24 questions · fictional cases
Can you spot the missing evidence?
A useful answer starts with a careful review.
Work through 24 fictional review decisions. Practise distinguishing a supported statement from an assumption, an accepted action from a proposal and a useful result from a confident guess.
An original learning exercise, not a validated assessment of a person or organisation.
Choose one answer to continue. Your selections stay on this page.
Your learning review
Keep the reasoning, not just the count.
This count describes the fictional questions in this exercise. It is not a diagnosis, certification or assessment of an organisation.
Review the reasoning.
The notes give an action but no deadline. What belongs in the summary?
Suggested answer: State that the deadline is not specified.
A missing date is part of the evidence. Preserve the action and the unknown rather than inventing a commitment.
A current policy and an archived policy disagree. What should the reviewer check first?
Suggested answer: Which version governs the task.
Authority and currency determine which rule applies; wording and document length do not settle the conflict.
A draft quotes a restricted document to someone without access. What is the issue?
Suggested answer: The response has crossed an access boundary.
A citation does not authorise disclosure. Source permissions need to apply to the answer as well as the document.
A meeting note says “we could try this.” How should the summary describe it?
Suggested answer: As a proposal, not an accepted action.
A suggestion is not a commitment. Preserve its status so the reader can see what still needs a decision.
Which trial makes a comparison easier to interpret?
Suggested answer: Use the same cases and acceptance conditions.
Comparable inputs and criteria help separate an actual difference from a change in the test itself.
The answer is fluent but its source cannot be opened. What should the reviewer do?
Suggested answer: Treat the claim as unverified until its support can be checked.
Readable prose is not evidence. An inaccessible source leaves a verification gap that should remain visible.
A person asks for a draft email. Which action needs separate authorisation?
Suggested answer: Sending the email.
Preparing content does not imply permission to contact someone. Keep the external action behind a separate decision.
The answer adds an owner who never accepted the task. What is the best correction?
Suggested answer: Mark ownership unresolved and retain the supporting note.
The summary should not create a new assignment. It should expose the missing ownership decision.
Which observation belongs in a review log?
Suggested answer: The draft omitted a specific accepted action.
A concrete failure can be investigated and retested. Broad impressions do not identify the change needed.
A draft takes two minutes, but checking it takes twenty. What should be measured?
Suggested answer: The complete preparation, drafting, review and correction effort.
A fast first draft can shift effort into review. Measure the full task against comparable quality requirements.
The source set does not answer a question. What is useful behaviour?
Suggested answer: Say what is unknown and what source would be needed.
An explicit unknown preserves the decision boundary and helps identify the next information need.
Which source set is appropriate for an initial learning exercise?
Suggested answer: A small set of clearly fictional documents.
Synthetic material lets people learn the review method before permissions for workplace information are agreed.
Two current sources conflict and neither clearly supersedes the other. What next?
Suggested answer: Surface both and ask the responsible owner to resolve the conflict.
The model should not silently settle an unresolved authority question. Make the conflict available for a human decision.
A workflow worked once on a simple case. What does that establish?
Suggested answer: That one observed result passed that case.
One successful observation has a narrow scope. Broader claims need a defined sample and repeatable evaluation.
Which review condition should be written before a trial?
Suggested answer: What would make the result unacceptable.
A failure condition keeps the trial open to evidence that contradicts the initial idea.
An employee loses access to a source. What should an access test include?
Suggested answer: Whether later retrieval and answers respect the change.
Permission changes matter after initial setup. A test should examine the intended behaviour after access is removed.
The team changes the source set and prompt at the same time. What becomes harder?
Suggested answer: Knowing which change caused the observed difference.
Changing one factor at a time can help explain a result. Record combined changes explicitly when separation is impractical.
A handoff says a date is requested but unconfirmed. What should the next reviewer see?
Suggested answer: The requested date and the unresolved confirmation.
Preserve both the request and its status. Removing uncertainty changes the meaning of the handoff.
A result receives a high self-reported confidence score. What still matters?
Suggested answer: Whether its claims match the authorised sources.
Confidence is not a substitute for evidence. Check the output against the sources and acceptance conditions.
A recurring error comes from inconsistent status labels. What should be examined?
Suggested answer: The source and workflow conventions as well as the prompt.
Some problems begin in the underlying work. Better instructions cannot reliably repair an undefined status convention.
What makes a stop condition useful?
Suggested answer: It names an observable trigger and a responsible decision-maker.
A specific trigger and owner make it possible to pause when evidence is missing or the boundary has been crossed.
A reviewer edits an unsupported statement. What should be retained for the next trial?
Suggested answer: The failure, supporting source and correction.
Keeping the failure makes the next trial repeatable and helps show whether the change addressed the actual problem.
Which conclusion can follow from a synthetic timing exercise?
Suggested answer: The arithmetic for those illustrative inputs.
Synthetic values support an illustration only. They do not establish measured outcomes or findings about a population.
The evidence does not justify a wider trial. What is a valid next decision?
Suggested answer: Stop or revise the bounded exercise.
An evaluation is useful when it supports a justified decision, including a decision not to continue.
Read the case before choosing.
Each question describes a small fictional task or review decision. Choose the response that preserves the evidence and the action boundary. You can go back to change an answer before finishing. The exercise remembers selections only while this page stays open; restarting clears them.
The result shows how many answers matched the explanations written for this exercise. It is not a benchmark, a prediction of job performance or a measure of an organisation’s readiness. Use the explanations to discuss the reasoning rather than treating the count as a credential.
Keep uncertainty visible.
Several questions deliberately leave out a date, owner or authoritative source. A useful answer should preserve those gaps. Filling them with a plausible detail changes what the source says and may create a commitment that nobody made.
When current sources disagree, show the disagreement and identify the decision that a responsible person needs to make. A confident tone cannot settle an unresolved policy conflict. The same principle applies when the requested answer is absent from the available evidence.
Separate preparation from permission.
The ability to prepare a draft does not establish permission to send it, update a record or change access. Several cases ask you to recognise the point where an internal exercise would become an external action.
For real work, agree on the approved tools, source permissions and reviewer responsibility before using workplace content. These questions use fictional material so you can practise the reasoning without sharing confidential information or acting on behalf of another person.
Turn an answer into a review habit.
After finishing, choose one explanation that changed your mind. Write a corresponding review question you could apply to another synthetic example: which sentence supports this claim, what remains unresolved or who can approve the next action?
The linked playbook provides a broader planning exercise, and the work review workbook helps make preparation and correction effort visible. Neither resource turns this quiz into a certification. The useful outcome is a clearer question and a more deliberate review.