Discover how to design an Azure AI chatbot that handles diverse user queries with precision. Learn the most effective strategies for leveraging Azure AI Language services and enhancing customer interaction in e-commerce applications.
Table of Contents
Question
Your organization, Xerigon Inc., is developing a chatbot for its e-commerce application using the question and answering capability of the Azure AI Language service. The chatbot is designed to assist customers by answering questions about the company’s products and policies. You have uploaded the document that contains information and questions on the electronic products.
One of the questions is, “Do you have laser printers?”. The chatbot should return the same responses to multiple questions asked by the user. Questions should be, “Can you display the list of laser printers?” and “Which laser printers are available?”.
Which of the following actions is most effective for handling multiple user query formats and ensuring that the chatbot provides the correct answer for all queries?
A. Delete the existing question, and add more specific questions to the knowledge base.
B. Let the question and answering capability of the service add additional questions automatically when users query.
C. Enable active learning to improve the model based on user interactions.
D. Add question-and-answer pairs manually to the knowledge base.
Answer
D. Add question-and-answer pairs manually to the knowledge base.
Explanation
In the given scenario, you would add question-and-answer (QnA) pairs manually to the knowledge base. Manually adding alternative question variations ensures that the chatbot responds accurately to different phrasings of user queries. This approach allows you to associate multiple similar queries (e.g., “Can you display the list of laser printers?” and “Which laser printers are available?”) with a single answer, improving accuracy and coverage. This is the most effective way to handle multiple user query formats.
To add QnA pairs manually to the knowledge base, you would follow the below-outlined steps:
- Log in to the Azure AI Language service.
- Click the knowledge base to add the QnA pair.
- Click Add QnA pair to add a new QnA pair on the Edit page of the knowledge base (refer to the exhibit).
- Add the specific question and any needed answer fields in the new QnA pair row.
- Alternate phrasing, metadata, and follow-up prompts are optional fields. However, you can change and fill in these fields later if required.
- Click Save and train to view any predictions including the new QnA pair.
You would not let the question and answering capability of the Azure AI Language service add additional questions automatically when users query in the given scenario. The service does not automatically update the knowledge base by adding new user queries. Instead, it provides tools such as active learning to suggest potential improvements, but these still require manual intervention. The chatbot cannot address variations automatically without you explicitly adding new question-and-answer pairs.
You would not delete the existing question and add more specific questions to the knowledge base in the given scenario. This is unnecessary and inefficient. The existing question should be retained, and new variations should be added as alternate questions under the same answer. This approach ensures that all related queries point to the correct response without losing valuable data.
You would not enable active learning to improve the model based on user interactions in the given scenario. Active learning helps refine the knowledge base by suggesting alternate questions based on user interactions. This allows the system to learn from real-world usage and improve over time. However, this requires manual review and approval of suggested changes, making it a supplementary step.
Microsoft Azure AI Engineer Associate AI-102 certification exam practice question and answer (Q&A) dump with detail explanation and reference available free, helpful to pass the Microsoft Azure AI Engineer Associate AI-102 exam and earn Microsoft Azure AI Engineer Associate AI-102 certification.