1.Pre-Deployment Testing #
After completing the bot configuration, storefront owners must conduct a comprehensive test of all core features through the Bot Test interface to ensure it performs as expected before installing it on the storefront. This process does not affect store data.
1.1 Pre-Sales Assistance #
- Ask ShopMate for product recommendations, e.g., “Can you recommend some summer dresses?”
- Ask it to compare products, e.g., “Which of these two phones takes better photos?”
- Ask about product details, e.g., “What material is this clothing made of?”“Is the sizing of these shoes true to size?”
- Ask about promotions, e.g., “Are there any current discounts?”“Is this item on sale?”
- (Proactive popup prompts are coming soon; testing will be available later.)
1.2 Order Tracking #
- Try entering an order number, e.g., “Please check the status of order 123456.”
- Ask about delivery progress, e.g., “When will my package arrive?”
1.3 FAQ Handling #
- Ask about returns and shipping times, e.g., “What is your return policy?”“How long does shipping take?”
- Ask about store policies, e.g., “Will international orders be taxed?”
- Ask questions from your Q&A knowledge base to verify if the bot can answer correctly.
- Try casual conversation, e.g., “Hi”, “Can you help me?”, “I didn’t understand, can you repeat that?”
1.4 Human-AI Collaboration #
- Ask ShopMate something it doesn’t understand and check whether it informs you and transfers to a human agent.
- Intentionally express frustration or dissatisfaction to see if it can recognize the tone and initiate a handoff.
- Click “Not satisfied” on a response and confirm whether ShopMate initiates a handoff.
- Say “I want to talk to a human” and check if it immediately initiates a transfer.
1.5 Multilingual Functionality #
- Change your browser language settings and confirm that ShopMate responds in the correct language.
- Switch languages mid-conversation to test if ShopMate can instantly switch and continue the dialogue.
Please test with varied question formats, multiple languages, and diverse scenarios to ensure ShopMate can respond accurately and hand off smoothly when needed.
Once all tests pass, ShopMate could be deployed to the storefront to deliver a better shopping experience to your customers! 😊
2.Post-Deployment Optimization #
2.1 Review Conversation History to Evaluate AI Performance #
Log in to the Chat History to view all conversations between customers and ShopMate. Focus on whether the bot accurately answered user questions, and identify any missed or incorrect responses. This will help uncover issues and provide a solid foundation for ongoing optimization.
2.2 Upvote and Downvote Feedback #
Customers can give “upvote” or “downvote” feedbacks on the Bot’s responses during conversations. Storefront managers can also view the historical chat records and provide “upvote” or “downvote” feedbacks accordingly. These feedbacks will be directly sent to the ShopMate product algorithm team to help continuously improve AI training.
2.3 Improve the Q&A Knowledge Base (by adding missing knowledge) #
By integrating conversation records and user feedback, when ShopMate fails to answer questions accurately or misses relevant topics entirely, relevant content should be promptly added to the Q&A Knowledge Base. This helps expand ShopMate’s knowledge scope and covers a broader range of user questions.
2.4 Edit and Update Q&A Knowledge (By maintaining existing content) #
Regularly review and inspect the existing Q&A Knowledge Base. Integrating user feedback and changes in business policies, update and optimize the Q&A content to ensure that ShopMate’s responses always remain accurate and aligned with the latest business requirements.