Microsoft AB-731 Practice Test 17
Question 160
At Cedar Analytics, a product manager wants an AI feature accepted because its demo was impressive. Evidence: No documented rubric or blind comparison exists. The business goal is to use a repeatable evaluation rubric with representative raters. The project owner requires a measurable acceptance criterion before signoff. Which evidence would most convincingly demonstrate that the requirement has been met?
- The returned wording matches the approved source and its citation
- Inter-rater agreement and quality meet predefined thresholds
- Forecast error beats a defined naive baseline across key items
- Net benefit remains positive after all incremental costs and errors are included
Explanation
Correct Answer
B:
No documented rubric or blind comparison exists. Inter-rater agreement and quality meet predefined thresholds.
Incorrect Answers
A:
“The returned wording matches the approved source and its citation” validates an outcome in a case where the evidence is: Paraphrasing would change obligations and cannot replace the signed source. This question instead requires an action that can use a repeatable evaluation rubric with representative raters. No documented rubric or blind comparison exists. Inter-rater agreement and quality meet predefined thresholds.
C:
“Forecast error beats a defined naive baseline across key items” validates an outcome in a case where the evidence is: Historical seasonality and promotions drive numerical forecasts. This question instead requires an action that can use a repeatable evaluation rubric with representative raters. No documented rubric or blind comparison exists. Inter-rater agreement and quality meet predefined thresholds.
D:
“Net benefit remains positive after all incremental costs and errors are included” validates an outcome in a case where the evidence is: No baseline handling time, error rate or deployment cost is recorded. This question instead requires an action that can use a repeatable evaluation rubric with representative raters. No documented rubric or blind comparison exists. Inter-rater agreement and quality meet predefined thresholds.
Question 161
At Meridian Imports, an enterprise RAG index contains a malicious imported document. Evidence: The document includes text telling the assistant to export hidden records. The business goal is to treat retrieved content as untrusted and isolate tool permissions. The project owner requires a measurable acceptance criterion before signoff. Which evidence would most convincingly demonstrate that the requirement has been met?
- Both exact and semantic requirements retrieve correct approved text
- Final deck remains understandable using keyboard and screen reader
- Chosen model meets quality and cost limits for its intended task
- No unauthorized tool call or sensitive disclosure occurs
Explanation
Correct Answer
D:
The document includes text telling the assistant to export hidden records. No unauthorized tool call or sensitive disclosure occurs. Instructions inside retrieved documents are untrusted data, not authorization to call tools or expose records. Enforce tool permissions outside the model and treat attempted instruction redirection as hostile content.
Incorrect Answers
A:
“Both exact and semantic requirements retrieve correct approved text” validates an outcome in a case where the evidence is: Vector-only matching risks missing literal identifiers. This question instead requires an action that can treat retrieved content as untrusted and isolate tool permissions. The document includes text telling the assistant to export hidden records. No unauthorized tool call or sensitive disclosure occurs. Instructions inside retrieved documents are untrusted data, not authorization to call tools or expose records. Enforce tool permissions outside the model and treat attempted instruction redirection as hostile content.
B:
“Final deck remains understandable using keyboard and screen reader” validates an outcome in a case where the evidence is: Generated visuals may omit alt text or use low-contrast design. This question instead requires an action that can treat retrieved content as untrusted and isolate tool permissions. The document includes text telling the assistant to export hidden records. No unauthorized tool call or sensitive disclosure occurs. Instructions inside retrieved documents are untrusted data, not authorization to call tools or expose records. Enforce tool permissions outside the model and treat attempted instruction redirection as hostile content.
C:
“Chosen model meets quality and cost limits for its intended task” validates an outcome in a case where the evidence is: Candidate models differ in modality, region, cost and accuracy. This question instead requires an action that can treat retrieved content as untrusted and isolate tool permissions. The document includes text telling the assistant to export hidden records. No unauthorized tool call or sensitive disclosure occurs. Instructions inside retrieved documents are untrusted data, not authorization to call tools or expose records. Enforce tool permissions outside the model and treat attempted instruction redirection as hostile content.
Question 162
At Lakeshore Commerce, an internal assistant confuses HR benefits questions with IT support tickets. Evidence: Different user intents need separate approved answer paths. The business goal is to design agent topics and disambiguation for distinct intents. 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 intent confusion and fallback rate
- Monitor stale citations and missed owner review cycles
- Monitor fabricated chart values and inconsistent slide versions
- Monitor evaluation drift as source documents change
Explanation
Correct Answer
A:
Different user intents need separate approved answer paths. Monitor intent confusion and fallback rate.
