Microsoft AB-731 Practice Test 39
Question 380
At Tamarack Logistics, 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 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 image resolution, required detection precision and deployment limits
- Distinguish exact extraction from creative brainstorming needs
- Identify which safe metadata can correlate a request with its output
- Test missed queries using synonyms and exact identifiers
Explanation
Correct Answer
C:
There is no trace from answer to model version and retrieved sources. Identify which safe metadata can correlate a request with its output.
Incorrect Answers
A:
“Inspect image resolution, required detection precision and deployment limits” is a diagnostic step for the reported condition: 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. Identify which safe metadata can correlate a request with its output.
B:
“Distinguish exact extraction from creative brainstorming needs” is a diagnostic step for the reported condition: Creative variation adds no value and can introduce mismatches. 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. Identify which safe metadata can correlate a request with its output.
D:
“Test missed queries using synonyms and exact identifiers” is a diagnostic step for the reported condition: Matching literal keywords alone fails for synonymous wording. 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. Identify which safe metadata can correlate a request with its output.
Question 381
At Prairie Health, 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 sponsor needs an evidence-based diagnosis rather than a guessed fix. Which initial investigation best identifies the controlling constraint before any solution is chosen?
- Separate short routing tasks from analysis-heavy summaries
- Record data residency, license, support and evaluation requirements
- Check index refresh time against the policy effective date
- Check whether known categories exist and assess data volume
Explanation
Correct Answer
A:
The two tasks have different turnaround and explanation requirements. Separate short routing tasks from analysis-heavy summaries. 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
B:
“Record data residency, license, support and evaluation requirements” is a diagnostic step for the reported condition: Contractual hosting, customization and governance requirements differ. 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. Separate short routing tasks from analysis-heavy summaries. 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.
C:
“Check index refresh time against the policy effective date” is a diagnostic step for the reported condition: The knowledge index was last refreshed before the recent policy revision. 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. Separate short routing tasks from analysis-heavy summaries. 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:
“Check whether known categories exist and assess data volume” is a diagnostic step for the reported condition: There are no predefined labels and the first goal is discovering patterns. 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. Separate short routing tasks from analysis-heavy summaries. 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 382
At Crownridge Travel, a utility customer asks a bot to negotiate a nonstandard billing adjustment. Evidence: The request is beyond published policy and automated authority. The business goal is to provide human escalation from governed agents. 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 task completion, deflections, errors and escalations
- Identify out-of-scope intents and appropriate case owners
- List connector privileges, approval owners and audit requirements
- Inspect image resolution, skew and required invoice fields
Explanation
Correct Answer
B:
The request is beyond published policy and automated authority. Identify out-of-scope intents and appropriate case owners.
Incorrect Answers
A:
“Inspect task completion, deflections, errors and escalations” is a diagnostic step for the reported condition: Usage metrics do not reveal whether questions were resolved correctly. This question instead requires an action that can provide human escalation from governed agents. The request is beyond published policy and automated authority. Identify out-of-scope intents and appropriate case owners.
C:
“List connector privileges, approval owners and audit requirements” is a diagnostic step for the reported condition: Reading policy is different from invoking a high-impact connector action. This question instead requires an action that can provide human escalation from governed agents. The request is beyond published policy and automated authority. Identify out-of-scope intents and appropriate case owners.
D:
“Inspect image resolution, skew and required invoice fields” is a diagnostic step for the reported condition: The original files are scans and require text extraction. This question instead requires an action that can provide human escalation from governed agents. The request is beyond published policy and automated authority. Identify out-of-scope intents and appropriate case owners.
Question 383
At Redwood Consulting, a claims AI pilot reports high answer accuracy but sometimes exposes private fields. Evidence: Data protection failures are not reflected in the accuracy metric. The business goal is to use multidimensional quality and risk gates. 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 wide confidence intervals and missing difficult subgroups
- Monitor each risk dimension independently after rollout
- Monitor data distribution drift and uncovered roles
- Monitor per-class accuracy and drift after new complaint types appear
Explanation
Correct Answer
B:
Data protection failures are not reflected in the accuracy metric. Monitor each risk dimension independently after rollout. An answer-quality metric misses unauthorized disclosure and unsafe actions. Use separate privacy, authorization and safety thresholds alongside utility metrics before promotion.
