Databricks Generative AI Engineer Associate Question 122
Single answerYou are building a customer support chatbot using a generative AI model on Databricks. The chatbot receives a user query: 'I want to track my order.' The model’s default prompt focuses on general customer support. However, to improve the model's response, you decide to augment the prompt with additional context to make it specific to order tracking. Which approach would be the most effective for augmenting the prompt?
- A
Add a fixed phrase such as 'Provide detailed information' to the prompt for all queries.
- B
Extract key terms like 'track' and 'order' from the user's query and append them to the prompt to specify the intent.
- C
Use a generic template for customer interactions without modifying the prompt based on the user's input.
- D
Include relevant fields like 'tracking ID' and 'order date' from the user's input or follow-up, if available, to refine the prompt.
Show answer and explanation
Correct answer: B
Explanation
To improve the AI model's response to the user’s query, augmenting the prompt with key terms and intents derived from the user's input, such as 'track' and 'order,' ensures that the response aligns with the specific context. This approach allows the model to generate more accurate and relevant answers while maintaining flexibility for different queries.
- A. Incorrect.
Adding a fixed phrase like 'Provide detailed information' does not address the specific intent of the user's query and will likely result in generic responses.
- B. Correct.
Extracting key terms like 'track' and 'order' from the user's input helps the model understand the specific context and intent, leading to more relevant and accurate responses.
- C. Incorrect.
Using a generic template for all customer support interactions does not leverage the user's specific input, which can result in less personalized and effective responses.
- D. Incorrect.
Including detailed fields like 'tracking ID' and 'order date' is only effective if such information is explicitly provided by the user. Since the query does not include these fields, this option is not applicable.