CCAO-F - Governance, Risk, and Responsible Use (15% of the exam) - Section 6.1

Identify appropriate and inappropriate use cases.

Telling apart work Claude suits, such as drafting, summarising and analysis with human review, from uses that are inappropriate, such as unreviewed decisions about people or anything prohibited by policy.

appropriate use casesinappropriate use casesdecisions about people

Practice question for this objective

Free sampleGovernance, Risk, and Responsible Usemedium

A procurement officer at a local council has 14 bids for a grounds-maintenance contract, and the evaluation panel meets on Thursday. The council's AI-use policy permits Claude for drafting and summarising internal documents but states that AI must not be used to score or rank tender bids. How should the officer use Claude to prepare for the panel?

AUse Claude to draft scoring guidance for the panel from the published evaluation criteria, leaving all scoring to the panel Correct
BHave Claude score each bid against the criteria, then ask the panel to review and adjust any score that looks wrong
CHave Claude score only the price element of each bid, since that is arithmetic rather than a judgement about quality
DHave Claude rank the bids privately to guide the officer's own preparation, without sharing the ranking with the panel

When policy prohibits a use outright, adding human review does not make it acceptable; use Claude only for the tasks the policy permits. Human review turns many uses into appropriate ones, but it cannot override a policy that rules a use out entirely. The council has drawn a line at AI scoring or ranking of bids, so the officer looks for the permitted part of the job, here turning the published criteria into guidance. That keeps the evaluation defensible while still saving preparation time.

Why A is correct: Drafting guidance from the published criteria is a permitted drafting task, and it helps the panel score consistently while every score and ranking is produced by people, as the policy requires.

Why B is wrong: Human review is what usually makes Claude's output acceptable, so this looks responsible. But the policy prohibits AI scoring outright, and a panel adjusting Claude's scores is still relying on AI scores it was told not to use.

Why C is wrong: Price feels mechanical, so it seems outside the spirit of the rule. But price scoring is still scoring a bid, the policy makes no exception for it, and the price score feeds directly into the ranking.

Why D is wrong: Keeping the ranking out of the panel's papers can feel like a safe compromise. But it is still a prohibited use, and a private ranking can shape how the officer briefs the panel without anyone knowing.

See more CCAO-F practice questions, answers explained.

Exam traps in Governance, Risk, and Responsible Use

Answers that look right on this material and are not. Each one is a distractor from a different question in the CCAO-F bank for this domain.

Whether the task stays inside the firm, since internal work carries little risk even when Claude's output is used without review.
Why it is wrong: Internal work often feels lower risk, so this test is appealing. But removing a subcontractor from the approved list is internal and still has real consequences, and unreviewed output can mislead internal decisions as easily as external ones.
Agree, and write all 40 letters by hand, because any letter about a selection outcome counts as an AI decision about a person.
Why it is wrong: This is tempting because it sounds cautious, but it confuses communicating a decision with making one. The panel decided; drafting the letters with review is ordinary drafting work that policy already permits, so abandoning the tool is an over-correction.
Send them as drafted, because thank-you replies are routine and carry little risk
Why it is wrong: Routine messages feel safe to skip review on, especially under a deadline. The policy covers all AI-drafted external messages, and a routine label does not create an exception.

Examworthy is not affiliated with or endorsed by Anthropic. Original, blueprint-aligned practice material only.