Anthropic CCAO-F (Claude Certified Associate - Foundations) Practice Test 5

Question 40

Claude cannot retrieve a restricted file from a connected drive even though another employee can open it in their own account. Which option best resolves the stated issue?

  1. Avoid duplicating documents with conflicting titles
  2. Refresh an uploaded file after a controlled update
  3. Audit connector usefulness before keeping it enabled
  4. Check underlying permission when a connector returns no document

Explanation

Correct Answer

D:

Connectors operate within authorized access boundaries. The missing result may reflect permissions rather than a model failure.

Incorrect Answers

A:

Duplicate overlapping references can create ambiguity; maintain a single controlled document or explicit precedence rules, while the requirement here is distinguishing retrieval authorization from generation quality instead of resolving conflicts among repeated knowledge files.

B:

Uploaded material needs lifecycle maintenance; replace the old copy and re-test prompts referencing changed sections; that distinction matters because the specific gap is distinguishing retrieval authorization from generation quality, not keeping static knowledge synchronized with revisions.

C:

Remove unnecessary integrations and periodically check owner and purpose; that is relevant to minimizing unused connector exposure, not the stated need for distinguishing retrieval authorization from generation quality.

Question 41

A scheduling study needs peak-hour counts but the supplied dataset also has government ID numbers for each participant. Which approach best satisfies this requirement?

  1. Use aggregated values when individual rows are unnecessary
  2. Classify data before connecting a new source
  3. Honor a contractual restriction on third-party processing
  4. Strip unique identifiers not needed for the question

Explanation

Correct Answer

D:

Only collect or process fields necessary for the approved purpose. Unused sensitive columns add risk without analytical benefit.

Incorrect Answers

A:

Aggregation can reduce personal-data exposure when designed to prevent disclosure of small groups; that is relevant to reducing person-level detail in analytical inputs, not the stated need for minimizing unnecessary identity attributes.

B:

Determine sensitivity and permitted processing before enabling access; connector availability is not blanket data authorization; that distinction matters because the specific gap is minimizing unnecessary identity attributes, not assessing source sensitivity before access.

C:

Contractual controls define what processing is permitted; obtain appropriate approval or use a compliant alternative; prompt wording cannot waive the restriction, while the requirement here is minimizing unnecessary identity attributes instead of respecting external processing restrictions.

Question 42

A regulator-facing statement has approved dates and amounts but needs less promotional wording; full rewrites keep altering the numbers. What is the most appropriate next action?

  1. Update the core template for a recurring issue
  2. Make change reasons explicit in the defect log
  3. Preserve accepted factual content during a tone revision
  4. Re-run a small acceptance suite after prompt edits

Explanation

Correct Answer

C:

Scope editing to voice and wording while explicitly retaining approved dates, amounts, and obligations. Rechecking protected facts avoids accidental changes to information already verified.

Incorrect Answers

A:

Fix the persistent specification rather than depend on repeated follow-ups; this helps with eliminating a repeated instruction omission while leaving protecting checked facts during editorial change unsolved.

B:

Link each change to a reproducible issue and verification result; this supports evidence-based iteration and rollback, a useful capability for tracing improvements to observed failures but not a solution to protecting checked facts during editorial change.

D:

Regression cases check both targeted fixes and unaffected requirements; prompt success must be measured across representative tasks, yet the scenario asks for protecting checked facts during editorial change rather than verifying that repairs did not break other behavior.

Question 43

A service dataset contains duplicate tickets and inconsistent date formats; Claude is asked to estimate resolution time without data preparation. What is the most defensible approach?

  1. Separate extraction of facts from interpretation
  2. Build a preliminary rubric before evaluating submissions
  3. Insert stakeholder validation between draft and rollout
  4. Design a data-cleaning step before calculating metrics

Explanation

Correct Answer

D:

Cleaning and defining records first stops duplicate observations and malformed dates from corrupting the metric. Calculation quality depends on input quality.

