Senior Product Designer

Making Conversation Between Sellers & Buyers Smarter with AI

Personalized, AI-driven quick actions for faster, smarter shopping conversations.

Overview

Rethinking the Chatbot:

How Can We Make Conversation Between Sellers & Buyers Faster & Smarter?

Rethinking the Chatbot:

How Can We Make Conversation Between Sellers & Buyers Faster & Smarter?

Shopee’s shop chatbot plays a crucial role in connecting buyers and sellers, helping users get quick answers about products, orders, and store policies. However, the existing chatbot experience felt outdated and lacked intelligence.

As the lead designer for this revamp, my goal was to enhance the chatbot’s usability by leveraging AI to provide smarter, more personalized conversation starters. This meant:

✅ Making the chatbot feel intuitive and responsive

✅ Delivering personalized quick actions based on user context

✅ Reducing friction in conversations and helping users get what they need faster

This case study walks through (high level) my process of transforming a static chatbot into a dynamic, AI-powered experience.

As the lead designer for this revamp, my goal was to enhance the chatbot’s usability by leveraging AI to provide smarter, more personalized conversation starters. This meant:

✅ Making the chatbot feel intuitive and responsive

✅ Delivering personalized quick actions based on user context

✅ Reducing friction in conversations and helping users get what they need faster

This case study walks through (high level) my process of transforming a static chatbot into a dynamic, AI-powered experience.

Role

Product designer

Team

1 Product Designer

2 Product Managers

1 Design Lead

Development Team

1 Product Designer

2 Product Managers

1 Design Lead

Development Team

1 Product Designer

2 Product Managers

1 Design Lead

Development Team

Timeline

4 months (still undergoing development as of Apr 2025)

Dive Into The Problem Space

A Traditional One-Size-Fits-All Chat Experience

Before the revamp, the chatbot’s conversation starter was basic and uninspiring:

👉🏻

A simple chat bubble with a generic “You may want to ask” suggestion

👉🏻

No personalization based on the user’s entry point, scenario, or needs

👉🏻

An outdated UI that didn’t align with Shopee’s evolving brand

👉🏻

Missed opportunities for engagement and conversion

The results?

Users struggling to find relevant information quickly

Sellers receiving repetitive inquiries

Potential revenue loss due to poor chatbot experience

Research & Insights

Looking Around - Analysing What Are Others Doing?

To guide our approach, I conducted a competitive analysis of both direct and indirect competitors, studying how they structured their chat experiences. Key questions I explored:

？

How do other platforms use AI to personalize chat interactions?

？

What types of quick actions and prompts drive engagement?

？

How can we balance automation with user control?

Landscape Research

Finding common patterns and ways other platforms introduce the chatbots and showcase the chatbot capabilities

Design Logic

These insights helped shape a framework for structuring dynamic conversation starters that felt helpful, not intrusive.

Where we started

Exploring & Adding Dynamic Elements into Static Conversations

With our research findings in mind, I began mapping out the new possible chatbot experience:

Early Design & Exploration

Testing different directions to align with product vision

Experiments on how should we introduce AI in chatbot

Testing out different tone of voice and content direction

Mockups & Prototyping

Creating interactive prototypes, including animations, to showcase design and get buy-ins

UI Experiement

Exploring different UI elements and layouts that showcase newly added AI capabilities

How to give an elevated look and feel

How to present functions

How to showcase contextual suggestions

Approach & Process

How did we tackle the problem?

Buying digital ads was confusing, inefficient, and heavily dependent on account managers. Advertisers struggled with fragmented workflows, unclear pricing, and a steep learning curve. Our goal was to create a seamless, intuitive platform that enabled businesses to buy ads independently, without needing expert guidance.

Solution: The AI-Powered Solution

Elevated Shopping Conversation with A Chatbot That Knows What You Needed

With the revamp, we moved away from static, rule-based suggestions to a dynamic, AI-powered experience. The chatbot now adapts based on where the user starts the conversation—whether it’s from the shop homepage, product details page, or order details page. Each entry point presents a tailored experience that looks, feels, and functions smarter.

