Skip to course content
Free prompt workflow course

Prompting and AI Workflow Design

Module 08 Knowledge Check

Questions: 10 Pass mark: 8 out of 10 Retries: allowed after reading feedback

Instructions

Choose the best answer. The goal is to check whether you can apply evaluation and output review to your own work, not whether you can recall the wording of a unit.

Questions

1. Review should be treated as…

A. what happens if there is time B. a workflow step with an owner, an output and a time budget C. the reader's job D. optional for internal work

2. If checking the output would take longer than doing the task yourself…

A. check anyway B. that is a finding: this task is a poor fit for AI assistance C. skip the check D. shorten the output

3. The worst reviewer of an output is…

A. a junior colleague B. the person who wrote the prompt C. someone outside the team D. the approver

4. The five review questions are accuracy, evidence, tone, usefulness and…

A. length B. completeness C. format D. speed

5. The quickest way to make review efficient is to…

A. read faster B. do the cheap checks first, so quick disqualifiers come first C. check only accuracy D. use a longer rubric

6. It is quicker and more reliable to verify…

A. the conclusion B. the assumptions it depends on C. the tone D. the formatting

7. Asking how the answer would change if each assumption were wrong turns…

A. a list into a ranking B. a rubric into a score C. review into approval D. a draft into a final

8. A three-level scale is preferred to a five-level one because…

A. it is faster to write B. the middle of a five-point scale carries no information C. it is more accurate D. tools prefer it

9. Review must end in…

A. an impression B. one of five decisions: keep, edit, verify, ask again, or reject C. a revision D. an approval

10. If you are editing more than about a third of an output, the right response is…

A. keep editing B. ask again, because it is usually a prompt problem C. reject it D. lower your standards

Answers

1. B

It gets skipped mostly because it was never scheduled.

2. B

Better to learn it before spending the time.

3. B

They read what they meant, like an author missing their own typos.

4. B

It is the hardest, because it asks about what is absent.

5. B

There is no point tracing citations in an output that fails usefulness.

6. B

If an assumption is wrong, the answer usually is too.

7. B

You then check the top two.

8. B

Three forces a decision.

9. B

Naming the decision is what makes the step real.

10. B

You are otherwise doing the work twice.

What to Review