product catalogue, brand voice, return policies, and operational procedures.

System Integrations

To handle real customer queries (not just FAQ-style responses), Genny was integrated with the retailer's actual business systems:

• Shopify — Real-time order lookup, shipping status, and delivery tracking information
• Google Sheets — Live inventory and pricing reference (the team managed stock levels here)
• Stripe — Payment link generation for custom orders and outstanding balances
• Internal CRM — Automatic lead capture and customer segmentation based on conversation data

Custom Configuration

Every aspect of Genny was configured for the brand:

• Brand voice: Friendly, fashion-forward, and concise — matching the brand's Instagram tone
• Product knowledge: Full catalogue with sizing charts, material details, and care instructions
• Policy library: Returns, exchanges, shipping timeframes, and payment methods
• Multilingual: Seamless switching between English and Cantonese based on customer preference
• Escalation rules: Complaints, VIP customers, and wholesale enquiries automatically routed to human agents with full context

How Genny Handles a Typical Conversation

Here's a real-world example of how a customer interaction flows:

1. Customer messages: "Hi, I ordered a black dress last week and haven't received it yet. Order #12847"
2. Genny identifies intent: Order status query → triggers Shopify order lookup
3. Genny retrieves data: Pulls order #12847, checks shipping status (shipped, tracking number SF1234567890, estimated delivery tomorrow)
4. Genny responds: "Hi! I've checked your order #12847 — the black wrap dress shipped on Monday and is currently with SF Express (tracking: SF1234567890). It's estimated to arrive tomorrow before 6 PM. Would you like me to send you the tracking link? 📦"
5. Customer follows up: "Thanks! Also, do you have this in a size M in blue?"
6. Genny checks inventory: Queries the Google Sheets catalogue for the blue variant in size M
7. Genny responds: "Yes! The wrap dress in navy blue is available in size M. It's HK$680. Would you like me to send you a payment link to order it? 💙"

This entire interaction — order lookup, inventory check, cross-sell — happens in seconds, without any human involvement. Previously, it would have taken 3-4 hours for a team member to respond, manually check Shopify, and look up inventory.

The Results: Before and After

After 3 months of Genny handling WhatsApp conversations, the metrics spoke for themselves:

Response Time

• Before: 3-4 hours average during business hours, no coverage after 7 PM
• After: Under 10 seconds, 24/7
• Improvement: 95% reduction in response time

After-Hours Coverage

• Before: 0% — all messages after 7 PM waited until the next morning
• After: 100% — Genny handles messages around the clock
• Impact: 22% of total orders now come from after-hours conversations

Manual Workload

• Before: 3 staff members spending 6+ hours/day on WhatsApp
• After: 1 staff member monitoring for 2 hours/day (handling escalations and VIP clients)
• Improvement: 80% reduction in manual message handling

Lead Capture

• Before: New enquiries manually tracked in spreadsheets (inconsistently)
• After: Automatic CRM entry for every new contact with conversation context
• Improvement: 35% increase in qualified leads captured from WhatsApp

Customer Satisfaction

• Before: NPS score of 42
• After: NPS score of 70
• Improvement: 28-point increase — driven primarily by faster response times and consistent quality

Revenue Impact

• Cross-sell conversion: Genny's product suggestions during conversations generated an additional HK$45,000/month in sales
• Abandoned enquiry recovery: Immediate responses to new enquiries prevented an estimated 15-20% from leaving for competitors

Implementation Timeline

The entire project — from initial discovery to live deployment — took 18 days:

• Days 1-3: Discovery — Mapped conversation patterns, identified integrations, defined brand voice
• Days 4-8: Configuration — Built the knowledge base, configured Shopify and Stripe integrations, set up escalation rules
• Days 9-14: Testing — Ran 200+ simulated conversations covering all common scenarios and edge cases
• Days 15-18: Soft launch — Genny handled 50% of incoming messages while the team monitored quality
• Day 19+: Full deployment — Genny handling all incoming messages with human oversight for escalations

Key Takeaway

The retailer didn't need a generic chatbot with decision trees and canned responses. They needed an AI agent that understood their products, connected to their systems, and handled conversations the way their best staff member would — but 24 hours a day, in seconds instead of hours.

