Microsoft AB-731 Practice Test 12
Question 110
At Marigold Education, a manager blames an AI recommendation for a denied benefit. Evidence: No human owner accepted responsibility for the decision. The business goal is to assign accountable humans for consequential outcomes. The operating team needs an ongoing check after the initial rollout. Which post-rollout signal is most important for detecting regression or governance drift?
- Monitor unauthorized dataset additions and contractual changes
- Monitor unused commitments and high variable charges
- Monitor decisions without owner or appeal process
- Audit sensitive-data leakage and collection creep
Explanation
Correct Answer
C:
No human owner accepted responsibility for the decision. Monitor decisions without owner or appeal process.
Incorrect Answers
A:
“Monitor unauthorized dataset additions and contractual changes” monitors this signal: Rights to read documents do not automatically permit model training. This question instead requires an action that can assign accountable humans for consequential outcomes. No human owner accepted responsibility for the decision. Monitor decisions without owner or appeal process.
B:
“Monitor unused commitments and high variable charges” monitors this signal: Demand is uncertain and production volume may be seasonal. This question instead requires an action that can assign accountable humans for consequential outcomes. No human owner accepted responsibility for the decision. Monitor decisions without owner or appeal process.
D:
“Audit sensitive-data leakage and collection creep” monitors this signal: The project only needs agreed project decisions. This question instead requires an action that can assign accountable humans for consequential outcomes. No human owner accepted responsibility for the decision. Monitor decisions without owner or appeal process.
Question 111
At Redwood Consulting, an AI team proposes tuning a model with licensed reports and confidential contracts. Evidence: Rights to read documents do not automatically permit model training. The business goal is to confirm consent and rights for training data reuse. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Test a mock feature-unavailable ticket under two license profiles
- Route two fictional low- and high-risk proposals through governance
- Test a fictional data inventory for restricted reuse licenses
- Test a fictional disputed eligibility decision with review steps
Explanation
Correct Answer
C:
Rights to read documents do not automatically permit model training. Test a fictional data inventory for restricted reuse licenses.
Incorrect Answers
A:
“Test a mock feature-unavailable ticket under two license profiles” tests a proposed change for the condition: Users do not know whether the issue is access, training or configuration. This question instead requires an action that can confirm consent and rights for training data reuse. Rights to read documents do not automatically permit model training. Test a fictional data inventory for restricted reuse licenses.
B:
“Route two fictional low- and high-risk proposals through governance” tests a proposed change for the condition: A low-risk note formatter and high-stakes hiring scorer carry different risks. This question instead requires an action that can confirm consent and rights for training data reuse. Rights to read documents do not automatically permit model training. Test a fictional data inventory for restricted reuse licenses.
D:
“Test a fictional disputed eligibility decision with review steps” tests a proposed change for the condition: No human owner accepted responsibility for the decision. This question instead requires an action that can confirm consent and rights for training data reuse. Rights to read documents do not automatically permit model training. Test a fictional data inventory for restricted reuse licenses.
Question 112
At Ashwood Public Services, an organization wants descriptions of training images for accessibility. Evidence: Some images contain diagrams whose meaning needs human review. The business goal is to use vision-assisted descriptions as reviewed accessibility drafts. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Simulate a report plan with one deliberate input correction
- Compare old and new models on an approved regression set
- Test fictional slides with a simple chart and decorative photo
- Test synthetic product photos paired with conflicting notes
Explanation
Correct Answer
C:
Some images contain diagrams whose meaning needs human review. Test fictional slides with a simple chart and decorative photo. Vision-generated alternative text is a draft interpretation, especially for diagrams with implied meaning. Human review should verify essential relationships, labels and omissions before publication.
Incorrect Answers
A:
“Simulate a report plan with one deliberate input correction” tests a proposed change for the condition: Restarting could duplicate work and bypass reviewed steps. This question instead requires an action that can use vision-assisted descriptions as reviewed accessibility drafts. Some images contain diagrams whose meaning needs human review. Test fictional slides with a simple chart and decorative photo. Vision-generated alternative text is a draft interpretation, especially for diagrams with implied meaning. Human review should verify essential relationships, labels and omissions before publication.
