Microsoft AB-731 Practice Test 22
Question 210
At Elmwood Banking, 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 operating team needs an ongoing check after the initial rollout. Which post-rollout signal is most important for detecting regression or governance drift?
- Monitor missing traces, excess retention and alert thresholds
- Track tails of latency and degraded analysis under load
- Monitor occlusion failures and unseen packaging changes
- Monitor permission inheritance changes and cross-user leakage
Explanation
Correct Answer
A:
There is no trace from answer to model version and retrieved sources. Monitor missing traces, excess retention and alert thresholds.
Incorrect Answers
B:
“Track tails of latency and degraded analysis under load” monitors this signal: The two tasks have different turnaround and explanation requirements. 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. Monitor missing traces, excess retention and alert thresholds.
C:
“Monitor occlusion failures and unseen packaging changes” monitors this signal: Visual interpretation matters more than long-form text creativity. 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. Monitor missing traces, excess retention and alert thresholds.
D:
“Monitor permission inheritance changes and cross-user leakage” monitors this signal: Retrieval happens before checking the caller's permissions. 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. Monitor missing traces, excess retention and alert thresholds.
Question 211
At Tamarack Logistics, a policy director needs a comparative industry briefing from many sources. Evidence: The question requires multi-step evidence gathering rather than quick lookup. The business goal is to select Researcher for deeper cited multi-source synthesis where available. 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 stale sources and unsupported conclusions
- Monitor dropped escalations and unauthorized concession promises
- Monitor missing key relationships and misleading descriptions
- Monitor new anonymous links and inheritance changes
Explanation
Correct Answer
A:
The question requires multi-step evidence gathering rather than quick lookup. Monitor stale sources and unsupported conclusions.
Incorrect Answers
B:
“Monitor dropped escalations and unauthorized concession promises” monitors this signal: The request is beyond published policy and automated authority. This question instead requires an action that can select Researcher for deeper cited multi-source synthesis where available. The question requires multi-step evidence gathering rather than quick lookup. Monitor stale sources and unsupported conclusions.
C:
“Monitor missing key relationships and misleading descriptions” monitors this signal: Some images contain diagrams whose meaning needs human review. This question instead requires an action that can select Researcher for deeper cited multi-source synthesis where available. The question requires multi-step evidence gathering rather than quick lookup. Monitor stale sources and unsupported conclusions.
D:
“Monitor new anonymous links and inheritance changes” monitors this signal: Copilot follows existing file permissions, including mistaken grants. This question instead requires an action that can select Researcher for deeper cited multi-source synthesis where available. The question requires multi-step evidence gathering rather than quick lookup. Monitor stale sources and unsupported conclusions.
Question 212
At Clearwater Partners, a consumer company expects tenfold seasonal chat volume. Evidence: The current manual process cannot expand rapidly without hiring. The business goal is to evaluate whether a bounded AI workflow can scale safely. The proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Include sanitized example inputs and outputs alongside schema rules
- Assign evaluated response modes and service-level objectives by task type
- Design elastic capacity with fallbacks and human handoff
- Return traceable references and abstain where evidence is absent
Explanation
Correct Answer
C:
The current manual process cannot expand rapidly without hiring. Design elastic capacity with fallbacks and human handoff.
Incorrect Answers
A:
“Include sanitized example inputs and outputs alongside schema rules” is a design or implementation response to the condition: Written instructions alone yield inconsistent section order. This question instead requires an action that can evaluate whether a bounded AI workflow can scale safely. The current manual process cannot expand rapidly without hiring. Design elastic capacity with fallbacks and human handoff.
B:
“Assign evaluated response modes and service-level objectives by task type” is a design or implementation response to the condition: The two tasks have different turnaround and explanation requirements. This question instead requires an action that can evaluate whether a bounded AI workflow can scale safely. The current manual process cannot expand rapidly without hiring. Design elastic capacity with fallbacks and human handoff.
D:
“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 evaluate whether a bounded AI workflow can scale safely. The current manual process cannot expand rapidly without hiring. Design elastic capacity with fallbacks and human handoff.
