How to build an AI chatbot for retail and ecommerce customers: A step-by-step guide

Won Suh author
Won Suh
Solutions Architect
  • Tutorial Type: Basics
  • Reading Time: 10 min
  • Building Time: 20 min
Supercharge your customer journey with custom AI chatbots!

Building an AI chatbot for retail

In this tutorial, we will discover how to build a functional AI chatbot for retail. This tutorial will help you to:

  • Understand how Sendbird’s AI Chatbot enhances retail and e-commerce customer engagement

  • Learn to create and configure AI Chatbots for retail use cases

  • Explore testing, deployment, and sharing of AI chatbots in a retail context

This is relevant because, in today’s competitive retail landscape, delivering exceptional customer service is essential. Sendbird’s AI Chatbot empowers retailers to:

  • Provide instant answers to common customer queries based on grounded answers

  • Offer personalized product recommendations

  • Streamline order tracking, returns, and exchanges

  • Enhance overall customer satisfaction and drive sales conversions

This tutorial will guide you through setting up a retail AI Chatbot using Sendbird’s Bot Studio, from basic configuration to knowledge base settings for well-grounded responses.

Graident background

Build & deploy a custom AI chatbot in minutes for free

What will be covered

  1. Benefits of Sendbird AI Chatbot for Retail

  2. Getting Started with Bot Studio

  3. Creating a Retail-Focused AI Chatbot from Scratch

  4. Testing and Deployment

  5. Enhancing Retail Knowledge with Knowledge Center

  6. Best Practices for Retail Chatbots

  7. Conclusion

1. Benefits of Sendbird AI Chatbot for Retail

Sendbird AI Chatbot can revolutionize the retail customer experience by:

  • Reducing wait times with instant responses.

  • Increasing engagement through personalized interactions.

  • Improving customer support by handling routine inquiries.

  • Boosting sales with tailored product suggestions.

Graident background

Build & deploy a custom AI chatbot in minutes for free

2. Getting Started with Bot Studio

2.1 Accessing Bot Studio

1. Log in to the Sendbird Dashboard using your credentials.

2. Navigate to “AI Chatbot” in the left-hand menu.

3. Select “Manage Bots” to access Bot Studio.

Getting started with Bot Studio

2.2 Understanding the Interface

  • Bot Profile: Configure your chatbot’s image and name.
  • First Message: Set the welcome message customers see.
  • AI Engine & Instructions: Choose the AI model and define your chatbot’s behavior.

3. Creating a Retail-Focused Chatbot

3.1 Setting Up the Bot Profile

1. Click on “Create Bot” in Bot Studio.

2. Under Bot Profile:

  • Profile Image: Upload your brand logo or a friendly avatar.
  • Bot Name: For example, “ShopAssist” or “FashionHelper”.
  • Enable Feedback: Toggle on to allow customer feedback.
Setting Up the Bot Profile

3.2 Crafting the First Message

  • Welcome Message:

Unset

"Hello! Welcome to [Your Store Name]. I'm here to help you find the perfect products. How can I assist you today?"

  • Suggested Replies:
    • “Show me the latest arrivals.”
    • “I need help with my order.”
    • “What are your store hours?”
Chat starters and replies

3.3 Configuring the AI Engine & Instructions

1. Select AI Engine: Choose an engine like GPT4o

2. Instructions: You are a helpful assistant for [Your Store Name], specializing in [your products, e.g., electronics, fashion]. Assist customers with product inquiries, orders, and support. Use a friendly and professional tone.

3. Example Responses:

  • Customer: “Do you have the new [Product Name]?”
  • Bot: “Yes, it’s available in various sizes/colors. Would you like to add it to your cart?”

4. Advanced Settings:

  • Streaming Response: Enable for real-time typing effect.

  • Temperature: Set between 0.2 to 0.5. Otherwise keep default settings

  • Maximum Tokens: Use default unless specific needs arise.

5. Click “Apply Changes”.

AI engine and instructions

4. Testing and Deployment

4.1 Testing the Chatbot

1. Use the built-in chat widget on the right to test interactions.

2. Sample Queries:

  • “Can you recommend a gift for a teenager?”

  • “How do I track my shipment?”

  • “What is your return policy?”

Testing the Chatbot

Note: Without storing a knowledge base of the business, the Ai chatbot will always make up answers or “hallucinate.” This is expected behavior and optimizing prompts or instructions is vital to steering the Bot responses for retail business use.

4.2 Refining Responses

  • Handle Off-Topic Questions:
    • Update Instructions:
  • If a question is unrelated to our products or services, politely inform the customer that you can assist with store-related inquiries.
  • Apply Changes and retest.

4.3 Review Instruction History

  • In AI Engine & Instructions:
    • Select Show history to display 5 recent instruction changes

Review Instruction History

4.4 Deploying the Chatbot

1. Click on “Add to My Website”.

2. Copy the Embed Code provided.

3. Paste the script into your website’s HTML (e.g., before </body> tag).

4. Save and Publish your website (WordPress, blog, etc).

Deploying the chatbot

5. Sharing the AI Chatbot

Direct URL Sharing:

1. In Bot Studio, click “Share”.

2. Copy the URL.

3. Distribute via email newsletters, social media, or QR codes.

Share chatbot

6. Enhancing Retail Knowledge with Knowledge Center

6.1 Accessing Knowledge Center

1. Navigate to “AI Chatbot” in the dashboard.

2. Select “Knowledge center”.

3. Select “Add Source +”.

Add knowledge source

6.2 Adding Knowledge Sources

  • Product Catalogs:
    • Upload CSV or JSON files with product details.
  • FAQs:
    • Include common customer questions and answers.
  • Policy Documents:
    • Add return policies, shipping info, etc.

6.3 Uploading Files (CSV/JSON)

1. Add File. Upload a CSV file with question, answer headers. Export as CSV from an excel file if needed. 

Download link: CSV

Uploading Files

2. Wait for the uploading to be completed

Knowledge center

3. Select the uploaded file for preview details

Select the uploaded file for preview details

6.4 Attaching Knowledge Source to Bot

1. Select “Knowledge source” in the Bot created from before.

Attaching Knowledge Source to Bot

2. Apply Changes

3. Test changes in the Bot Studio

  • Refresh the widget before proceeding
Test changes in the Bot Studio

7. Best Practices for Retail Chatbots

7.1 Crafting Effective Prompts

Define the Audience: Tailor responses to your customer demographics.

Use Clear Language: Be direct and avoid ambiguity.

Provide Examples: Guide the chatbot with sample interactions. (good vs bad)

7.2 Organizing Knowledge Sources

Regular Updates: Keep product info and policies current to minimize hallucination.

Structured Data: Use appropriate formats (CSV for FAQs, Markdown for documents).

Focused Content: Include only relevant information in the knowledge source to prevent confusion.

Graident background

Build & deploy a custom AI chatbot in minutes for free

8. Conclusion

By following this tutorial, you’ve learned how to:

  • Create a retail-focused AI Chatbot to enhance customer engagement.

  • Configure knowledge sources to provide accurate information.

  • Deploy the chatbot on your website and share it across channels.

Regularly update your chatbot to meet evolving customer needs and stay ahead in the retail market.

In the next tutorial, we will cover advanced concepts to extend the capabilities of our retail chatbot using workflows, forms and function calls. 

Please start a free trial to get started right away! You can also request a demo or contact us to learn more!