B:
“Compare old and new models on an approved regression set” tests a proposed change for the condition: The successor may change behavior and cost. This question instead requires an action that can use vision-assisted descriptions as reviewed accessibility drafts. Some images contain diagrams whose meaning needs human review. Test fictional slides with a simple chart and decorative photo. Vision-generated alternative text is a draft interpretation, especially for diagrams with implied meaning. Human review should verify essential relationships, labels and omissions before publication.
D:
“Test synthetic product photos paired with conflicting notes” tests a proposed change for the condition: Text-only models cannot inspect image evidence. This question instead requires an action that can use vision-assisted descriptions as reviewed accessibility drafts. Some images contain diagrams whose meaning needs human review. Test fictional slides with a simple chart and decorative photo. Vision-generated alternative text is a draft interpretation, especially for diagrams with implied meaning. Human review should verify essential relationships, labels and omissions before publication.
Question 113
At BrightPath Media, an insurance group needs consistent structured tone in routine loss summaries. Evidence: Grounding documents are current but outputs repeatedly violate a defined writing format. The business goal is to distinguish behavior adaptation from retrieval of fresh facts. The project owner requires a measurable acceptance criterion before signoff. Which evidence would most convincingly demonstrate that the requirement has been met?
- Both exact identifiers and meaning-based questions find correct sources
- Inter-rater agreement and quality meet predefined thresholds
- Selection remains explainable and aligns with weights agreed in advance
- Formatting improvements hold on unseen cases without erasing factual limits
Explanation
Correct Answer
D:
Grounding documents are current but outputs repeatedly violate a defined writing format. Formatting improvements hold on unseen cases without erasing factual limits.
Incorrect Answers
A:
“Both exact identifiers and meaning-based questions find correct sources” validates an outcome in a case where the evidence is: Matching literal keywords alone fails for synonymous wording. This question instead requires an action that can distinguish behavior adaptation from retrieval of fresh facts. Grounding documents are current but outputs repeatedly violate a defined writing format. Formatting improvements hold on unseen cases without erasing factual limits.
B:
“Inter-rater agreement and quality meet predefined thresholds” validates an outcome in a case where the evidence is: No documented rubric or blind comparison exists. This question instead requires an action that can distinguish behavior adaptation from retrieval of fresh facts. Grounding documents are current but outputs repeatedly violate a defined writing format. Formatting improvements hold on unseen cases without erasing factual limits.
C:
“Selection remains explainable and aligns with weights agreed in advance” validates an outcome in a case where the evidence is: Ideas differ in readiness, benefits and regulatory exposure. This question instead requires an action that can distinguish behavior adaptation from retrieval of fresh facts. Grounding documents are current but outputs repeatedly violate a defined writing format. Formatting improvements hold on unseen cases without erasing factual limits.
Question 114
At Evergreen Research, a field manager needs short approved Copilot guidance while traveling. Evidence: The desktop-only instructions may not describe mobile access or capabilities. The business goal is to match the task to supported mobile features and company controls. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Prepare a mock reply with one unapproved promotion hidden in the thread
- Test a non-sensitive query on managed mobile and web sessions
- Trial with public articles and fictional internal project notes
- Test synthetic policy-bypass prompts under controlled conditions
Explanation
Correct Answer
B:
The desktop-only instructions may not describe mobile access or capabilities. Test a non-sensitive query on managed mobile and web sessions. A desktop feature may not behave identically in mobile clients or under corporate controls. Test the supported mobile workflow and account entitlement rather than assuming parity.
Incorrect Answers
A:
“Prepare a mock reply with one unapproved promotion hidden in the thread” tests a proposed change for the condition: A proposed discount is mentioned but not yet approved. This question instead requires an action that can match the task to supported mobile features and company controls. The desktop-only instructions may not describe mobile access or capabilities. Test a non-sensitive query on managed mobile and web sessions. A desktop feature may not behave identically in mobile clients or under corporate controls. Test the supported mobile workflow and account entitlement rather than assuming parity.