Question 213
At Summit Manufacturing, a policy unit changes internal procedures every week. Evidence: A static model snapshot cannot reliably know the latest approved document. The business goal is to choose retrieval-based grounding for changing private knowledge. The proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Chunk and retrieve relevant passages with source labels
- Retrieve approved current policy passages at query time with citations
- Evaluate suitable vision capabilities with human cross-checks
- Use authorized transcription then a grounded summary with uncertainty labels
Explanation
Correct Answer
B:
A static model snapshot cannot reliably know the latest approved document. Retrieve approved current policy passages at query time with citations.
Incorrect Answers
A:
“Chunk and retrieve relevant passages with source labels” is a design or implementation response to the condition: Critical terms from the oldest section may be omitted or truncated. This question instead requires an action that can choose retrieval-based grounding for changing private knowledge. A static model snapshot cannot reliably know the latest approved document. Retrieve approved current policy passages at query time with citations.
C:
“Evaluate suitable vision capabilities with human cross-checks” is a design or implementation response to the condition: Visual interpretation matters more than long-form text creativity. This question instead requires an action that can choose retrieval-based grounding for changing private knowledge. A static model snapshot cannot reliably know the latest approved document. Retrieve approved current policy passages at query time with citations.
D:
“Use authorized transcription then a grounded summary with uncertainty labels” is a design or implementation response to the condition: Poor transcription can propagate incorrect speaker attribution into summaries. This question instead requires an action that can choose retrieval-based grounding for changing private knowledge. A static model snapshot cannot reliably know the latest approved document. Retrieve approved current policy passages at query time with citations.
Question 214
At Silverpine Group, a department wants agents published directly from personal test environments. Evidence: Unreviewed deployment could expose confidential data or unauthorized tools. The business goal is to establish managed development, test and production controls. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Run fictional queries with deliberately distractor paragraphs
- Try a simulated inventory update with denied and allowed mock roles
- Compare a fictional work-grounded summary against source documents
- Publish a synthetic demo agent only to a test audience
Explanation
Correct Answer
D:
Unreviewed deployment could expose confidential data or unauthorized tools. Publish a synthetic demo agent only to a test audience. A prototype without separated environments and promotion approval can expose real data or invoke live tools. Distinct test/production scopes and controlled release gates contain those changes.
Incorrect Answers
A:
“Run fictional queries with deliberately distractor paragraphs” tests a proposed change for the condition: First-stage keyword search has poor top-result precision. This question instead requires an action that can establish managed development, test and production controls. Unreviewed deployment could expose confidential data or unauthorized tools. Publish a synthetic demo agent only to a test audience. A prototype without separated environments and promotion approval can expose real data or invoke live tools. Distinct test/production scopes and controlled release gates contain those changes.
B:
“Try a simulated inventory update with denied and allowed mock roles” tests a proposed change for the condition: Reading policy is different from invoking a high-impact connector action. This question instead requires an action that can establish managed development, test and production controls. Unreviewed deployment could expose confidential data or unauthorized tools. Publish a synthetic demo agent only to a test audience. A prototype without separated environments and promotion approval can expose real data or invoke live tools. Distinct test/production scopes and controlled release gates contain those changes.
C:
“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 establish managed development, test and production controls. Unreviewed deployment could expose confidential data or unauthorized tools. Publish a synthetic demo agent only to a test audience. A prototype without separated environments and promotion approval can expose real data or invoke live tools. Distinct test/production scopes and controlled release gates contain those changes.
Question 215
At Cedar Analytics, an employee forwards a vendor page containing instructions to ignore system policies. Evidence: The page is untrusted external content rather than a user-approved command. The business goal is to treat lower-trust input as data and defend against instruction injection. The proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Build a multi-year TCO comparison with sensitivity ranges
- Start with one repeatable approved workflow and owner-defined success criteria
- Set ingestion and retirement rules for approved policy documents
- Ignore page instructions and answer only from relevant evidence
Explanation
Correct Answer
D:
The page is untrusted external content rather than a user-approved command. Ignore page instructions and answer only from relevant evidence.
Incorrect Answers
A:
“Build a multi-year TCO comparison with sensitivity ranges” is a design or implementation response to the condition: Integration, human review, logging and employee training are omitted. This question instead requires an action that can treat lower-trust input as data and defend against instruction injection. The page is untrusted external content rather than a user-approved command. Ignore page instructions and answer only from relevant evidence.
