先試後買

A. adding a quality check on outputs before they enter the downstream process

B. extracting the standard reference material into a Project to avoid repeating it

C. reorganizing the workflow document's section headings to improve navigation

D. updating the team's shared workspace name to reflect the new workflow process

E. labeling each workflow step with descriptive names to improve readability for new team members

A. a clerical typo in a routine internal calendar invitation

B. inconsistent use of the company's preferred dash style across documents

C. decisions that materially affect a person's employment without human accountability

D. inconsistent tone and formatting across internal HR communications sent to employees

E. outputs that systematically disadvantage members of a protected group

A. Iteration is most productive when each change targets a specifically diagnosed weakness.

B. Repeatedly resubmitting the same prompt is the most reliable way to improve the output.

C. Comparing the output against the original criteria validates whether the iteration helped.

D. Iteration should stop only after every possible alternative phrasing has been tried.

E. More iterations always produce a better output, regardless of what was changed.

A. Summarizing performance-review notes across a team to support a manager's promotion decisions.

B. Drafting a preliminary symptom summary to support a clinician's diagnostic review.

C. Drafting a legal summary that an attorney will review and edit before sharing it with a client.

D. Synthesizing themes across approved customer feedback under aggregation rules.

E. Drafting an internal training summary from approved internal documentation.

A. The step where individual contributors report the highest workload in their weekly status updates

B. The step with the longest queue or the largest backlog of items waiting to enter it

C. The step most recently added to the workflow, based on the assumption that process changes introduce the most friction

D. The step with the highest average cycle time per item across the past month's runs

A. Any model, because reasoning depth does not vary across the lineup.

B. Opus, because it is positioned as the most capable for advanced reasoning on complex, specialized tasks.

C. A non-Anthropic model, because the Claude lineup has no option for complex reasoning.

D. Haiku, because lightweight models are always preferred regardless of complexity.

A. Speed of generation, visual presentation quality, and consistency with the formatting used in previous status reports.

B. Accuracy of facts against the charter, completeness against required reporting fields, and relevance to the audience receiving the report.

C. Length, formatting consistency, and tone appropriateness for a steering committee audience.

D. Readability, narrative flow, and whether the report tells a coherent story about the project's progress.

A. Republish the brief unchanged because the claim sounds authoritative.

B. Accept the assertion because the phrase "well-known" is reassuring.

C. Treat the assertion as unverified, and either source it from an authoritative reference or remove it.

D. Strengthen the assertion by adding more confident wording.

A. The response is a brief internal announcement about an upcoming team all-hands meeting.

B. The response is an internal draft summarizing the associate's notes from a customer call for their own reference.

C. The response is an unchanged draft for a paying customer that includes a service-level commitment.

D. The response is a casual rephrasing of an internal team status update.