AI-102 Question 255
Select 4You are developing a customer service chatbot for an e-commerce company using Azure AI Language. The chatbot must recognize user intents such as 'Check Order Status', 'Cancel Order', and 'Return Item'. Additionally, the chatbot should identify specific entities like 'Order Number' and 'Reason for Return'. Which steps should you take to implement and manage the language understanding model?
- A
Create a custom project in Azure AI Language and define intents and entities.
- B
Train the language understanding model using example utterances for each intent.
- C
Deploy the model to a staging environment and test it with real customer data.
- D
Use Azure Cognitive Search to preprocess customer data before training the model.
- E
Export the language understanding model and integrate it with the chatbot framework.
Show answer and explanation
Correct answers: A, B, C, E
Explanation
To implement and manage a language understanding model using Azure AI Language, you must first define intents and entities to capture the desired behaviors and data. Afterward, training the model with example utterances allows it to learn effectively. Once trained, deploying the model to a staging environment helps validate its performance. Finally, exporting the model and integrating it with the chatbot framework ensures that the chatbot can use the model to process user queries. Azure Cognitive Search, while a powerful tool, is not relevant to this specific task.
- A. Correct.
Correct: Defining intents and entities is a foundational step when creating a language understanding model to map user input to specific actions and extract relevant data.
- B. Correct.
Correct: Training the model with example utterances allows it to learn patterns and improve its ability to recognize intents and extract entities accurately.
- C. Correct.
Correct: Deploying the model to a staging environment enables you to test its performance and make adjustments before using it in production.
- D. Incorrect.
Incorrect: Azure Cognitive Search is a tool for indexing and searching data, not specifically for preprocessing data for training a language understanding model.
- E. Correct.
Correct: Exporting the trained model and integrating it with the chatbot framework ensures that the chatbot can use the model to interpret user input.