Part 1: Quiz
Take the Quiz on Prompt Engineering.
Prompt Engineering — Multiple-Choice Test
In the context of raiaAI, what best defines “prompt engineering”? A) Tuning GPU settings for faster inference B) Designing and refining the agent’s instructions to guide its behavior C) Writing marketing copy for AI products D) Converting PDFs into vector embeddings
Which factor most directly determines an agent’s success? A) Choice of programming language B) Quality of the agent’s instructions C) Size of the knowledge base D) Number of team members on the project
Which is NOT listed as a benefit of effective prompt engineering? A) Handles complex scenarios B) Maintains a consistent personality C) Guarantees zero escalations D) Follows business rules
Which section defines the agent’s core identity and primary function? A) Response Guidelines and Constraints B) Escalation Procedures C) Agent Role and Purpose D) Knowledge Sources and Routing
What is the main purpose of the “Communication Style and Tone” section? A) To map query types to documents B) To dictate the agent’s personality and voice C) To set escalation thresholds D) To define logging and analytics
Which guidance belongs in Response Guidelines and Constraints (not Escalation Procedures)? A) “If two failed attempts, escalate.” B) “If user asks for a manager, hand off.” C) “Never promise unreleased features.” D) “Transfer to human on extreme frustration.”
“Always cite the source of your information” is primarily about: A) Personality consistency B) Knowledge routing accuracy C) Response formatting and constraints D) Escalation thresholds
Which is the BEST reason routing improves accuracy? A) It reduces token count B) It directs the agent to the most relevant knowledge sources for each query type C) It increases temperature D) It bypasses human review
In a lead-qualification workflow, which sequence is MOST aligned with the course? A) Pitch → Close → Discount → Upsell B) Greet → Qualify (budget/timeline) → Recommend → Offer to schedule C) Escalate → Apologize → Refund → Close D) Search docs → Send link → End chat → Survey
Which is a correct escalation trigger per the material? A) Any billing question B) User expresses extreme frustration or asks for a manager C) First failed attempt at solving a problem D) Any request longer than 3 sentences
What is the recommended escalation phrasing? A) “That’s not my job; ask someone else.” B) “I can’t help. Goodbye.” C) “I understand this is a complex issue. Let me connect you with a member of our team who can better assist you.” D) “Hold on.”
Which is NOT a listed benefit of AI-assisted instruction writing (Manus/ChatGPT)? A) Faster development cycles B) Consistent structure across projects C) Guaranteed 100% accuracy without review D) Easy iteration and refinement
Which set correctly lists typical client specification components? A) Business/Goals, Functional Requirements, Brand/Communication, Technical Specs B) Networking, GPU clocks, Containerization, DevOps C) Sales quotes, Social posts, Invoices, HR forms D) Only Business Context and Objectives
Why does raiaAI prefer Markdown for instructions? A) It enables encrypted execution B) It provides clean, structured formatting for easy reading and pasting C) It auto-generates citations D) It prevents model hallucinations
Which Markdown elements are emphasized for raiaAI integration? A) Inline CSS, iframes, scripts B) Headings, bullet lists, clear sectioning (e.g., Role, Routing, Escalation) C) Only bold text D) Tables only
In the hands-on lab, which step comes FIRST? A) Test escalation triggers B) Upload knowledge base C) Create a new agent in Launch Pad D) Paste instructions
Which step is part of the lab’s testing phase? A) Change the model architecture B) Ask sample questions and verify citations to the knowledge file C) Rebuild the vector index manually D) Disable escalation
In routing design, a query “I can’t log in and I’m being charged for something I didn’t buy” should route to: A) product_marketing.md only B) password_reset_procedures.md AND billing_dispute_resolution.md C) careers.md AND brand_voice.md D) general_information.md only
Which pairing is correct? A) Persona/Role-play → Adopt a specific, branded character to shape tone and behavior B) Chain-of-thought → Hide reasoning steps from the model C) Few-shot → Provide one example only D) Dynamic instructions → Never change prompts after deployment
When a user asks about “the new product,” the BEST first response is to: A) Guess and give details about any recent launch B) Ask a clarifying question to disambiguate (e.g., “Do you mean SmartWidget 2.0 or GigaGadget Pro?”) C) Apologize and end the chat D) Escalate immediately
Last updated