Microsoft AB-731 Practice Test 29
Question 280
At Ashwood Public Services, an organization imposes the same lengthy review on all AI use cases. Evidence: A low-risk note formatter and high-stakes hiring scorer carry different risks. The business goal is to scale governance depth to documented impact and autonomy. 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 model change with one missing reviewer approval
- Model fictional employee cohorts under several license mixes
- Route two fictional low- and high-risk proposals through governance
- Simulate a mock workflow with optional and unnecessary tracking
Explanation
Correct Answer
C:
A low-risk note formatter and high-stakes hiring scorer carry different risks. Route two fictional low- and high-risk proposals through governance.
Incorrect Answers
A:
“Simulate a model change with one missing reviewer approval” tests a proposed change for the condition: No second approver checks the release evidence. This question instead requires an action that can scale governance depth to documented impact and autonomy. A low-risk note formatter and high-stakes hiring scorer carry different risks. Route two fictional low- and high-risk proposals through governance.
B:
“Model fictional employee cohorts under several license mixes” tests a proposed change for the condition: Usage patterns vary and some workflows may use included chat capabilities. This question instead requires an action that can scale governance depth to documented impact and autonomy. A low-risk note formatter and high-stakes hiring scorer carry different risks. Route two fictional low- and high-risk proposals through governance.
D:
“Simulate a mock workflow with optional and unnecessary tracking” tests a proposed change for the condition: The proposed system analyzes communications and behavior at scale. This question instead requires an action that can scale governance depth to documented impact and autonomy. A low-risk note formatter and high-stakes hiring scorer carry different risks. Route two fictional low- and high-risk proposals through governance.
Question 281
At Tamarack Logistics, an international service desk wants one summarizer for multilingual requests. Evidence: Evaluation data covers only English and may hide translation losses. The business goal is to validate language capability on actual supported locales. 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 locale-specific error rates and language fallbacks
- Monitor vendor policy changes and data upload exceptions
- Monitor retrieval precision and drifting embedding models
- Monitor recurring vague prompts and unapproved recommendations
Explanation
Correct Answer
A:
Evaluation data covers only English and may hide translation losses. Monitor locale-specific error rates and language fallbacks.
Incorrect Answers
B:
“Monitor vendor policy changes and data upload exceptions” monitors this signal: The vendor retention and training terms were not evaluated. This question instead requires an action that can validate language capability on actual supported locales. Evaluation data covers only English and may hide translation losses. Monitor locale-specific error rates and language fallbacks.
C:
“Monitor retrieval precision and drifting embedding models” monitors this signal: Similarity matching is distinct from composing a final recommendation. This question instead requires an action that can validate language capability on actual supported locales. Evaluation data covers only English and may hide translation losses. Monitor locale-specific error rates and language fallbacks.
D:
“Monitor recurring vague prompts and unapproved recommendations” monitors this signal: The request gives no objective, audience, boundaries or evidence. This question instead requires an action that can validate language capability on actual supported locales. Evaluation data covers only English and may hide translation losses. Monitor locale-specific error rates and language fallbacks.
Question 282
At Evergreen Research, a business analyst wants to connect meeting notes to related documents. Evidence: People, files and collaboration events have permissioned relationships. The business goal is to use available work graph context for authorized cross-app insights. 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 silent duplication and changed sheet structures
- Monitor missing associations and changed access boundaries
- Monitor intent confusion and fallback rate
- Monitor safety policy updates and filter false positives
Explanation
Correct Answer
B:
People, files and collaboration events have permissioned relationships. Monitor missing associations and changed access boundaries. Work graph relationships can connect authorized people, files and collaboration history, but inherited access boundaries still apply. Test whether the requester can view each cited item.
Incorrect Answers
A:
“Monitor silent duplication and changed sheet structures” monitors this signal: Incorrect join keys could double-count expense records. This question instead requires an action that can use available work graph context for authorized cross-app insights. People, files and collaboration events have permissioned relationships. Monitor missing associations and changed access boundaries. Work graph relationships can connect authorized people, files and collaboration history, but inherited access boundaries still apply. Test whether the requester can view each cited item.
C:
“Monitor intent confusion and fallback rate” monitors this signal: Different user intents need separate approved answer paths. This question instead requires an action that can use available work graph context for authorized cross-app insights. People, files and collaboration events have permissioned relationships. Monitor missing associations and changed access boundaries. Work graph relationships can connect authorized people, files and collaboration history, but inherited access boundaries still apply. Test whether the requester can view each cited item.