Incorrect Answers

A:

Extracting verifiable events first creates an evidentiary base; interpretation can then be labeled and checked separately instead of contaminating the timeline; this helps with distinguishing events from causal judgments while leaving fixing record quality before calculation unsolved.

B:

An agreed rubric aligns fields, scoring definitions, and exclusions across all submissions; that distinction matters because the specific gap is fixing record quality before calculation, not standardizing evaluation dimensions ahead of review.

C:

A pilot review with users reveals operational gaps before the script becomes standard, a useful capability for testing a proposed process with actual users but not a solution to fixing record quality before calculation.

Question 44

Claude summarizes a transcript as though the CFO approved a budget, but the transcript records only a request for further analysis. What is the best response to this situation?

  1. Check omission of accessibility needs
  2. Investigate fabricated meeting commitments
  3. Identify confirmation bias in evidence selection
  4. Validate that an attribution is not composite

Explanation

Correct Answer

B:

Decision extraction must distinguish proposals, discussion, approvals, and follow-ups. Inventing authorization can trigger costly downstream actions.

Incorrect Answers

A:

A useful plan must account for affected user groups rather than silently adopting a narrow default; that distinction matters because the specific gap is distinguishing proposals from recorded decisions, not including accessibility requirements.

C:

Leading instructions distort evidence selection; ask for a neutral assessment with supportive and adverse indicators and consistent thresholds, while the requirement here is distinguishing proposals from recorded decisions instead of preventing a leading prompt from skewing evaluation.

D:

A quotation must trace to one actual speaker and passage; composite paraphrases should not be presented as verbatim attributable speech; that is relevant to preventing misattributed composite quotes, not the stated need for distinguishing proposals from recorded decisions.

Question 45

A cheap model's drafts need ten minutes of correction while a slightly costlier model's drafts are accepted after one minute. Which method best fits these constraints?

  1. Use a small validated model for simple extraction
  2. Reserve expert review for high-risk conclusions
  3. Measure error-adjusted cost rather than price per request only
  4. Batch similar tasks where practical

Explanation

Correct Answer

C:

Total workflow cost includes human rework and rejected outputs. Per-request price alone can misrepresent economic value.

Incorrect Answers

A:

Routine extraction may not require the highest reasoning tier; validate against representative forms and track exceptions; this helps with reducing model expense on well-defined extraction while leaving comparing end-to-end work cost unsolved.

B:

Quality controls should be risk-weighted rather than eliminated for cost savings, a useful capability for protecting high-impact review gates but not a solution to comparing end-to-end work cost.

D:

Grouping similar work reduces process overhead while keeping task boundaries and traceability, yet the scenario asks for comparing end-to-end work cost rather than improving throughput without losing record boundaries.

Question 46

A service team wants to test whether agents can understand a proposed issue-triage display before engineers build the production interface. What is the best response to this situation?

  1. Draft acceptance tests for a planned solution
  2. Design a human approval checkpoint
  3. Escalate production API and identity architecture
  4. Prototype a screen concept as an Artifact

Explanation

Correct Answer

D:

A lightweight prototype makes layout and decision flow tangible for stakeholder feedback. It is not itself a secure deployed system.

Incorrect Answers

A:

Acceptance tests turn ambitions into observable input-output cases; they guide implementation review without assuming success, yet the scenario asks for exploring interface usability before development rather than defining testable behavior for a solution.

B:

An authorized reviewer at the publication boundary controls external commitments; this helps with keeping approval before external release while leaving exploring interface usability before development unsolved.

C:

Authentication, authorization, and failure recovery need engineering design and organizational controls, a useful capability for routing complex systems engineering to specialists but not a solution to exploring interface usability before development.

Question 47

A new proposal Project has only 'be professional' as its instruction and yields inconsistent results across different writers. Which approach best satisfies this requirement?

  1. Define source precedence explicitly
  2. State escalation conditions for out-of-scope requests
  3. Specify acceptable tone without conflicting prohibitions
  4. Describe role, task and output expectations

Explanation

Correct Answer

D:

Specific role, audience, format, and sourcing rules make behavior testable. Vague adjectives do not define success.