B:
“Start with one repeatable approved workflow and owner-defined success criteria” tests a proposed change for the condition: High deployment effort conceals whether one use case can succeed. This question instead requires an action that can treat lower-trust input as data and defend against instruction injection. The page is untrusted external content rather than a user-approved command. Ignore page instructions and answer only from relevant evidence.
C:
“Set ingestion and retirement rules for approved policy documents” is a design or implementation response to the condition: The knowledge index was last refreshed before the recent policy revision. This question instead requires an action that can treat lower-trust input as data and defend against instruction injection. The page is untrusted external content rather than a user-approved command. Ignore page instructions and answer only from relevant evidence.
Question 216
At Solstice Retail, an assistant reads OCR scans where table columns merge incorrectly. Evidence: Corrupt source extraction can make even a grounded answer wrong. The business goal is to validate document extraction before retrieval and generation. 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?
- Inspect schema error types and required field constraints
- Review endpoint dependencies, timeouts and critical workflow priority
- Record baseline cost per resolved case, volume and quality
- Compare OCR output and table cells to sample original scans
Explanation
Correct Answer
D:
Corrupt source extraction can make even a grounded answer wrong. Compare OCR output and table cells to sample original scans. If extraction drops table columns or corrupts words, even a faithfully cited answer can be false. Validate extraction fidelity and parse failures before indexing the content.
Incorrect Answers
A:
“Inspect schema error types and required field constraints” is a diagnostic step for the reported condition: An inconsistent structure causes automation failures. This question instead requires an action that can validate document extraction before retrieval and generation. Corrupt source extraction can make even a grounded answer wrong. Compare OCR output and table cells to sample original scans. If extraction drops table columns or corrupts words, even a faithfully cited answer can be false. Validate extraction fidelity and parse failures before indexing the content.
B:
“Review endpoint dependencies, timeouts and critical workflow priority” is a diagnostic step for the reported condition: A single dependent service outage stops all customer help. This question instead requires an action that can validate document extraction before retrieval and generation. Corrupt source extraction can make even a grounded answer wrong. Compare OCR output and table cells to sample original scans. If extraction drops table columns or corrupts words, even a faithfully cited answer can be false. Validate extraction fidelity and parse failures before indexing the content.
C:
“Record baseline cost per resolved case, volume and quality” is a diagnostic step for the reported condition: No baseline handling time, error rate or deployment cost is recorded. This question instead requires an action that can validate document extraction before retrieval and generation. Corrupt source extraction can make even a grounded answer wrong. Compare OCR output and table cells to sample original scans. If extraction drops table columns or corrupts words, even a faithfully cited answer can be false. Validate extraction fidelity and parse failures before indexing the content.
Question 217
At Silverpine Group, a startup proposes training a foundation model from scratch for FAQ answers. Evidence: The available budget and data are far smaller than required for foundation training. The business goal is to prefer evaluating pretrained hosted models when they meet the need. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Compare prebuilt answers against a mock custom prototype
- Test mock contributor and reviewer identities against deployment actions
- Run a low-risk FAQ prototype with approved public content
- Process mock invoices with rotated and low-contrast pages
Explanation
Correct Answer
C:
The available budget and data are far smaller than required for foundation training. Run a low-risk FAQ prototype with approved public content.
Incorrect Answers
A:
“Compare prebuilt answers against a mock custom prototype” tests a proposed change for the condition: A prebuilt agent may already support search without bespoke code. This question instead requires an action that can prefer evaluating pretrained hosted models when they meet the need. The available budget and data are far smaller than required for foundation training. Run a low-risk FAQ prototype with approved public content.
B:
“Test mock contributor and reviewer identities against deployment actions” tests a proposed change for the condition: Excess access can lead to unreviewed deployments and data exposure. This question instead requires an action that can prefer evaluating pretrained hosted models when they meet the need. The available budget and data are far smaller than required for foundation training. Run a low-risk FAQ prototype with approved public content.
D:
“Process mock invoices with rotated and low-contrast pages” tests a proposed change for the condition: The original files are scans and require text extraction. This question instead requires an action that can prefer evaluating pretrained hosted models when they meet the need. The available budget and data are far smaller than required for foundation training. Run a low-risk FAQ prototype with approved public content.