Incorrect Answers
B:
“Monitor stale citations and missed owner review cycles” monitors this signal: The source library was not refreshed after recent rule changes. This question instead requires an action that can design agent topics and disambiguation for distinct intents. Different user intents need separate approved answer paths. Monitor intent confusion and fallback rate.
C:
“Monitor fabricated chart values and inconsistent slide versions” monitors this signal: Slides need reviewable editing in the target presentation. This question instead requires an action that can design agent topics and disambiguation for distinct intents. Different user intents need separate approved answer paths. Monitor intent confusion and fallback rate.
D:
“Monitor evaluation drift as source documents change” monitors this signal: The published scores do not measure the company's actual document tasks. This question instead requires an action that can design agent topics and disambiguation for distinct intents. Different user intents need separate approved answer paths. Monitor intent confusion and fallback rate.
Question 163
At Northlake Services, a sales team needs instant short call tags and slower detailed briefs. Evidence: The two tasks have different turnaround and explanation requirements. The business goal is to use different quality-latency targets for distinct AI workflows. 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 retry storms, latency spikes and fallback effectiveness
- Watch fabricated precision and overly confident executive narratives
- Track tails of latency and degraded analysis under load
- Monitor overfitting to exemplar content and stale template wording
Explanation
Correct Answer
C:
The two tasks have different turnaround and explanation requirements. Track tails of latency and degraded analysis under load. A quick status response and a defensible compliance analysis need different latency, evidence and review budgets. Set separate acceptance targets rather than optimize all requests for one average speed.
Incorrect Answers
A:
“Monitor retry storms, latency spikes and fallback effectiveness” monitors this signal: A single dependent service outage stops all customer help. This question instead requires an action that can use different quality-latency targets for distinct AI workflows. The two tasks have different turnaround and explanation requirements. Track tails of latency and degraded analysis under load. A quick status response and a defensible compliance analysis need different latency, evidence and review budgets. Set separate acceptance targets rather than optimize all requests for one average speed.
B:
“Watch fabricated precision and overly confident executive narratives” monitors this signal: The model lacks an approved statistical forecast and uncertainty bounds. This question instead requires an action that can use different quality-latency targets for distinct AI workflows. The two tasks have different turnaround and explanation requirements. Track tails of latency and degraded analysis under load. A quick status response and a defensible compliance analysis need different latency, evidence and review budgets. Set separate acceptance targets rather than optimize all requests for one average speed.
D:
“Monitor overfitting to exemplar content and stale template wording” monitors this signal: Written instructions alone yield inconsistent section order. This question instead requires an action that can use different quality-latency targets for distinct AI workflows. The two tasks have different turnaround and explanation requirements. Track tails of latency and degraded analysis under load. A quick status response and a defensible compliance analysis need different latency, evidence and review budgets. Set separate acceptance targets rather than optimize all requests for one average speed.
Question 164
At Solstice Retail, a transformation team receives dozens of unrelated AI requests. Evidence: No process ranks ideas by value, readiness and risk. The business goal is to govern intake and prioritization for sustainable AI investment. The project owner requires a measurable acceptance criterion before signoff. Which evidence would most convincingly demonstrate that the requirement has been met?
- Only authorized users see approved contextual content
- Similar qualified applicants receive comparable screening treatment
- Funded pilots have defensible priority and named accountability
- Chosen commercial model meets cost and performance requirements
Explanation
Correct Answer
C:
No process ranks ideas by value, readiness and risk. Funded pilots have defensible priority and named accountability.
Incorrect Answers
A:
“Only authorized users see approved contextual content” validates an outcome in a case where the evidence is: Unreviewed permissions could expose confidential material to broader employees. This question instead requires an action that can govern intake and prioritization for sustainable AI investment. No process ranks ideas by value, readiness and risk. Funded pilots have defensible priority and named accountability.
B:
“Similar qualified applicants receive comparable screening treatment” validates an outcome in a case where the evidence is: Performance differs sharply across legally relevant applicant groups. This question instead requires an action that can govern intake and prioritization for sustainable AI investment. No process ranks ideas by value, readiness and risk. Funded pilots have defensible priority and named accountability.
D:
“Chosen commercial model meets cost and performance requirements” validates an outcome in a case where the evidence is: Demand is uncertain and production volume may be seasonal. This question instead requires an action that can govern intake and prioritization for sustainable AI investment. No process ranks ideas by value, readiness and risk. Funded pilots have defensible priority and named accountability.