Solution: The Redesigned Experience

What did we create, and how did it improve the user experience?

1

When user enters from the shop homepage

➡

Smarter First Impression

✨Personalised greeting

Instead of a generic “Hi there” or "Dear Customer", users now see a more human, scenario-aware greeting that sets a friendlier tone.

🎨 Modernised UI

The overall visual direction was refreshed with gradients, cleaner layouts, and intelligent content groupings to signal a “smarter” experience.

🧠 AI-powered suggestions

Previously, sellers manually configured FAQs and hot topics, resulting in limited usefulness. Now, AI automatically surfaces relevant features and quick actions based on the user’s platform-wide behaviour.

2

When user enters from a product details page

➡

A Personal Shopping Assistant

🧾 Product-aware suggestions

The chatbot understands the product shown and adjusts its quick actions accordingly.

💬 Smart guidance

The assistant also highlights questions users commonly ask about this product type, helping to reduce typing effort and encourage discovery.

This makes the conversation feel like it’s tailored to the product—not just the platform.

3

When user enters from an order details page

➡

Fast Help, No Guesswork

📦 Order-aware quick actions

The assistant recognizes the order status (e.g., shipped, delivered, pending) and dynamically presents relevant functions:

If the order is still processing → “Cancel my order”
If it’s shipped → “Track my delivery”
If it’s delivered → “Request a refund”

🤖 Guided prompts

These suggestions reduce confusion and help users resolve post-purchase issues quickly without needing to search or scroll.

By surfacing the right tools at the right moment, the chatbot becomes a reliable post-purchase companion.

Impacts & Results

A Faster, More Relevant Chat Experience

The revamped chatbot is designed to improve user experience and seller efficiency. While the feature was still in the rollout phase, we expected to see:

🚀

Increased Engagement

More users clicking on AI-generated quick actions

Fewer Repetitive Inquiries

Sellers spending less time answering the same questions

📉

Better Information Discovery

Users finding what they need faster, reducing frustration

🔍

📉

Fewer Repetitive Inquiries

Sellers spending less time answering the same questions

🔍

Better Information Discovery

Users finding what they need faster, reducing frustration

Impacts & Results

A Faster, More Relevant Chat Experience

The revamped chatbot is designed to improve user experience and seller efficiency. While the feature was still in the rollout phase, we expected to see:

Learnings & Takeaways

Designing for AI and Real Users

This project reinforced key takeaways about designing AI-driven experiences:

AI should enhance, not replace, human interaction

Users still need control over their conversations

Iteration is key

Testing with both animations and static designs helped us refine the final experience

Context is everything

A personalized chatbot experience is only as good as the data that powers it

Learnings & Takeaways

Designing for AI and Real Users

This project reinforced key takeaways about designing AI-driven experiences:

✈️ Future Opportunities

Further refine AI accuracy based on real user interactions
Expand personalization by incorporating user shopping history
Experiment with conversational UI elements for richer engagement

Further refine AI accuracy based on real user interactions
Expand personalization by incorporating user shopping history
Experiment with conversational UI elements for richer engagement

Next Project

Empowering Advertiser with a Self-Serve Media Buying Platform

Making Conversation Between Sellers & Buyers Smarter with AI

Rethinking the Chatbot:

How Can We Make Conversation Between Sellers & Buyers Faster & Smarter?

Role

Team

Timeline

A Traditional One-Size-Fits-All Chat Experience

Looking Around - Analysing What Are Others Doing?

How did we tackle the problem?

What did we create, and how did it improve the user experience?

1

✨Personalised greeting

🎨 Modernised UI

🧠 AI-powered suggestions

2

🧾 Product-aware suggestions

🧾 Product-aware suggestions

💬 Smart guidance

💬 Smart guidance

3

📦 Order-aware quick actions

📦 Order-aware quick actions

🤖 Guided prompts

🤖 Guided prompts

Fewer Repetitive Inquiries

Better Information Discovery

A Faster, More Relevant Chat Experience

Designing for AI and Real Users

Next Project

Next Project

Contact Me

Contact Me