Anthropic CCAR-P (Claude Certified Architect - Professional) Practice Test 90

Question 890

A customer success delivery team is designing a Claude-supported workflow because contract terms restrict where customer records may be processed or stored. Which architectural decision most directly meets that requirement?

  1. Adopt Privacy impact and lawful-basis review
  2. Adopt Least-data-required architecture
  3. Adopt HIPAA responsibility and contract assessment
  4. Adopt Data residency deployment review

Explanation

Correct Answer

D:

The mechanism is specific: the control layer can verify regional processing, logs, backups and support access as a whole. A credible implementation requires contractual residency definition and vendor processing disclosures, while its acceptance test monitors forbidden region transfers found in data-flow audits. The alternative approaches leave the risk that selecting one regional database does not constrain every connected service.

Incorrect Answers

A:

When EU customer profiling may process identifiable personal data, the control layer can involve privacy counsel in lawful basis, purpose and data minimization assessment; whereas the research administration case must guard against the outcome that selecting one regional database does not constrain every connected service.

B:

For a workload where classification only needs a narrow subset of full case records, the control layer can transform and suppress unnecessary personal fields before invoking tools; that does not provide the capability actually needed: the control layer can verify regional processing, logs, backups and support access as a whole.

C:

When a US clinical workflow handles protected health information through vendors, the control layer can verify applicable HIPAA obligations and required business associate arrangements; the alternative cannot replace this key requirement: the control layer can verify regional processing, logs, backups and support access as a whole.

Question 891

An independent assessment of a logistics coordination Claude system reports that one stalled low-priority tool can dominate overall response time. Which change best closes the exposed control gap?

  1. Deploy Adaptive tool timeout budget
  2. Deploy Result-driven model escalation
  3. Deploy Parallelizable workflow optimization
  4. Deploy Pre-generation context pruning

Explanation

Correct Answer

A:

The remediation acts at the relevant design boundary: the evaluation process can use per-dependency deadlines and continue only if required evidence is available. That counteracts the observed hazard that one stalled low-priority tool can dominate overall response time. Rollout should include dependency importance rules and explicit fallback messages, with regression evidence from timeout impact on critical-path latency and grounded-answer rate.

Incorrect Answers

B:

This addresses the failure where swapping every request to a premium tier increases cost without matching benefit, because the evaluation process can raise model capacity selectively after evidence-based difficulty checks; yet it would not correct the observed weakness: one stalled low-priority tool can dominate overall response time.

C:

Where three independent classifications currently run sequentially, the evaluation process can parallelize work while joining outputs under a bounded concurrency cap; but this remedy operates on another boundary, not the case where one stalled low-priority tool can dominate overall response time.

D:

Where redundant retrieved passages consume most of the prompt budget, the evaluation process can deduplicate and rank passages before model invocation; whereas the logistics coordination defect follows the path in which one stalled low-priority tool can dominate overall response time.

Question 892

An enterprise knowledge management service owner is reviewing Map-reduce summarization after a canary rollout. The critical regression would occur if concatenating every note increases cost and can truncate important evidence. What observable acceptance signal belongs in the dashboard?

  1. Track citation support rate on final findings
  2. Track failed handoffs caused by missing upstream artifacts
  3. Track retained key-fact recall and final contradiction rate
  4. Track exception detection recall and unsafe bypass counts

Explanation

Correct Answer

C:

An appropriate gate measures retained key-fact recall and final contradiction rate. It tests the mechanism in which the architecture can summarize bounded chunks then aggregate their structured findings, rather than a cosmetic demonstration. That distinction matters because concatenating every note increases cost and can truncate important evidence.

Incorrect Answers

A:

The alternative indicator checks whether the architecture can gather and validate evidence before asking Claude to draw conclusions when complex risk reports contain claims from many separate source systems; but the acceptance gate must instead inspect retained key-fact recall and final contradiction rate to detect the targeted failure.