Question 165
At Bridgewater Logistics, a manager asks Copilot to summarize a restricted executive folder. Evidence: The user lacks access even though the files exist in Microsoft 365. The business goal is to respect Microsoft Graph-connected permissions during grounding. 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 broadened group memberships and source access drift
- Monitor stale work context and access-control changes
- Monitor regression reports and unsupported API changes
- Monitor stale references and accidental collaborator exposure
Explanation
Correct Answer
A:
The user lacks access even though the files exist in Microsoft 365. Monitor broadened group memberships and source access drift.
Incorrect Answers
B:
“Monitor stale work context and access-control changes” monitors this signal: The answer depends on authorized Microsoft 365 work context. This question instead requires an action that can respect Microsoft Graph-connected permissions during grounding. The user lacks access even though the files exist in Microsoft 365. Monitor broadened group memberships and source access drift.
C:
“Monitor regression reports and unsupported API changes” monitors this signal: The successor may change behavior and cost. This question instead requires an action that can respect Microsoft Graph-connected permissions during grounding. The user lacks access even though the files exist in Microsoft 365. Monitor broadened group memberships and source access drift.
D:
“Monitor stale references and accidental collaborator exposure” monitors this signal: Unorganized ad hoc chats repeatedly lose contextual references. This question instead requires an action that can respect Microsoft Graph-connected permissions during grounding. The user lacks access even though the files exist in Microsoft 365. Monitor broadened group memberships and source access drift.
Question 166
At Monarch Energy, an AI outreach draft consistently favors one demographic group. Evidence: Training examples omit important customer contexts. The business goal is to evaluate fairness across representative groups before use. The operating team needs an ongoing check after the initial rollout. Which post-rollout signal is most important for detecting regression or governance drift?
- Watch capacity saturation, failed handoffs and unit cost under peaks
- Monitor disparate error rates and newly biased patterns
- Monitor false visual assertions and camera-domain drift
- Monitor service-location changes and unexpected cross-border transfers
Explanation
Correct Answer
B:
Training examples omit important customer contexts. Monitor disparate error rates and newly biased patterns. Representative test cases are needed to expose subgroup performance gaps and avoid unfair conclusions from narrow training examples.
Incorrect Answers
A:
“Watch capacity saturation, failed handoffs and unit cost under peaks” monitors this signal: The current manual process cannot expand rapidly without hiring. This question instead requires an action that can evaluate fairness across representative groups before use. Training examples omit important customer contexts. Monitor disparate error rates and newly biased patterns. Representative test cases are needed to expose subgroup performance gaps and avoid unfair conclusions from narrow training examples.
C:
“Monitor false visual assertions and camera-domain drift” monitors this signal: Text descriptions alone omit visual indicators needed for triage. This question instead requires an action that can evaluate fairness across representative groups before use. Training examples omit important customer contexts. Monitor disparate error rates and newly biased patterns. Representative test cases are needed to expose subgroup performance gaps and avoid unfair conclusions from narrow training examples.
D:
“Monitor service-location changes and unexpected cross-border transfers” monitors this signal: Data location and transfer obligations may constrain service selection. This question instead requires an action that can evaluate fairness across representative groups before use. Training examples omit important customer contexts. Monitor disparate error rates and newly biased patterns. Representative test cases are needed to expose subgroup performance gaps and avoid unfair conclusions from narrow training examples.
Question 167
At Summit Manufacturing, 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 proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Remediate high-risk data access and confirm pilot prerequisites
- Use AI to prepare evidence but require approved human decision
- Choose an approved in-app experience and reduce unnecessary handoffs
- Assign adoption, technical, change and risk workstream owners
Explanation
Correct Answer
A:
Enterprise repositories contain broad anonymous sharing links. Remediate high-risk data access and confirm pilot prerequisites.
Incorrect Answers
B:
“Use AI to prepare evidence but require approved human decision” is a design or implementation response to the condition: Ambiguous policy exceptions can harm individuals. This question instead requires an action that can run readiness checks before connecting work-grounded assistants. Enterprise repositories contain broad anonymous sharing links. Remediate high-risk data access and confirm pilot prerequisites.
C:
“Choose an approved in-app experience and reduce unnecessary handoffs” is a design or implementation response to the condition: The assistant is not integrated into their ordinary work process. This question instead requires an action that can run readiness checks before connecting work-grounded assistants. Enterprise repositories contain broad anonymous sharing links. Remediate high-risk data access and confirm pilot prerequisites.
D:
“Assign adoption, technical, change and risk workstream owners” is a design or implementation response to the 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. Remediate high-risk data access and confirm pilot prerequisites.