C:
“Trial with public articles and fictional internal project notes” tests a proposed change for the condition: Public web claims and workplace records have different provenance. This question instead requires an action that can match the task to supported mobile features and company controls. The desktop-only instructions may not describe mobile access or capabilities. Test a non-sensitive query on managed mobile and web sessions. A desktop feature may not behave identically in mobile clients or under corporate controls. Test the supported mobile workflow and account entitlement rather than assuming parity.
D:
“Test synthetic policy-bypass prompts under controlled conditions” tests a proposed change for the condition: Performance alone does not demonstrate responsible deployment. This question instead requires an action that can match the task to supported mobile features and company controls. The desktop-only instructions may not describe mobile access or capabilities. Test a non-sensitive query on managed mobile and web sessions. A desktop feature may not behave identically in mobile clients or under corporate controls. Test the supported mobile workflow and account entitlement rather than assuming parity.
Question 115
At Cedar Analytics, a bank proposes testing credit-risk fairness only on AI-generated applicants. Evidence: Synthetic samples may reproduce omissions and hide rare real-world issues. The business goal is to distinguish synthetic test utility from representative production evidence. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Apply a fictional scoring rubric to three competing proposals
- Update one fictional product spec and verify retrieval change
- Run repeat extraction on invoices with the same hidden reference
- Test rare fictional boundary cases alongside sampled actual distributions
Explanation
Correct Answer
D:
Synthetic samples may reproduce omissions and hide rare real-world issues. Test rare fictional boundary cases alongside sampled actual distributions. Synthetic examples are good for safe boundary testing, but generated cases may reproduce the omissions of the prompt. Keep a separately governed representative evaluation set for real-world coverage.
Incorrect Answers
A:
“Apply a fictional scoring rubric to three competing proposals” tests a proposed change for the condition: Ideas differ in readiness, benefits and regulatory exposure. This question instead requires an action that can distinguish synthetic test utility from representative production evidence. Synthetic samples may reproduce omissions and hide rare real-world issues. Test rare fictional boundary cases alongside sampled actual distributions. Synthetic examples are good for safe boundary testing, but generated cases may reproduce the omissions of the prompt. Keep a separately governed representative evaluation set for real-world coverage.
B:
“Update one fictional product spec and verify retrieval change” tests a proposed change for the condition: Its vector index does not reflect new catalog documents. This question instead requires an action that can distinguish synthetic test utility from representative production evidence. Synthetic samples may reproduce omissions and hide rare real-world issues. Test rare fictional boundary cases alongside sampled actual distributions. Synthetic examples are good for safe boundary testing, but generated cases may reproduce the omissions of the prompt. Keep a separately governed representative evaluation set for real-world coverage.
C:
“Run repeat extraction on invoices with the same hidden reference” tests a proposed change for the condition: Creative variation adds no value and can introduce mismatches. This question instead requires an action that can distinguish synthetic test utility from representative production evidence. Synthetic samples may reproduce omissions and hide rare real-world issues. Test rare fictional boundary cases alongside sampled actual distributions. Synthetic examples are good for safe boundary testing, but generated cases may reproduce the omissions of the prompt. Keep a separately governed representative evaluation set for real-world coverage.
Question 116
At Orion Distribution, an AI-assisted employee service provides no way to challenge incorrect answers. Evidence: Users may be unable to correct harmful mistakes. The business goal is to create accessible escalation and correction procedures. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Route two fictional low- and high-risk proposals through governance
- Score two fictional vendors using a documented risk rubric
- Submit a fictional erroneous benefit answer to the appeal path
- Pilot staged communications with a fictional regional rollout
Explanation
Correct Answer
C:
Users may be unable to correct harmful mistakes. Submit a fictional erroneous benefit answer to the appeal path.
Incorrect Answers
A:
“Route two fictional low- and high-risk proposals through governance” tests a proposed change for the condition: A low-risk note formatter and high-stakes hiring scorer carry different risks. This question instead requires an action that can create accessible escalation and correction procedures. Users may be unable to correct harmful mistakes. Submit a fictional erroneous benefit answer to the appeal path.