Incorrect Answers
A:
“Monitor wide confidence intervals and missing difficult subgroups” monitors this signal: The sample cannot represent real volume and edge cases. This question instead requires an action that can use multidimensional quality and risk gates. Data protection failures are not reflected in the accuracy metric. Monitor each risk dimension independently after rollout. An answer-quality metric misses unauthorized disclosure and unsafe actions. Use separate privacy, authorization and safety thresholds alongside utility metrics before promotion.
C:
“Monitor data distribution drift and uncovered roles” monitors this signal: The target deployment covers regions with very different job families. This question instead requires an action that can use multidimensional quality and risk gates. Data protection failures are not reflected in the accuracy metric. Monitor each risk dimension independently after rollout. An answer-quality metric misses unauthorized disclosure and unsafe actions. Use separate privacy, authorization and safety thresholds alongside utility metrics before promotion.
D:
“Monitor per-class accuracy and drift after new complaint types appear” monitors this signal: The desired answer is a stable label rather than a free-form narrative. This question instead requires an action that can use multidimensional quality and risk gates. Data protection failures are not reflected in the accuracy metric. Monitor each risk dimension independently after rollout. An answer-quality metric misses unauthorized disclosure and unsafe actions. Use separate privacy, authorization and safety thresholds alongside utility metrics before promotion.
Question 384
At Riverbend Publishing, a manager shares a reusable Copilot template containing client account identifiers. Evidence: A prompt may be reusable but embedded personal details are not. The business goal is to share safe generalized prompts without leaking source content. 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?
- List references, publication dates and excluded perspectives
- Review templates for names, passwords, private links and assumptions
- Trace date claims to individual file versions and sender approvals
- Review human relevance labels and current ranking
Explanation
Correct Answer
B:
A prompt may be reusable but embedded personal details are not. Review templates for names, passwords, private links and assumptions. A reusable prompt should preserve task structure while removing identities, secret values and customer-specific evidence. Share generalized instructions rather than confidential examples.
Incorrect Answers
A:
“List references, publication dates and excluded perspectives” is a diagnostic step for the reported condition: The agent draws from a small biased sample of industry articles. This question instead requires an action that can share safe generalized prompts without leaking source content. A prompt may be reusable but embedded personal details are not. Review templates for names, passwords, private links and assumptions. A reusable prompt should preserve task structure while removing identities, secret values and customer-specific evidence. Share generalized instructions rather than confidential examples.
C:
“Trace date claims to individual file versions and sender approvals” is a diagnostic step for the reported condition: Sources differ in authority and may require separate confirmation. This question instead requires an action that can share safe generalized prompts without leaking source content. A prompt may be reusable but embedded personal details are not. Review templates for names, passwords, private links and assumptions. A reusable prompt should preserve task structure while removing identities, secret values and customer-specific evidence. Share generalized instructions rather than confidential examples.
D:
“Review human relevance labels and current ranking” is a diagnostic step for the reported condition: First-stage keyword search has poor top-result precision. This question instead requires an action that can share safe generalized prompts without leaking source content. A prompt may be reusable but embedded personal details are not. Review templates for names, passwords, private links and assumptions. A reusable prompt should preserve task structure while removing identities, secret values and customer-specific evidence. Share generalized instructions rather than confidential examples.
Question 385
At Pinecrest Services, a company wants a governed workspace to evaluate and deploy AI models. Evidence: The team must compare models, tools and safety controls as one lifecycle. The business goal is to use Microsoft Foundry for suitable governed AI development workloads. Only non-sensitive synthetic data and approved sandbox systems may be used. Which bounded trial with synthetic records best tests the proposed approach?
- Query a mock malicious file requesting out-of-scope export
- Publish a synthetic demo agent only to a test audience
- Run a low-risk FAQ prototype with approved public content
- Trial a fictional workload using a limited nonproduction project
Explanation
Correct Answer
D:
The team must compare models, tools and safety controls as one lifecycle. Trial a fictional workload using a limited nonproduction project.
Incorrect Answers
A:
“Query a mock malicious file requesting out-of-scope export” tests a proposed change for the condition: The document includes text telling the assistant to export hidden records. This question instead requires an action that can use Microsoft Foundry for suitable governed AI development workloads. The team must compare models, tools and safety controls as one lifecycle. Trial a fictional workload using a limited nonproduction project.
B:
“Publish a synthetic demo agent only to a test audience” tests a proposed change for the condition: Unreviewed deployment could expose confidential data or unauthorized tools. This question instead requires an action that can use Microsoft Foundry for suitable governed AI development workloads. The team must compare models, tools and safety controls as one lifecycle. Trial a fictional workload using a limited nonproduction project.
