Microsoft MD-102 Practice Test 72 - Prepaway
Practice Exams:

Pass Microsoft MD-102 Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers!

30 Days Free Updates, Instant Download!

Verified By Experts
Premium Bundle
$29.98
$79.99 $109.97

MD-102 Premium Bundle


  • Premium File 463 Questions & Answers. Last update: Nov 12, 2026
  • Training Course 84 Lectures
  • Study Guide 512 Pages

Microsoft MD-102 Practice Test 72

Question 710

An Intune custom compliance script reports the string false for an endpoint check, but the rules expect a Boolean. Which TWO shortcuts conflict with the stated management requirements?

  1. Inspect PowerShell JSON serialization, custom rule file, script output, Intune report, and device execution log.
  2. Change the rule to accept any text so failing security agent checks always pass.
  3. A failing pilot produces a Boolean false with an expected noncompliant rule result and correct supporting evidence.
  4. Print formatted colored progress messages instead of compliant JSON and assume Intune infers the desired Boolean.
  5. Emit valid schema-compliant JSON using supported primitive types and align rules evaluation to actual output.

Explanation

Correct Answer

B:

“Change the rule to accept any text so failing security agent checks always pass” would violate the requirement. Custom compliance evaluates typed script output, so ambiguous strings cannot substitute for declared Boolean conditions.

D:

“Print formatted colored progress messages instead of compliant JSON and assume Intune infers the desired Boolean” would violate the requirement. Custom compliance evaluates typed script output, so ambiguous strings cannot substitute for declared Boolean conditions.

Incorrect Answers

A:

“Inspect PowerShell JSON serialization, custom rule file, script output, Intune report, and device execution log” is the relevant troubleshooting evidence, not a reason to bypass safeguards. Custom compliance evaluates typed script output, so ambiguous strings cannot substitute for declared Boolean conditions.

C:

A failing pilot produces a Boolean false with an expected noncompliant rule result and correct supporting evidence. This is a bounded acceptance test of the intended outcome, not an unsafe production shortcut.

E:

“Emit valid schema-compliant JSON using supported primitive types and align rules evaluation to actual output” is the controlled implementation, not an unsafe workaround. Custom compliance evaluates typed script output, so ambiguous strings cannot substitute for declared Boolean conditions. Confirm that a failing pilot produces a Boolean false with an expected noncompliant rule result and correct supporting evidence.

Question 711

A Security Copilot agent recommends isolating a device after detecting suspicious credential theft behavior. What is the most appropriate supported endpoint management change?

  1. Ignore source telemetry and treat an AI-written summary as proof of a confirmed security breach.
  2. Review the agent's evidence and recommended scope before performing an authorized Defender or Intune containment action.
  3. Use a supported unattended authentication method with approved permissions and token refresh handling.
  4. Approve every suggested destructive device action blindly because an AI assistant supplied a confidence score.

Explanation

Correct Answer

B:

A Security Copilot agent recommends isolating a device after detecting suspicious credential theft behavior. Review the agent's evidence and recommended scope before performing an authorized Defender or Intune containment action. Agent-generated security recommendations support analyst judgment but do not automatically establish authority for destructive response. A specialist can validate the threat context and approve a targeted containment with a traceable audit record.

Incorrect Answers

A:

“Ignore source telemetry and treat an AI-written summary as proof of a confirmed security breach” would not meet the requirement. Instead, review the agent's evidence and recommended scope before performing an authorized Defender or Intune containment action.

C:

It would not produce the required result: A specialist can validate the threat context and approve a targeted containment with a traceable audit record.

D:

“Approve every suggested destructive device action blindly because an AI assistant supplied a confidence score” does not implement the control that the facts call for. Instead, review the agent's evidence and recommended scope before performing an authorized Defender or Intune containment action.

Question 712

A Security Copilot agent recommends isolating a device after detecting suspicious credential theft behavior. Which evidence most directly isolates the endpoint issue?

