AI-102 Question 322
Select 3You are developing a virtual assistant using Azure Bot Framework to handle customer inquiries about products. The assistant needs to conduct multi-turn conversations, where it asks follow-up questions to gather more information before providing a response. Which of the following steps are required to implement multi-turn conversations effectively?
- A
Use dialog classes like WaterfallDialog to define the flow of the conversation.
- B
Store user-specific conversation context in Azure Blob Storage.
- C
Implement state management using Bot Framework's ConversationState or UserState.
- D
Define intents and entities using Azure Language Understanding (LUIS) to understand user input.
- E
Use QnA Maker exclusively to handle multi-turn conversations.
Show answer and explanation
Correct answers: A, C, D
Explanation
To implement multi-turn conversations, you need to use tools and techniques like dialog classes (e.g., WaterfallDialog) to structure the flow, state management (ConversationState or UserState) to persist information across conversation turns, and LUIS to understand user intents and entities. While Azure Blob Storage and QnA Maker serve other purposes, they are not sufficient or specifically suited for handling multi-turn conversations.
- A. Correct.
WaterfallDialog and other dialog classes in the Bot Framework allow you to structure and manage the flow of multi-turn conversations effectively, making it a necessary component.
- B. Incorrect.
While Azure Blob Storage is a storage solution, it is not specifically designed for managing user-specific conversation context, making it an incorrect step for this scenario.
- C. Correct.
State management using ConversationState or UserState is essential to track and manage information across multiple turns in a conversation.
- D. Correct.
LUIS helps identify intents and extract entities from user input, which is critical for enabling the virtual assistant to understand and respond appropriately during multi-turn conversations.
- E. Incorrect.
QnA Maker is primarily designed for FAQ and single-turn question answering and does not fully support complex multi-turn conversation scenarios on its own.