Purpose
Use this lesson to turn an unclear request into instructions an AI can act on. The aim is to produce work that a colleague can review and act on, not merely text that sounds convincing.
Lesson Explanation
Scenario: A manager says: “Help with the client update.”
Better instruction: Write a 180-word client update covering the delayed launch, the new date, and the next decision needed.
State the job, the audience, the required facts, and the desired result. Prompt engineering is not magic wording; it is briefing work. A useful prompt removes the gap between what you mean and what the system receives. A good client update is accurate, short, and tells the reader what happens next.
Work in a short cycle: provide the relevant input, state the result and limits, inspect the draft against the acceptance criteria, then revise the instruction when a requirement is missing. Keep the final human decision with the person responsible for the work.
Practice Questions
-
A colleague asks an AI tool: “Help with the vendor email.” Is this a strong prompt? Why or why not?
View Answer
No – it leaves the job, audience, required facts, and desired result undefined, so the AI has to guess at all of them.
-
Rewrite “Help with the vendor email” into an operational instruction, inventing reasonable specifics.
View Answer
Something like: “Write a 100-word email to our vendor confirming the revised delivery date of March 15 and asking them to update the shipping paperwork accordingly.”
-
A manager says, “Make this sound more professional.” What is missing from this instruction, based on this lesson’s core principle?
View Answer
The job, audience, required facts, and desired result are all left unstated – “more professional” alone gives no concrete target.
-
Someone claims prompt engineering is mainly about finding the right “magic phrase” that makes AI produce great results. Is this consistent with this lesson?
View Answer
No – this lesson states prompt engineering is not magic wording, it is briefing work: stating the job, audience, facts, and result.
-
A request states the job and the audience but never says what the final result should look like. What is likely to go wrong?
View Answer
The AI still has room to guess at the desired result, producing something that may not match what was actually needed.
-
What four elements does this lesson say a useful prompt should state?
View Answer
The job, the audience, the required facts, and the desired result.
-
A junior employee submits an AI-drafted document to a client without reviewing it first, assuming the AI’s polish means it’s ready. What does this lesson say about this practice?
View Answer
It conflicts with the principle that the final human decision stays with the person responsible for the work – review before use matters regardless of how polished the output looks.
-
A prompt produces a fluent, well-written paragraph, but it doesn’t answer the actual business question asked. What does this lesson say a “useful prompt” removes?
View Answer
The gap between what you mean and what the system receives – fluency alone doesn’t guarantee the actual intent was captured.
-
What is the recommended working cycle described in this lesson for using AI on a task?
View Answer
Provide the relevant input, state the result and limits, inspect the draft against the acceptance criteria, then revise the instruction when a requirement is missing.
-
A team lead says, “Just ask the AI to write a status update” with no further detail. Applying this lesson’s principle, what should be added before sending that instruction?
View Answer
The specific job (what the update covers), the audience, the required facts to include, and the desired result or length.
-
An AI produces a polished-sounding market analysis, and a manager forwards it to leadership without checking it against the original request. What principle from this lesson does this skip?
View Answer
Inspecting the draft against acceptance criteria before use, and keeping the final human decision with the person responsible for the work.
-
Why does this lesson distinguish prompt engineering from “text that sounds convincing”?
View Answer
Because the actual goal is work a colleague can review and act on – convincing-sounding text that doesn’t meet the real requirement isn’t useful, regardless of how good it reads.