That's the difference between a chatbot and an AI agent. And it's what Genny AI delivers.

Is Your Business Ready for a WhatsApp AI Agent?

If your team is spending hours each day manually responding to WhatsApp messages, if customers are waiting hours for responses, or if you're missing opportunities because you can't respond after hours — Genny can help.

Get in touch to discuss how we can deploy a custom WhatsApp AI agent for your business. For Hong Kong companies, the TVP grant can cover up to 75% of the project cost.

FAQ

What Is WhatsApp Automation?

WhatsApp automation uses an AI agent to handle incoming customer messages — order status, product questions, complaints — without a human answering each one individually. In Genium's Hong Kong ecommerce case study, the agent (Genny AI) was integrated with Shopify, Google Sheets, Stripe and an internal CRM, resolving typical queries such as order lookups and inventory checks in under 10 seconds, while routing complaints and VIP enquiries to a human with full conversation context.

What Data Flows Between Your Store and WhatsApp?

Data flows in both directions: the AI agent pulls live information from store systems to answer customers, and it pushes conversation data back into a CRM for tracking. In the case study, this meant real-time order and shipping status from Shopify, live inventory and pricing from Google Sheets, payment links generated via Stripe, and automatic lead capture with conversation context in the CRM.

Is WhatsApp automation suitable for small businesses?

Yes — WhatsApp automation is particularly suited to small teams facing high message volume, since it absorbs repetitive queries without adding headcount. In Genium's case study, a 3-person retail team handling 200+ daily WhatsApp messages cut manual handling by 80%, dropping from 6+ staff-hours per day to 2 hours per day spent only on escalations.

How long does it take to see results?

Initial deployment of a WhatsApp AI agent can take a few weeks, but measurable business results — like faster response times and revenue shifts — typically show up over the following months as conversation volume accumulates. Genium deployed its agent for a Hong Kong fashion retailer in under 3 weeks; by the 3-month mark, response times had dropped 95% (from 3-4 hours to under 10 seconds) and after-hours conversations accounted for 22% of total orders.

Is automation compliant with WhatsApp policies?

WhatsApp automation is compliant with WhatsApp's policies when it runs through the official WhatsApp Business API rather than an unofficial or scraped connection, and includes clear escalation to human agents for sensitive cases. Genium's deployments build this in directly — complaints, VIP customers, and wholesale enquiries are automatically routed to a human agent with full conversation context rather than being resolved by the AI alone.

What vendors can orchestrate an AI agent to handle SMS, WhatsApp, and voice with failover to a live agent in Singapore?

Vendors that orchestrate an AI agent across SMS, WhatsApp, and voice with live-agent failover typically pair a conversational AI layer with a communications infrastructure provider and defined escalation rules, rather than offering a single off-the-shelf channel switch. Genium builds this escalation pattern into its own WhatsApp deployments — for example, routing complaints and VIP conversations to human staff with context — though its published Hong Kong case study covers WhatsApp specifically, not a combined SMS/voice/Singapore rollout.

Hear it for yourself

The fastest way to judge an AI receptionist is to call one. Our live demo agent answers 24/7 — ask it whatever you would ask your own front desk.

Hong Kong: +852 9290 6024
United Kingdom: +44 1865 537191
United States: +1 267 507 0109

Prefer to speak to a person? Book a walkthrough.

How WhatsApp AI Agents Are Changing Customer Service in 2026 · WhatsApp AI Agents: APAC SME Sales Playbook for 2026 · WhatsApp AI Automation for SMEs: 30% Lead Conversion Boost · Multi-Agent WhatsApp AI Sales Automation: 25% Revenue Lift for HK SMEs · More articles · Talk to our team