B:

The proposed measure would suit a setting where downstream analyses sometimes run before upstream data cleansing finishes, since the architecture can model the task as a DAG and execute nodes only when prerequisites are satisfied; by contrast, the enterprise knowledge management release needs evidence from retained key-fact recall and final contradiction rate.

D:

Where typical requests succeed but rare regulatory exceptions cause serious errors, a relevant check would establish whether the architecture can separate normal and exception branches with stronger checks for the latter; but the acceptance gate must instead inspect retained key-fact recall and final contradiction rate to detect the targeted failure.

Question 893

The manufacturing quality release team needs a numerical gate for Relevance-threshold abstention. Rather than equating deployment with success, it wants to catch cases in which selecting the nearest vector does not make it relevant. Which indicator should it track?

  1. Audit unsupported-answer refusal rate and missed-answer rate
  2. Audit recall gain versus extra retrieval calls
  3. Audit version-correct answer rate on dated cases
  4. Audit exact-field match rate and permission denial correctness

Explanation

Correct Answer

A:

The acceptance measure focuses on unsupported-answer refusal rate and missed-answer rate at the task or workflow boundary. Model fluency would not establish whether the integration can require evidence strength before generating a confident factual answer, and it could conceal cases where selecting the nearest vector does not make it relevant.

Incorrect Answers

B:

The alternative indicator checks whether the integration can create bounded alternate queries and merge deduplicated matches when relevant policy is difficult to find because employees use many aliases; however, that proxy would not demonstrate unsupported-answer refusal rate and missed-answer rate for the selected intervention.

C:

When users ask what rule applied on a particular historical date, the integration can filter versions by effective interval before selecting passages; but the acceptance gate must instead inspect unsupported-answer refusal rate and missed-answer rate to detect the targeted failure.

D:

The proposed measure would suit a setting where customers ask for a current balance identified by account ID, since the integration can retrieve authorized structured records using parameterized queries; by contrast, the manufacturing quality release needs evidence from unsupported-answer refusal rate and missed-answer rate.

Question 894

During an acceptance review in field-service coordination, Reproduction-first debugging is operational but its benefit has not been measured. Auditors are concerned that guessing from one screenshot hides true upstream causes. Which evidence most directly settles the question?

  1. Measure reproducible incident rate and time to confirmed root cause
  2. Measure recurrence rate of formerly resolved incidents
  3. Measure error resolution and schema compliance by tool version
  4. Measure duration attributable to each span and culprit isolation time

Explanation

Correct Answer

A:

Tracking reproducible incident rate and time to confirmed root cause provides direct evidence that the system delivers its intended effect: the engineering workflow can capture minimal reproducible requests and traces before changing behavior. The measure is more informative than prompt counts because the failure being measured is that guessing from one screenshot hides true upstream causes.

Incorrect Answers

B:

When a production agent repeatedly reintroduces a previously fixed defect, the engineering workflow can turn the root cause into a permanent automated test and review the rollout path; but the acceptance gate must instead inspect reproducible incident rate and time to confirmed root cause to detect the targeted failure.

C:

The alternative indicator checks whether the engineering workflow can inspect typed tool errors and validate provider-response contracts when end users see failures only when one MCP connector returns malformed results; by contrast, the field-service coordination release needs evidence from reproducible incident rate and time to confirmed root cause.

D:

The proposed measure would suit a setting where a slow multi-tool run obscures its longest step, since the engineering workflow can correlate model, orchestration and external calls on the critical path; however, that proxy would not demonstrate reproducible incident rate and time to confirmed root cause for the selected intervention.

Question 895

A healthcare operations service passed its happy-path demo, and a reviewer asks to discard Memory separation by trust level. Which gap would the remaining design fail to address?

  1. The possibility that blindly filling the context window can bury critical evidence
  2. The possibility that raw replay increases cost and may reintroduce superseded requests
  3. The possibility that a character-count heuristic can understate model-specific token use
  4. The possibility that persistent contamination can make old tool content govern new decisions

Explanation

Correct Answer

D:

The risk is not a generic model inconvenience: persistent contamination can make old tool content govern new decisions. The proposed safeguard changes that boundary because the model configuration can store verified state separately from untrusted retrieved text and time-bound hints, and the release would need typed memory entries and expiry rules plus state contamination incidents in repeated-session tests to show the risk remains contained.