Question 168
At Crownridge Travel, employees abandon Copilot because it requires extra manual copying. Evidence: The assistant is not integrated into their ordinary work process. The business goal is to improve workflow fit rather than pushing usage targets. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Trial a fictional document drafting task before and after integration
- Pilot staged communications with a fictional regional rollout
- Simulate a model change with one missing reviewer approval
- Model fictional teams with different adoption and error rates
Explanation
Correct Answer
A:
The assistant is not integrated into their ordinary work process. Trial a fictional document drafting task before and after integration. Usage targets cannot compensate for a tool absent from the task workflow. Map where users currently work, remove unnecessary context switches and measure resolved tasks rather than logins.
Incorrect Answers
B:
“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 improve workflow fit rather than pushing usage targets. The assistant is not integrated into their ordinary work process. Trial a fictional document drafting task before and after integration. Usage targets cannot compensate for a tool absent from the task workflow. Map where users currently work, remove unnecessary context switches and measure resolved tasks rather than logins.
C:
“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 improve workflow fit rather than pushing usage targets. The assistant is not integrated into their ordinary work process. Trial a fictional document drafting task before and after integration. Usage targets cannot compensate for a tool absent from the task workflow. Map where users currently work, remove unnecessary context switches and measure resolved tasks rather than logins.
D:
“Model fictional teams with different adoption and error rates” tests a proposed change for the condition: Real costs include licenses, training, evaluation and human review. This question instead requires an action that can improve workflow fit rather than pushing usage targets. The assistant is not integrated into their ordinary work process. Trial a fictional document drafting task before and after integration. Usage targets cannot compensate for a tool absent from the task workflow. Map where users currently work, remove unnecessary context switches and measure resolved tasks rather than logins.
Question 169
At Willow Creek Finance, an IT director permits usage-based AI billing without controls. Evidence: Consumption can rise faster than the annual forecast. The business goal is to establish budgets, cost visibility and governance for consumption charges. The proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Demonstrate reviewed tasks, source checking and human escalation
- Set budgets and review usage through supported cost management
- Use region-specific approved playbooks and central minimum standards
- Design least privilege, isolation, authentication and audit controls
Explanation
Correct Answer
B:
Consumption can rise faster than the annual forecast. Set budgets and review usage through supported cost management.
Incorrect Answers
A:
“Demonstrate reviewed tasks, source checking and human escalation” is a design or implementation response to the condition: Users have no transparent examples of error correction or escalation. This question instead requires an action that can establish budgets, cost visibility and governance for consumption charges. Consumption can rise faster than the annual forecast. Set budgets and review usage through supported cost management.
C:
“Use region-specific approved playbooks and central minimum standards” is a design or implementation response to the condition: Availability, privacy law and employee consultation rules differ by region. This question instead requires an action that can establish budgets, cost visibility and governance for consumption charges. Consumption can rise faster than the annual forecast. Set budgets and review usage through supported cost management.
D:
“Design least privilege, isolation, authentication and audit controls” is a design or implementation response to the condition: The agent can call tools and access internal data repositories. This question instead requires an action that can establish budgets, cost visibility and governance for consumption charges. Consumption can rise faster than the annual forecast. Set budgets and review usage through supported cost management.
Question 170
At Cedar Analytics, a pilot reports rising user complaints but lacks request identifiers. Evidence: There is no trace from answer to model version and retrieved sources. The business goal is to make AI outcomes diagnosable without exposing secrets. The proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Retrieve approved current policy passages at query time with citations
- Apply an appropriate re-ranking stage with measured latency impact
- Remove unrelated sensitive fields and apply purpose-specific access
- Log sanitized traces, model version, retrieval IDs and quality outcomes
Explanation
Correct Answer
D:
There is no trace from answer to model version and retrieved sources. Log sanitized traces, model version, retrieval IDs and quality outcomes.
Incorrect Answers
A:
“Retrieve approved current policy passages at query time with citations” is a design or implementation response to the condition: A static model snapshot cannot reliably know the latest approved document. This question instead requires an action that can make AI outcomes diagnosable without exposing secrets. There is no trace from answer to model version and retrieved sources. Log sanitized traces, model version, retrieval IDs and quality outcomes.
B:
“Apply an appropriate re-ranking stage with measured latency impact” is a design or implementation response to the condition: Initial keyword matching is broad and top result quality is poor. This question instead requires an action that can make AI outcomes diagnosable without exposing secrets. There is no trace from answer to model version and retrieved sources. Log sanitized traces, model version, retrieval IDs and quality outcomes.
C:
“Remove unrelated sensitive fields and apply purpose-specific access” is a design or implementation response to the condition: Only scheduling constraints and contact details are needed. This question instead requires an action that can make AI outcomes diagnosable without exposing secrets. There is no trace from answer to model version and retrieved sources. Log sanitized traces, model version, retrieval IDs and quality outcomes.