  1. Inspect token expiry, authentication method, app permissions, scheduled job context, and 401 response logs.
  2. Inspect agent evidence links, Defender timeline, device owner, recommended action, and approval decisions.
  3. Inspect reliability metrics, planned update schedule, system event IDs, crash dumps, and device grouping.
  4. Inspect report filter, RBAC role, export destination permissions, selected columns, and retention logs.

Explanation

Correct Answer

B:

A Security Copilot agent recommends isolating a device after detecting suspicious credential theft behavior. Inspect agent evidence links, Defender timeline, device owner, recommended action, and approval decisions. Agent-generated security recommendations support analyst judgment but do not automatically establish authority for destructive response. A specialist can validate the threat context and approve a targeted containment with a traceable audit record.

Incorrect Answers

A:

This evidence primarily tests powershell graph authentication token renewal; it would not isolate the actual failure boundary in this scenario. The targeted checks are: Inspect agent evidence links, Defender timeline, device owner, recommended action, and approval decisions.

C:

This evidence primarily tests endpoint restart frequency baseline and trend; it would not isolate the actual failure boundary in this scenario. The targeted checks are: Inspect agent evidence links, Defender timeline, device owner, recommended action, and approval decisions.

D:

This evidence primarily tests reporting export permissions and sensitive endpoint data; it would not isolate the actual failure boundary in this scenario. The targeted checks are: Inspect agent evidence links, Defender timeline, device owner, recommended action, and approval decisions.

Question 713

A Security Copilot agent recommends isolating a device after detecting suspicious credential theft behavior. What measured result confirms the endpoint deployment now works?

  1. A regional user sees accurate scoped compliance counts matching their permitted managed devices.
  2. A specialist can validate the threat context and approve a targeted containment with a traceable audit record.
  3. A test deployment selects only eligible corporate devices and records membership changes for audit.
  4. Service health reporting aligns with the incident and devices recover management check-in after provider restoration.

Explanation

Correct Answer

B:

A specialist can validate the threat context and approve a targeted containment with a traceable audit record. Review the agent's evidence and recommended scope before performing an authorized Defender or Intune containment action.

Incorrect Answers

A:

This could validate intune workbook operational dashboard filters, but it does not prove the specific requirement has been met. Instead measure: A specialist can validate the threat context and approve a targeted containment with a traceable audit record.

C:

This could validate automated group membership with entra attribute, but it does not prove the specific requirement has been met. Instead measure: A specialist can validate the threat context and approve a targeted containment with a traceable audit record.

D:

This could validate tenant service health versus device issue, but it does not prove the specific requirement has been met. Instead measure: A specialist can validate the threat context and approve a targeted containment with a traceable audit record.

Question 714

A Security Copilot agent recommends isolating a device after detecting suspicious credential theft behavior. Which underlying Intune or Windows behavior is most relevant here?

  1. Startup performance scoring distinguishes boot and sign-in experience and benefits from cohort and change comparison.
  2. AI-generated management suggestions are not an authorization boundary for identity or compliance policy changes.
  3. Agent-generated security recommendations support analyst judgment but do not automatically establish authority for destructive response.
  4. A compliance dashboard alone does not guarantee timely escalation without alerting and defined operations responsibility.

Explanation

Correct Answer

C:

Inspect agent evidence links, Defender timeline, device owner, recommended action, and approval decisions. A specialist can validate the threat context and approve a targeted containment with a traceable audit record.

Incorrect Answers

A:

This concerns endpoint analytics startup score investigation, which is a different mechanism than the one causing the stated symptom. The principle that applies here is: Agent-generated security recommendations support analyst judgment but do not automatically establish authority for destructive response.

B:

This concerns security copilot remediation approval workflow, which is a different mechanism than the one causing the stated symptom. The principle that applies here is: Agent-generated security recommendations support analyst judgment but do not automatically establish authority for destructive response.