B:
“Score two fictional vendors using a documented risk rubric” tests a proposed change for the condition: Contract data and prompts may be handled under unclear commitments. This question instead requires an action that can create accessible escalation and correction procedures. Users may be unable to correct harmful mistakes. Submit a fictional erroneous benefit answer to the appeal path.
D:
“Pilot staged communications with a fictional regional rollout” tests a proposed change for the condition: Teams miss updates and do not understand why workflows change. This question instead requires an action that can create accessible escalation and correction procedures. Users may be unable to correct harmful mistakes. Submit a fictional erroneous benefit answer to the appeal path.
Question 117
At Northlake Services, an organization wants descriptions of training images for accessibility. Evidence: Some images contain diagrams whose meaning needs human review. The business goal is to use vision-assisted descriptions as reviewed accessibility drafts. The sponsor needs an evidence-based diagnosis rather than a guessed fix. Which initial investigation best identifies the controlling constraint before any solution is chosen?
- Test sample queries by exact number and paraphrased policy terms
- List required read and write operations separately
- Determine which visual information is essential to the lesson
- Identify relevant harm categories and platform filtering limits
Explanation
Correct Answer
C:
Some images contain diagrams whose meaning needs human review. Determine which visual information is essential to the lesson. Vision-generated alternative text is a draft interpretation, especially for diagrams with implied meaning. Human review should verify essential relationships, labels and omissions before publication.
Incorrect Answers
A:
“Test sample queries by exact number and paraphrased policy terms” is a diagnostic step for the reported condition: Vector similarity alone is not enough to preserve identifiers. This question instead requires an action that can use vision-assisted descriptions as reviewed accessibility drafts. Some images contain diagrams whose meaning needs human review. Determine which visual information is essential to the lesson. Vision-generated alternative text is a draft interpretation, especially for diagrams with implied meaning. Human review should verify essential relationships, labels and omissions before publication.
B:
“List required read and write operations separately” is a diagnostic step for the reported condition: An unnecessary permission increases risk without improving the task. This question instead requires an action that can use vision-assisted descriptions as reviewed accessibility drafts. Some images contain diagrams whose meaning needs human review. Determine which visual information is essential to the lesson. Vision-generated alternative text is a draft interpretation, especially for diagrams with implied meaning. Human review should verify essential relationships, labels and omissions before publication.
D:
“Identify relevant harm categories and platform filtering limits” is a diagnostic step for the reported condition: The release needs safety checks beyond normal accuracy tests. This question instead requires an action that can use vision-assisted descriptions as reviewed accessibility drafts. Some images contain diagrams whose meaning needs human review. Determine which visual information is essential to the lesson. Vision-generated alternative text is a draft interpretation, especially for diagrams with implied meaning. Human review should verify essential relationships, labels and omissions before publication.
Question 118
At Riverbend Publishing, a legal search returns relevant documents below older irrelevant results. Evidence: Initial keyword matching is broad and top result quality is poor. The business goal is to consider semantic ranking after initial retrieval. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Simulate an endpoint timeout during a fictional support request
- Compare smaller and larger models on stratified fictional examples
- Test fictional legal questions whose correct section is lower ranked
- Test retrieval of mock credentials by authorized and denied roles
Explanation
Correct Answer
C:
Initial keyword matching is broad and top result quality is poor. Test fictional legal questions whose correct section is lower ranked.
Incorrect Answers
A:
“Simulate an endpoint timeout during a fictional support request” tests a proposed change for the condition: A single dependent service outage stops all customer help. This question instead requires an action that can consider semantic ranking after initial retrieval. Initial keyword matching is broad and top result quality is poor. Test fictional legal questions whose correct section is lower ranked.
B:
“Compare smaller and larger models on stratified fictional examples” tests a proposed change for the condition: Simple tags may be handled by a lower-cost capable model. This question instead requires an action that can consider semantic ranking after initial retrieval. Initial keyword matching is broad and top result quality is poor. Test fictional legal questions whose correct section is lower ranked.