Incorrect Answers

A:

The competing hazard is that blindly filling the context window can bury critical evidence in a workload where a case includes more background than the model can productively use; whereas removing Memory separation by trust level primarily exposes the risk that persistent contamination can make old tool content govern new decisions.

B:

The competing hazard is that raw replay increases cost and may reintroduce superseded requests in a workload where a long case management dialogue continually replays stale turns; but this is not the immediate loss associated with omitting Memory separation by trust level: persistent contamination can make old tool content govern new decisions.

C:

The competing hazard is that a character-count heuristic can understate model-specific token use in a workload where templates unexpectedly exceed request limits after tool results are appended; however, the named safeguard addresses the distinct hazard that persistent contamination can make old tool content govern new decisions.

Question 896

A design authority for contract administration has established this requirement: multiple environments produce divergent behavior despite using one codebase. Which intervention should be prioritized?

  1. Deploy Prompt and model configuration manifest
  2. Deploy Operational runbook
  3. Deploy Interface and error contract
  4. Deploy End-to-end data-flow diagram

Explanation

Correct Answer

A:

This control changes the relevant system boundary: the delivery team can version prompts, model choices, tools and safety settings together. The team should establish approved enabling configuration and accountable owner before enabling it, then evaluate environment parity and reproducible rollback so that source-control of application code alone does not recreate inference behavior is not hidden by fluent outputs.

Incorrect Answers

B:

When incident responders rely on the original author for every production failure, the delivery team can document symptoms, safe mitigation, escalation and rollback procedures; the missing operational behavior is specific: the delivery team can version prompts, model choices, tools and safety settings together.

C:

For a workload where separate teams implement tool providers and consumers inconsistently, the delivery team can publish request schemas, response semantics, error categories and retries; the current decision instead depends on a specific capability: the delivery team can version prompts, model choices, tools and safety settings together.

D:

If privacy reviewers cannot see where source documents or traces travel, the delivery team can show all data stores, transformations, transfers and trust boundaries; whereas the contract administration case must guard against the outcome that source-control of application code alone does not recreate inference behavior.

Question 897

A procurement services implementation team has approved Data residency deployment review. Its expected behavior is that the control layer can verify regional processing, logs, backups and support access as a whole. What prerequisite belongs on the release-readiness checklist?

  1. Arrange agency authorization requirements and boundary documentation
  2. Arrange contractual residency definition and vendor processing disclosures
  3. Arrange legal review and provider eligibility verification
  4. Arrange controlled retention and immutable evidence identifiers

Explanation

Correct Answer

B:

Readiness depends on contractual residency definition and vendor processing disclosures, not just a design approval. That condition supports the mechanism in which the control layer can verify regional processing, logs, backups and support access as a whole. Without it, selecting one regional database does not constrain every connected service; monitoring forbidden region transfers found in data-flow audits helps verify the deployed result.

Incorrect Answers

A:

Where a US federal workload has explicit authorized-environment constraints, the control layer can validate the complete deployment and service boundary against agency requirements; for this release, the team must first verify contractual residency definition and vendor processing disclosures.

C:

This prerequisite supports a method used when a US clinical workflow handles protected health information through vendors, because the control layer can verify applicable HIPAA obligations and required business associate arrangements; but the production dependency for Data residency deployment review is contractual residency definition and vendor processing disclosures.

D:

This prerequisite supports a method used when auditors need to see when governance controls were enforced for decisions, because the control layer can keep traceable control evidence, access decisions and versioned approvals; for this release, the team must first verify contractual residency definition and vendor processing disclosures.

Question 898

A supply-chain tool orchestrator gives every connector the same long timeout even when only one low-priority lookup is slow. Which latency risk should a per-tool budget prevent?

