Answer: Context enrichment

## Explanation **Context enrichment** is the correct answer because: 1. **Definition**: Context enrichment involves adding relevant information or context to prompts to improve the AI model's understanding and response accuracy. 2. **Scenario Analysis**: - The FAQ bot was producing inconsistent answers to policy-related questions - The developer's solution was to update prompts to include excerpts from the company's policy document - This provides the AI model with specific, relevant context to reference when answering questions 3. **Why other options are incorrect**: - **Zero-shot prompting (A)**: This refers to asking a model to perform a task without providing any examples. The scenario shows the developer is adding context, not using zero-shot. - **Negative prompting (C)**: This involves telling the model what NOT to do, which isn't what's happening here. - **Temperature adjustment (D)**: This controls the randomness/creativity of responses, not adding contextual information to prompts. 4. **Best Practice**: Adding relevant documents, guidelines, or specific context to prompts is a common best practice for improving consistency and accuracy in AI responses, especially for domain-specific questions like policy inquiries.

Author: Jin H