D:
“Monitor safety policy updates and filter false positives” monitors this signal: The release needs safety checks beyond normal accuracy tests. This question instead requires an action that can use available work graph context for authorized cross-app insights. People, files and collaboration events have permissioned relationships. Monitor missing associations and changed access boundaries. Work graph relationships can connect authorized people, files and collaboration history, but inherited access boundaries still apply. Test whether the requester can view each cited item.
Question 283
At Elmwood Banking, 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. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Test customer comprehension using a fictional chat transcript
- Compare synthetic pre/post workflows with consistent quality scoring
- Test a fictional disputed eligibility decision with review steps
- Test fictional qualified applicants across relevant cohorts
Explanation
Correct Answer
C:
No human owner accepted responsibility for the decision. Test a fictional disputed eligibility decision with review steps.
Incorrect Answers
A:
“Test customer comprehension using a fictional chat transcript” tests a proposed change for the condition: Users cannot understand the limits or source of the interaction. This question instead requires an action that can assign accountable humans for consequential outcomes. No human owner accepted responsibility for the decision. Test a fictional disputed eligibility decision with review steps.
B:
“Compare synthetic pre/post workflows with consistent quality scoring” tests a proposed change for the condition: No customer quality, time savings or cost data were collected. This question instead requires an action that can assign accountable humans for consequential outcomes. No human owner accepted responsibility for the decision. Test a fictional disputed eligibility decision with review steps.
D:
“Test fictional qualified applicants across relevant cohorts” tests a proposed change for the condition: Performance differs sharply across legally relevant applicant groups. This question instead requires an action that can assign accountable humans for consequential outcomes. No human owner accepted responsibility for the decision. Test a fictional disputed eligibility decision with review steps.
Question 284
At Clearwater Partners, a communications group publishes an AI illustration as a photograph of an event. Evidence: The image may falsely imply a real situation occurred. The business goal is to require truthful labeling and suitability review for generated media. 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 accessibility complaints and new interface regressions
- Monitor misleading generated media and brand misuse
- Monitor near misses and unsafe response patterns
- Monitor champion misinformation and coaching consistency
Explanation
Correct Answer
B:
The image may falsely imply a real situation occurred. Monitor misleading generated media and brand misuse.
Incorrect Answers
A:
“Monitor accessibility complaints and new interface regressions” monitors this signal: Keyboard navigation and alternative text were not tested. This question instead requires an action that can require truthful labeling and suitability review for generated media. The image may falsely imply a real situation occurred. Monitor misleading generated media and brand misuse.
C:
“Monitor near misses and unsafe response patterns” monitors this signal: Incorrect advice could cause physical harm. This question instead requires an action that can require truthful labeling and suitability review for generated media. The image may falsely imply a real situation occurred. Monitor misleading generated media and brand misuse.
D:
“Monitor champion misinformation and coaching consistency” monitors this signal: Some champions share risky shortcuts and outdated feature claims. This question instead requires an action that can require truthful labeling and suitability review for generated media. The image may falsely imply a real situation occurred. Monitor misleading generated media and brand misuse.
Question 285
At Prairie Health, a leadership team schedules AI deployment before verifying source permissions. Evidence: Enterprise repositories contain broad anonymous sharing links. The business goal is to run readiness checks before connecting work-grounded assistants. 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 data categories, purposes and potential intrusive inference
- Collect use-case owners, value hypotheses, datasets and risk assessments
- Map stakeholders, skills and handoffs for rollout stages
- Audit permission hygiene, devices, licensing and support capacity
Explanation
Correct Answer
D:
Enterprise repositories contain broad anonymous sharing links. Audit permission hygiene, devices, licensing and support capacity.
Incorrect Answers
A:
“Identify data categories, purposes and potential intrusive inference” is a diagnostic step for the reported condition: The proposed system analyzes communications and behavior at scale. This question instead requires an action that can run readiness checks before connecting work-grounded assistants. Enterprise repositories contain broad anonymous sharing links. Audit permission hygiene, devices, licensing and support capacity.
B:
“Collect use-case owners, value hypotheses, datasets and risk assessments” is a diagnostic step for the reported condition: No process ranks ideas by value, readiness and risk. This question instead requires an action that can run readiness checks before connecting work-grounded assistants. Enterprise repositories contain broad anonymous sharing links. Audit permission hygiene, devices, licensing and support capacity.