  1. A failure where cheapest inference is not cheapest successful business outcome
  2. A failure where frequent incidental changes can erase the expected saving
  3. A failure where blindly increasing concurrency can overload downstream quotas
  4. A failure where one stalled low-priority tool can dominate overall response time

Explanation

Correct Answer

D:

This is the direct residual hazard: one stalled low-priority tool can dominate overall response time. The omitted intervention matters because the evaluation process can use per-dependency deadlines and continue only if required evidence is available. A reviewer should demand dependency importance rules and explicit fallback messages and measure timeout impact on critical-path latency and grounded-answer rate before treating the risk as controlled.

Incorrect Answers

A:

If shaving tokens degrades downstream extraction precision, cheapest inference is not cheapest successful business outcome; whereas removing Adaptive tool timeout budget primarily exposes the risk that one stalled low-priority tool can dominate overall response time.

B:

When a common long policy prefix appears in nearly every request, frequent incidental changes can erase the expected saving; but this is not the immediate loss associated with omitting Adaptive tool timeout budget: one stalled low-priority tool can dominate overall response time.

C:

The other exposure arises when three independent classifications currently run sequentially: blindly increasing concurrency can overload downstream quotas. Omitting the requested control instead leaves the more immediate risk that one stalled low-priority tool can dominate overall response time; however, the named safeguard addresses the distinct hazard that one stalled low-priority tool can dominate overall response time.

Question 899

Following a regression in insurance processing, the team establishes that concatenating every note increases cost and can truncate important evidence. Which proposed redesign most directly prevents recurrence?

  1. Introduce Reconciliation of conflicting sources
  2. Introduce Exception-first task splitting
  3. Introduce Map-reduce summarization
  4. Introduce Hierarchical work-breakdown structure

Explanation

Correct Answer

C:

The engineering fix changes the mechanism rather than the presentation: the architecture can summarize bounded chunks then aggregate their structured findings. It requires stable chunk boundaries and merge rules; a regression test should examine retained key-fact recall and final contradiction rate in cases where concatenating every note increases cost and can truncate important evidence.

Incorrect Answers

A:

Where two authoritative repositories return different effective dates, the architecture can separate conflict detection and resolution from answer drafting; but this remedy operates on another boundary, not the case where concatenating every note increases cost and can truncate important evidence.

B:

The alternative reduces the risk that an average-case prompt can hide high-impact failure cases because the architecture can separate normal and exception branches with stronger checks for the latter; however, the insurance processing defect follows the path in which concatenating every note increases cost and can truncate important evidence.

D:

Where a large modernization analysis overwhelms a single unconstrained prompt, the architecture can break the outcome into testable components with explicit dependencies; but this remedy operates on another boundary, not the case where concatenating every note increases cost and can truncate important evidence.

Question 900

An education services pilot cannot be accepted because selecting the nearest vector does not make it relevant. The owner can change one primary architectural control; what should it be?

  1. Introduce Time-aware policy retrieval
  2. Introduce Exact database query through a typed tool
  3. Introduce Selective SQL plus document retrieval
  4. Introduce Relevance-threshold abstention

Explanation

Correct Answer

D:

The reported failure calls for the control under which the integration can require evidence strength before generating a confident factual answer. A prompt-only reminder or unrelated metric would not address the risk that selecting the nearest vector does not make it relevant. Evidence of resolution requires approved enabling configuration and accountable owner and an improvement in unsupported-answer refusal rate and missed-answer rate.

Incorrect Answers

A:

The alternative reduces the risk that always using the newest document can answer a historical question incorrectly because the integration can filter versions by effective interval before selecting passages; however, the education services defect follows the path in which selecting the nearest vector does not make it relevant.

B:

Although the alternative reduces the risk that generative similarity search is unsuitable for exact transactional state because the integration can retrieve authorized structured records using parameterized queries; yet it would not correct the observed weakness: selecting the nearest vector does not make it relevant.

C:

When either retrieval source alone omits one class of necessary evidence, the integration can join exact database results with grounded policy excerpts in the application; but this remedy operates on another boundary, not the case where selecting the nearest vector does not make it relevant.

Claude Certified Architect - Professional (CCAR-P) Practice Tests