Question 218
At Redwood Consulting, an assistant summarizes decisions based on mixed drafts and signed policies. Evidence: The documents have different authority and effective dates. The business goal is to establish source-of-truth and lineage requirements. The proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Prefer approved current sources and show provenance in outputs
- Require independent release authorization and auditable rollback
- Remediate oversharing and publish clear data stewardship responsibilities
- Provide feedback capture, human review and response tracking
Explanation
Correct Answer
A:
The documents have different authority and effective dates. Prefer approved current sources and show provenance in outputs.
Incorrect Answers
B:
“Require independent release authorization and auditable rollback” is a design or implementation response to the condition: No second approver checks the release evidence. This question instead requires an action that can establish source-of-truth and lineage requirements. The documents have different authority and effective dates. Prefer approved current sources and show provenance in outputs.
C:
“Remediate oversharing and publish clear data stewardship responsibilities” is a design or implementation response to the condition: Unreviewed permissions could expose confidential material to broader employees. This question instead requires an action that can establish source-of-truth and lineage requirements. The documents have different authority and effective dates. Prefer approved current sources and show provenance in outputs.
D:
“Provide feedback capture, human review and response tracking” is a design or implementation response to the condition: Users may be unable to correct harmful mistakes. This question instead requires an action that can establish source-of-truth and lineage requirements. The documents have different authority and effective dates. Prefer approved current sources and show provenance in outputs.
Question 219
At Marigold Education, a regional government requests AI processing for restricted local records. Evidence: Data location and transfer obligations may constrain service selection. The business goal is to check residency and processing commitments before architecture approval. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Have diverse reviewers independently score fictional outputs
- Compare full-library prompting with top-k retrieval on mock files
- Test a synthetic geography-constrained architecture review
- Test one routine FAQ and one legal exception in a mock session
Explanation
Correct Answer
C:
Data location and transfer obligations may constrain service selection. Test a synthetic geography-constrained architecture review.
Incorrect Answers
A:
“Have diverse reviewers independently score fictional outputs” tests a proposed change for the condition: No documented rubric or blind comparison exists. This question instead requires an action that can check residency and processing commitments before architecture approval. Data location and transfer obligations may constrain service selection. Test a synthetic geography-constrained architecture review.
B:
“Compare full-library prompting with top-k retrieval on mock files” tests a proposed change for the condition: Large retrieved context increases token spending and response time. This question instead requires an action that can check residency and processing commitments before architecture approval. Data location and transfer obligations may constrain service selection. Test a synthetic geography-constrained architecture review.
D:
“Test one routine FAQ and one legal exception in a mock session” tests a proposed change for the condition: Its intended scope is simple FAQ and required documents. This question instead requires an action that can check residency and processing commitments before architecture approval. Data location and transfer obligations may constrain service selection. Test a synthetic geography-constrained architecture review.
Question 220
At Crownridge Travel, a multinational enterprise launches AI pilots in several departments without common oversight. Evidence: Different owners approve conflicting risk exceptions. The business goal is to create cross-functional governance accountability. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Model fictional employee cohorts under several license mixes
- Test a synthetic unauthorized purchase request
- Run a fictional policy exception through the proposed council workflow
- Run a tabletop exercise on a fictional incorrect-policy incident
Explanation
Correct Answer
C:
Different owners approve conflicting risk exceptions. Run a fictional policy exception through the proposed council workflow. Named cross-functional decision rights and escalation paths prevent departments from independently granting incompatible AI-risk exceptions.
Incorrect Answers
A:
Licensing determines which users can access Copilot capabilities, but it cannot resolve conflicting ownership of AI-risk exception decisions. Named sponsors and cross-functional approval roles are required.
B:
“Test a synthetic unauthorized purchase request” tests a proposed change for the condition: The agent's action scope could exceed its intended advisory role. This question instead requires an action that can create cross-functional governance accountability. Different owners approve conflicting risk exceptions. Run a fictional policy exception through the proposed council workflow. Named cross-functional decision rights and escalation paths prevent departments from independently granting incompatible AI-risk exceptions.
D:
“Run a tabletop exercise on a fictional incorrect-policy incident” tests a proposed change for the condition: There is no owner or process for AI-specific incidents. This question instead requires an action that can create cross-functional governance accountability. Different owners approve conflicting risk exceptions. Run a fictional policy exception through the proposed council workflow. Named cross-functional decision rights and escalation paths prevent departments from independently granting incompatible AI-risk exceptions.