C:
“Map stakeholders, skills and handoffs for rollout stages” is a diagnostic step for the reported condition: IT, HR, security and functional leaders run disconnected activities. This question instead requires an action that can run readiness checks before connecting work-grounded assistants. Enterprise repositories contain broad anonymous sharing links. Audit permission hygiene, devices, licensing and support capacity.
Question 286
At Cascade Utilities, an AI procurement assistant ranks suppliers with no supporting rationale. Evidence: The decision-maker must justify awards to an audit committee. The business goal is to require decision evidence and transparent evaluation criteria. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Try a synthetic contract in an approved test environment
- Evaluate mock suppliers with a clear scoring breakdown
- Simulate an endpoint timeout during a fictional support request
- Run a test with two dated versions of a fictional policy
Explanation
Correct Answer
B:
The decision-maker must justify awards to an audit committee. Evaluate mock suppliers with a clear scoring breakdown. A decision audit needs the criteria, source evidence and responsible decision-maker, not just a polished recommendation. Trace each material award claim to reviewed records and documented evaluation rules.
Incorrect Answers
A:
“Try a synthetic contract in an approved test environment” tests a proposed change for the condition: The vendor retention and training terms were not evaluated. This question instead requires an action that can require decision evidence and transparent evaluation criteria. The decision-maker must justify awards to an audit committee. Evaluate mock suppliers with a clear scoring breakdown. A decision audit needs the criteria, source evidence and responsible decision-maker, not just a polished recommendation. Trace each material award claim to reviewed records and documented evaluation rules.
C:
“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 require decision evidence and transparent evaluation criteria. The decision-maker must justify awards to an audit committee. Evaluate mock suppliers with a clear scoring breakdown. A decision audit needs the criteria, source evidence and responsible decision-maker, not just a polished recommendation. Trace each material award claim to reviewed records and documented evaluation rules.
D:
“Run a test with two dated versions of a fictional policy” tests a proposed change for the condition: A static model snapshot cannot reliably know the latest approved document. This question instead requires an action that can require decision evidence and transparent evaluation criteria. The decision-maker must justify awards to an audit committee. Evaluate mock suppliers with a clear scoring breakdown. A decision audit needs the criteria, source evidence and responsible decision-maker, not just a polished recommendation. Trace each material award claim to reviewed records and documented evaluation rules.
Question 287
At Clearwater Partners, an international service desk wants one summarizer for multilingual requests. Evidence: Evaluation data covers only English and may hide translation losses. The business goal is to validate language capability on actual supported locales. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Test fictional tickets in languages from target regions
- Run a dummy transfer instruction against a read-only sandbox
- Test matched fictional cases with only irrelevant demographics varied
- Test semantically related fictional cases with different wording
Explanation
Correct Answer
A:
Evaluation data covers only English and may hide translation losses. Test fictional tickets in languages from target regions.
Incorrect Answers
B:
“Run a dummy transfer instruction against a read-only sandbox” tests a proposed change for the condition: Retrieving an account statement does not authorize transactions. This question instead requires an action that can validate language capability on actual supported locales. Evaluation data covers only English and may hide translation losses. Test fictional tickets in languages from target regions.
C:
“Test matched fictional cases with only irrelevant demographics varied” tests a proposed change for the condition: Training examples omit important customer contexts. This question instead requires an action that can validate language capability on actual supported locales. Evaluation data covers only English and may hide translation losses. Test fictional tickets in languages from target regions.
D:
“Test semantically related fictional cases with different wording” tests a proposed change for the condition: Similarity matching is distinct from composing a final recommendation. This question instead requires an action that can validate language capability on actual supported locales. Evaluation data covers only English and may hide translation losses. Test fictional tickets in languages from target regions.
Question 288
At BrightPath Media, a customer assistant occasionally invents instructions for risky equipment use. Evidence: Incorrect advice could cause physical harm. The business goal is to define safety boundaries and reliable fallback behavior. The project owner requires a measurable acceptance criterion before signoff. Which evidence would most convincingly demonstrate that the requirement has been met?
- Staff understand purpose, boundaries and measurable success
- The deployment is blocked until required signoff exists
- Unsafe situations trigger safe handoff rather than fabricated instructions
- Only proportionate approved data is processed
Explanation
Correct Answer
C:
Incorrect advice could cause physical harm. Unsafe situations trigger safe handoff rather than fabricated instructions.