C:
“Run a low-risk FAQ prototype with approved public content” tests a proposed change for the condition: The available budget and data are far smaller than required for foundation training. This question instead requires an action that can use Microsoft Foundry for suitable governed AI development workloads. The team must compare models, tools and safety controls as one lifecycle. Trial a fictional workload using a limited nonproduction project.
Question 386
At Redwood Consulting, a firm trains managers centrally but frontline employees avoid the tools. Evidence: Local help and peer examples are not available. The business goal is to build a representative champions network. 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?
- Review user-facing labels and implied authority
- Identify enthusiastic credible colleagues across job families
- Interview staff about failures and consequential risks
- Identify meters, user groups, reporting lag and spending owners
Explanation
Correct Answer
B:
Local help and peer examples are not available. Identify enthusiastic credible colleagues across job families.
Incorrect Answers
A:
“Review user-facing labels and implied authority” is a diagnostic step for the reported condition: Users cannot understand the limits or source of the interaction. This question instead requires an action that can build a representative champions network. Local help and peer examples are not available. Identify enthusiastic credible colleagues across job families.
C:
“Interview staff about failures and consequential risks” is a diagnostic step for the reported condition: Users have no transparent examples of error correction or escalation. This question instead requires an action that can build a representative champions network. Local help and peer examples are not available. Identify enthusiastic credible colleagues across job families.
D:
“Identify meters, user groups, reporting lag and spending owners” is a diagnostic step for the reported condition: Consumption can rise faster than the annual forecast. This question instead requires an action that can build a representative champions network. Local help and peer examples are not available. Identify enthusiastic credible colleagues across job families.
Question 387
At Silverpine Group, an enterprise sponsor is comparing models for regulated internal use. Evidence: Contractual hosting, customization and governance requirements differ. The business goal is to evaluate model sourcing against security, quality and lifecycle criteria. The proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Compare eligible models using the same approved benchmarks and terms
- Require staged design, evaluation, pilot and governed deployment gates
- Publish a documented ranking and choose a feasible limited first trial
- Use a suitable reviewed multimodal workflow with human verification
Explanation
Correct Answer
A:
Contractual hosting, customization and governance requirements differ. Compare eligible models using the same approved benchmarks and terms. A hosted, open-weight or custom model has different obligations for data location, integration, evaluation and upkeep. Compare the options against the actual contract, privacy and lifecycle requirements before buying or building.
Incorrect Answers
B:
“Require staged design, evaluation, pilot and governed deployment gates” is a design or implementation response to the condition: There is no documented validation, deployment or monitoring plan. This question instead requires an action that can evaluate model sourcing against security, quality and lifecycle criteria. Contractual hosting, customization and governance requirements differ. Compare eligible models using the same approved benchmarks and terms. A hosted, open-weight or custom model has different obligations for data location, integration, evaluation and upkeep. Compare the options against the actual contract, privacy and lifecycle requirements before buying or building.
C:
“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 evaluate model sourcing against security, quality and lifecycle criteria. Contractual hosting, customization and governance requirements differ. Compare eligible models using the same approved benchmarks and terms. A hosted, open-weight or custom model has different obligations for data location, integration, evaluation and upkeep. Compare the options against the actual contract, privacy and lifecycle requirements before buying or building.
D:
“Use a suitable reviewed multimodal workflow with human verification” is a design or implementation response to the condition: Text descriptions alone omit visual indicators needed for triage. This question instead requires an action that can evaluate model sourcing against security, quality and lifecycle criteria. Contractual hosting, customization and governance requirements differ. Compare eligible models using the same approved benchmarks and terms. A hosted, open-weight or custom model has different obligations for data location, integration, evaluation and upkeep. Compare the options against the actual contract, privacy and lifecycle requirements before buying or building.
Question 388
At Sapphire Mobility, a legal department searches by exact contract number and semantically similar clauses. Evidence: Vector-only matching risks missing literal identifiers. The business goal is to consider hybrid search combining keyword and semantic retrieval. 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?
- List which facts need public search versus work file access
- Compare claims to quoted source passages and classify unsupported leaps
- Measure exact identifier and conceptual query relevance separately
- Compare current failure patterns against baseline benchmarks
Explanation
Correct Answer
C:
Vector-only matching risks missing literal identifiers. Measure exact identifier and conceptual query relevance separately. Vector similarity can find concepts while missing exact part numbers or codes. Combining lexical and semantic retrieval retains identifier precision and can improve relevant passage recall.