Incorrect Answers

A:

Priority rules should name the governing source and behavior on conflict, while the requirement here is creating actionable reusable behavioral guidance instead of resolving which instruction or source controls.

B:

Instructions can define approval thresholds and when to refer a case to a manager; that is relevant to identifying the point where a human must authorize, not the stated need for creating actionable reusable behavioral guidance.

C:

Coherent instructions avoid contradictory tone rules; resolve conflicts based on the actual audience and purpose, yet the scenario asks for creating actionable reusable behavioral guidance rather than removing inconsistent style directives.

Question 48

A published example omits names but includes a rare job title, a specific incident date, and a tiny office location. What should the responsible employee choose?

  1. Treat model instructions as distinct from technical privacy controls
  2. Confirm permitted cross-border processing
  3. Check indirect reidentification in supposedly anonymous data
  4. Apply the relevant retention and records policy

Explanation

Correct Answer

C:

Combinations of quasi-identifiers can reveal a person. Evaluate linkage risk before publication or AI processing.

Incorrect Answers

A:

A request in the prompt does not establish contractual, account, or system-level handling guarantees; that is relevant to using enforceable privacy mechanisms, not the stated need for preventing identity inference from sparse records.

B:

Authorization must include applicable transfer and processing constraints, yet the scenario asks for preventing identity inference from sparse records rather than validating geographic processing obligations.

D:

Retention, deletion, and access obligations may apply to both input and output, while the requirement here is preventing identity inference from sparse records instead of managing sensitive generated records across their lifecycle.

Question 49

A returns assistant handles standard orders but fails whenever a product was purchased under a time-limited promotion. Which method best fits these constraints?

  1. Add tests for an omitted exception rule
  2. Ask for alternate versions with explicit differences
  3. Escalate when a prompt fix cannot grant data access
  4. Stop using a failing shortcut when error costs rise

Explanation

Correct Answer

A:

Encode the exception and test both promoted and ordinary purchases. Narrow tests expose whether the rule works across conditions.

Incorrect Answers

B:

Name the dimensions to vary and compare versions against the same source facts; this helps with producing meaningfully different drafting alternatives while leaving correcting a missing special-case rule unsolved.

C:

Authorization problems require an approved access change or different workflow, a useful capability for resolving a system access boundary but not a solution to correcting a missing special-case rule.

D:

Measured error costs justify redesign or escalation of hard cases; price per call is not the only optimization criterion; that distinction matters because the specific gap is correcting a missing special-case rule, not correcting a cost-saving method that harms quality.

Question 50

A research summary names three competitors but omits citations for pricing figures; a reviewer can identify the precise unsupported sentences. What is the most appropriate next action?

  1. Compare the revision against an explicit acceptance checklist
  2. Append additional raw notes but leave the required action unspecified during prompt refinement
  3. Request a substantive self-critique against criteria
  4. Revise the prompt when its original goal was ambiguous
  5. Point to the missing evidence and request a correction

Explanation

Correct Answer

E:

Targeted feedback directs repair to identifiable claims and asks for verifiable support. A broad 'improve' instruction does not identify the failure mode.

Incorrect Answers

A:

A checklist catches mandatory omissions independently of how polished the text becomes; that distinction matters because the specific gap is repairing specifically unsupported claims, not verifying that mandatory content survived revision.

B:

More pasted material enlarges context without telling Claude which decision or output matters; it leaves the specified need for repairing specifically unsupported claims unresolved.

C:

A structured critique can surface weaknesses for follow-up; it is a review aid rather than independent proof that the draft is accurate, while the requirement here is repairing specifically unsupported claims instead of identifying risks for subsequent human checking.

D:

Repairing the original definition removes a systematic ambiguity from the task, a useful capability for correcting an ambiguous initial metric but not a solution to repairing specifically unsupported claims.

Claude Certified Associate - Foundations (CCAO-F) Practice Tests