Prompt Engineering Techniques AIE Practice Questions (Page 6)

Questions 26–30

1. 26expert · hard

   A marketing team uses an AI to generate ad copy. They have a rubric that scores copy on clarity, brand voice, and call-to-action strength. After running a prompt, the copy scores high on clarity but low on brand voice and call-to-action. The team wants to improve the prompt. What is the most effective next step?

   • AAsk the AI to 'improve the copy' without specifying what is wrong.
   • BAnalyze the rubric scores, identify the specific weaknesses (brand voice and call-to-action), and refine the prompt to include brand voice guidelines and a strong call-to-action requirement, then retest.
   • CProvide a few examples of ad copy that scored high on all criteria.
   • DAccept the copy because clarity is the most important criterion.
   Reveal Solution Discussion Select an answer first
2. 27foundation · easy

   What is the core idea behind iterative refinement of prompts?

   • AWriting a perfect prompt on the first attempt and never changing it.
   • BTesting a prompt, observing the output, and then adjusting the prompt to improve the result.
   • CUsing the same prompt for every task to ensure consistency.
   • DAsking the AI to refine its own prompt without any human input.
   Reveal Solution Discussion Select an answer first
4. 28application · medium

   A user asks an AI: 'Should I take the job offer?' The AI responds with a list of general pros and cons. The user wanted a decision based on their specific salary, commute, and career goals. What is the main problem with the prompt?

   • AThe user should ask a different AI.
   • BThe AI is not capable of making decisions.
   • CThe AI is biased toward general advice.
   • DThe prompt lacks the user's specific context, such as salary, commute, and career goals, making it impossible for the AI to give a tailored answer.
   Reveal Solution Discussion Select an answer first
5. 29expert · hard

   A healthcare compliance team wants an AI to classify patient feedback as 'compliant' or 'non-compliant' with privacy regulations. The team has a set of feedback items that were manually reviewed. They are concerned about edge cases where feedback contains both compliant and non-compliant elements. Which prompt strategy is most effective?

   • ATell the AI to classify the feedback as 'non-compliant' if it is uncertain.
   • BInstruct the AI to classify based on the presence of any patient name.
   • CProvide a few examples of feedback items with mixed content and their correct classifications, and instruct the AI to explain its reasoning for each classification.
   • DAsk the AI to use its general knowledge of privacy regulations.
   Reveal Solution Discussion Select an answer first
6. 30expert · hard

   A legal team wants an AI to draft non-disclosure agreements (NDAs). The prompt must ensure the AI uses the correct jurisdiction, party names, and effective date. The team also wants to avoid the AI inventing terms not in the template. Which prompt design best meets these needs?

   • AAsk the AI to 'draft a standard NDA' and include the party names.
   • BTell the AI to 'use its knowledge of NDAs' to draft the agreement.
   • CProvide the jurisdiction and ask the AI to 'make it legally binding'.
   • DProvide the template text and instruct the AI to fill in only the placeholders with the provided details.
   Reveal Solution Discussion Select an answer first