Incorrect Answers
A:
“List which facts need public search versus work file access” is a diagnostic step for the reported condition: Public web claims and workplace records have different provenance. This question instead requires an action that can consider hybrid search combining keyword and semantic retrieval. Vector-only matching risks missing literal identifiers. Measure exact identifier and conceptual query relevance separately. Vector similarity can find concepts while missing exact part numbers or codes. Combining lexical and semantic retrieval retains identifier precision and can improve relevant passage recall.
B:
“Compare claims to quoted source passages and classify unsupported leaps” is a diagnostic step for the reported condition: Citations exist but conclusions are not entailed by the passages. This question instead requires an action that can consider hybrid search combining keyword and semantic retrieval. Vector-only matching risks missing literal identifiers. Measure exact identifier and conceptual query relevance separately. Vector similarity can find concepts while missing exact part numbers or codes. Combining lexical and semantic retrieval retains identifier precision and can improve relevant passage recall.
D:
“Compare current failure patterns against baseline benchmarks” is a diagnostic step for the reported condition: New request types were absent from evaluation data. This question instead requires an action that can consider hybrid search combining keyword and semantic retrieval. Vector-only matching risks missing literal identifiers. Measure exact identifier and conceptual query relevance separately. Vector similarity can find concepts while missing exact part numbers or codes. Combining lexical and semantic retrieval retains identifier precision and can improve relevant passage recall.
Question 389
At Redwood Consulting, a team embeds Foundry API secrets in customer-facing browser code. Evidence: Exposed credentials could be reused for unauthorized inference. The business goal is to protect model endpoints with least privilege and managed credentials. 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 repeated single-source bias and unsupported certainty
- Monitor drift, safety issues and rollback thresholds
- Monitor throttling, retry storms and peak cost
- Monitor leaked tokens and abnormal endpoint traffic
Explanation
Correct Answer
D:
Exposed credentials could be reused for unauthorized inference. Monitor leaked tokens and abnormal endpoint traffic.
Incorrect Answers
A:
“Monitor repeated single-source bias and unsupported certainty” monitors this signal: The agent draws from a small biased sample of industry articles. This question instead requires an action that can protect model endpoints with least privilege and managed credentials. Exposed credentials could be reused for unauthorized inference. Monitor leaked tokens and abnormal endpoint traffic.
B:
“Monitor drift, safety issues and rollback thresholds” monitors this signal: New request types were absent from evaluation data. This question instead requires an action that can protect model endpoints with least privilege and managed credentials. Exposed credentials could be reused for unauthorized inference. Monitor leaked tokens and abnormal endpoint traffic.
C:
“Monitor throttling, retry storms and peak cost” monitors this signal: Quota limits and token throughput may prevent desired traffic. This question instead requires an action that can protect model endpoints with least privilege and managed credentials. Exposed credentials could be reused for unauthorized inference. Monitor leaked tokens and abnormal endpoint traffic.
Question 390
At Ashwood Public Services, a service desk updates one instruction and later cannot explain changed answers. Evidence: Prompt templates have no version owner or change record. The business goal is to control prompt changes like other production artifacts. The proposed solution must respect documented authority and organizational controls. Which approach most directly meets the stated business objective within the approved authority?
- Keep arithmetic controls in validated code and use AI for exception explanation
- Use an approved redacted extract or seek authorized access
- Choose available compliant deployment regions and contractual controls
- Store versioned prompts with owner, regression tests and rollback
Explanation
Correct Answer
D:
Prompt templates have no version owner or change record. Store versioned prompts with owner, regression tests and rollback.
Incorrect Answers
A:
“Keep arithmetic controls in validated code and use AI for exception explanation” is a design or implementation response to the condition: The reconciliation is deterministic arithmetic with known control rules. This question instead requires an action that can control prompt changes like other production artifacts. Prompt templates have no version owner or change record. Store versioned prompts with owner, regression tests and rollback.
B:
“Use an approved redacted extract or seek authorized access” is a design or implementation response to the condition: The user can access the approved summary but not full agreement text. This question instead requires an action that can control prompt changes like other production artifacts. Prompt templates have no version owner or change record. Store versioned prompts with owner, regression tests and rollback.
C:
“Choose available compliant deployment regions and contractual controls” is a design or implementation response to the condition: Data location and transfer obligations may constrain service selection. This question instead requires an action that can control prompt changes like other production artifacts. Prompt templates have no version owner or change record. Store versioned prompts with owner, regression tests and rollback.