Incorrect Answers
A:
“Staff understand purpose, boundaries and measurable success” validates an outcome in a case where the evidence is: Employees receive mixed signals about priority and approved use. This question instead requires an action that can define safety boundaries and reliable fallback behavior. Incorrect advice could cause physical harm. Unsafe situations trigger safe handoff rather than fabricated instructions.
B:
“The deployment is blocked until required signoff exists” validates an outcome in a case where the evidence is: No second approver checks the release evidence. This question instead requires an action that can define safety boundaries and reliable fallback behavior. Incorrect advice could cause physical harm. Unsafe situations trigger safe handoff rather than fabricated instructions.
D:
“Only proportionate approved data is processed” validates an outcome in a case where the evidence is: The proposed system analyzes communications and behavior at scale. This question instead requires an action that can define safety boundaries and reliable fallback behavior. Incorrect advice could cause physical harm. Unsafe situations trigger safe handoff rather than fabricated instructions.
Question 289
At Prairie Health, a client success leader asks for the source of AI-generated renewal totals. Evidence: Some underlying work records were revised after the first draft. The business goal is to make source identities and versions visible in answers. The project owner requires a measurable acceptance criterion before signoff. Which evidence would most convincingly demonstrate that the requirement has been met?
- The result excludes restricted relationships and identities
- The external message is sent only after approved corrections
- The model separates supported visual facts from inferred causes
- Every total reconciles to approved source records
Explanation
Correct Answer
D:
Some underlying work records were revised after the first draft. Every total reconciles to approved source records. Dated source links let a reviewer trace each reported figure to the correct data version and identify missing inputs.
Incorrect Answers
A:
Graph permission scoping protects private collaboration relationships. It does not show which revised record version supplied an AI-generated renewal total; dated source-linked calculations do.
B:
“The external message is sent only after approved corrections” validates an outcome in a case where the evidence is: Generated text could contain binding promises beyond approval. This question instead requires an action that can make source identities and versions visible in answers. Some underlying work records were revised after the first draft. Every total reconciles to approved source records. Dated source links let a reviewer trace each reported figure to the correct data version and identify missing inputs.
C:
“The model separates supported visual facts from inferred causes” validates an outcome in a case where the evidence is: Text-only signals cannot describe the damage pattern. This question instead requires an action that can make source identities and versions visible in answers. Some underlying work records were revised after the first draft. Every total reconciles to approved source records. Dated source links let a reviewer trace each reported figure to the correct data version and identify missing inputs.
Question 290
At Cascade Utilities, a released AI classifier becomes less accurate after a product expansion. Evidence: New request types were absent from evaluation data. The business goal is to monitor deployed models and refresh evaluation with new examples. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Evaluate candidate models on identical fictional customer tickets
- Compare a fictional work-grounded summary against source documents
- Replay synthetic old and new product requests
- Test mock package images with scratches and harmless markings
Explanation
Correct Answer
C:
New request types were absent from evaluation data. Replay synthetic old and new product requests. Previously unseen request types can degrade a model after deployment while old benchmark scores remain stable. Capture representative production failures and update the evaluation set before changes are promoted.
Incorrect Answers
A:
“Evaluate candidate models on identical fictional customer tickets” tests a proposed change for the condition: Candidate models differ in modality, region, cost and accuracy. This question instead requires an action that can monitor deployed models and refresh evaluation with new examples. New request types were absent from evaluation data. Replay synthetic old and new product requests. Previously unseen request types can degrade a model after deployment while old benchmark scores remain stable. Capture representative production failures and update the evaluation set before changes are promoted.
B:
“Compare a fictional work-grounded summary against source documents” tests a proposed change for the condition: The answer depends on authorized Microsoft 365 work context. This question instead requires an action that can monitor deployed models and refresh evaluation with new examples. New request types were absent from evaluation data. Replay synthetic old and new product requests. Previously unseen request types can degrade a model after deployment while old benchmark scores remain stable. Capture representative production failures and update the evaluation set before changes are promoted.
D:
“Test mock package images with scratches and harmless markings” tests a proposed change for the condition: Text-only signals cannot describe the damage pattern. This question instead requires an action that can monitor deployed models and refresh evaluation with new examples. New request types were absent from evaluation data. Replay synthetic old and new product requests. Previously unseen request types can degrade a model after deployment while old benchmark scores remain stable. Capture representative production failures and update the evaluation set before changes are promoted.