D:
“Test retrieval of mock credentials by authorized and denied roles” tests a proposed change for the condition: The credentials enable unauthorized paid inference requests. This question instead requires an action that can consider semantic ranking after initial retrieval. Initial keyword matching is broad and top result quality is poor. Test fictional legal questions whose correct section is lower ranked.
Question 119
At Monarch Energy, a finance assistant answers a query and is also asked to transfer funds. Evidence: Retrieving an account statement does not authorize transactions. The business goal is to separate read-only grounding from high-impact actions. The proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Compare forecasting baselines against seasonal ML alternatives
- Return traceable references and abstain where evidence is absent
- Allow grounded answers but require explicit privileged payment authorization
- Publish a documented ranking and choose a feasible limited first trial
Explanation
Correct Answer
C:
Retrieving an account statement does not authorize transactions. Allow grounded answers but require explicit privileged payment authorization.
Incorrect Answers
A:
“Compare forecasting baselines against seasonal ML alternatives” is a design or implementation response to the condition: Historical seasonality and promotions drive numerical forecasts. This question instead requires an action that can separate read-only grounding from high-impact actions. Retrieving an account statement does not authorize transactions. Allow grounded answers but require explicit privileged payment authorization.
B:
“Return traceable references and abstain where evidence is absent” is a design or implementation response to the condition: Retrieval produced a plausible paragraph but not supporting evidence. This question instead requires an action that can separate read-only grounding from high-impact actions. Retrieving an account statement does not authorize transactions. Allow grounded answers but require explicit privileged payment authorization.
D:
“Publish a documented ranking and choose a feasible limited first trial” tests a proposed change for the condition: Ideas differ in readiness, benefits and regulatory exposure. This question instead requires an action that can separate read-only grounding from high-impact actions. Retrieving an account statement does not authorize transactions. Allow grounded answers but require explicit privileged payment authorization.
Question 120
At Solstice Retail, a research platform retrieves many superficially similar but irrelevant paragraphs. Evidence: First-stage keyword search has poor top-result precision. The business goal is to use semantic reranking when it measurably improves relevance. The sponsor needs an evidence-based diagnosis rather than a guessed fix. Which initial investigation best identifies the controlling constraint before any solution is chosen?
- Identify editor permissions, authoritative draft and version owner
- Review human relevance labels and current ranking
- Check chosen model, quota, region and data-processing terms
- Define representative input examples and expected outcomes
Explanation
Correct Answer
B:
First-stage keyword search has poor top-result precision. Review human relevance labels and current ranking. Semantic reranking reorders an initial candidate set by contextual relevance. It can improve top results only if useful candidates were retrieved first and the measured quality gain justifies latency.
Incorrect Answers
A:
“Identify editor permissions, authoritative draft and version owner” is a diagnostic step for the reported condition: Parallel copies cause conflicting versions of one plan. This question instead requires an action that can use semantic reranking when it measurably improves relevance. First-stage keyword search has poor top-result precision. Review human relevance labels and current ranking. Semantic reranking reorders an initial candidate set by contextual relevance. It can improve top results only if useful candidates were retrieved first and the measured quality gain justifies latency.
C:
“Check chosen model, quota, region and data-processing terms” is a diagnostic step for the reported condition: Not every model and deployment option exists in all regions. This question instead requires an action that can use semantic reranking when it measurably improves relevance. First-stage keyword search has poor top-result precision. Review human relevance labels and current ranking. Semantic reranking reorders an initial candidate set by contextual relevance. It can improve top results only if useful candidates were retrieved first and the measured quality gain justifies latency.
D:
“Define representative input examples and expected outcomes” is a diagnostic step for the reported condition: The published scores do not measure the company's actual document tasks. This question instead requires an action that can use semantic reranking when it measurably improves relevance. First-stage keyword search has poor top-result precision. Review human relevance labels and current ranking. Semantic reranking reorders an initial candidate set by contextual relevance. It can improve top results only if useful candidates were retrieved first and the measured quality gain justifies latency.