D:

This concerns compliance drift alert rule thresholds, which is a different mechanism than the one causing the stated symptom. The principle that applies here is: Agent-generated security recommendations support analyst judgment but do not automatically establish authority for destructive response.

Question 715

A Security Copilot agent recommends isolating a device after detecting suspicious credential theft behavior. Which TWO unsafe endpoint management actions should be rejected?

  1. Ignore source telemetry and treat an AI-written summary as proof of a confirmed security breach.
  2. A specialist can validate the threat context and approve a targeted containment with a traceable audit record.
  3. Approve every suggested destructive device action blindly because an AI assistant supplied a confidence score.
  4. Review the agent's evidence and recommended scope before performing an authorized Defender or Intune containment action.
  5. Inspect agent evidence links, Defender timeline, device owner, recommended action, and approval decisions.

Explanation

Correct Answer

A:

“Ignore source telemetry and treat an AI-written summary as proof of a confirmed security breach” would violate the requirement. The appropriate control instead is to review the agent's evidence and recommended scope before performing an authorized Defender or Intune containment action.

C:

“Approve every suggested destructive device action blindly because an AI assistant supplied a confidence score” would violate the requirement.

Incorrect Answers

B:

A specialist can validate the threat context and approve a targeted containment with a traceable audit record. This is a bounded acceptance test of the intended outcome, not an unsafe production shortcut.

D:

“Review the agent's evidence and recommended scope before performing an authorized Defender or Intune containment action” is the controlled implementation, not an unsafe workaround.

E:

“Inspect agent evidence links, Defender timeline, device owner, recommended action, and approval decisions” is the relevant troubleshooting evidence, not a reason to bypass safeguards.

Question 716

An endpoint analytics agent recommends broad memory upgrades after identifying slow app startups on a subset of PCs. Which Intune or Entra configuration best solves the endpoint requirement?

  1. Validate the suggested cohort and telemetry before choosing a staged configuration, application, or hardware improvement.
  2. Disable performance measurements after accepting the AI recommendation so no contradictory evidence appears.
  3. Analyze restart event categories and historical baselines before deploying a device fix to affected cohorts.
  4. Purchase new memory for every corporate laptop based only on one anomalous example summarized by the agent.

Explanation

Correct Answer

A:

An endpoint analytics agent recommends broad memory upgrades after identifying slow app startups on a subset of PCs. Validate the suggested cohort and telemetry before choosing a staged configuration, application, or hardware improvement. AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes. A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.

Incorrect Answers

B:

“Disable performance measurements after accepting the AI recommendation so no contradictory evidence appears” would not meet the requirement. AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes. Instead, validate the suggested cohort and telemetry before choosing a staged configuration, application, or hardware improvement.

C:

It would not produce the required result: A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.

D:

“Purchase new memory for every corporate laptop based only on one anomalous example summarized by the agent” does not implement the control that the facts call for. AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes. Instead, validate the suggested cohort and telemetry before choosing a staged configuration, application, or hardware improvement.

Question 717

An endpoint analytics agent recommends broad memory upgrades after identifying slow app startups on a subset of PCs. What should the administrator inspect to diagnose this behavior?

  1. Inspect endpoint delta support, delta token, response metadata, 429 retry-after headers, and sync timestamps.
  2. Inspect agent reasoning, endpoint startup metrics, device model, background app behavior, and test comparisons.
  3. Inspect report filter, RBAC role, export destination permissions, selected columns, and retention logs.
  4. Inspect script exit codes, detection output, remediation status, assignment, and local management extension logs.

Explanation

Correct Answer

B:

An endpoint analytics agent recommends broad memory upgrades after identifying slow app startups on a subset of PCs. Inspect agent reasoning, endpoint startup metrics, device model, background app behavior, and test comparisons. AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes. A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.

Incorrect Answers

A:

This evidence primarily tests graph delta synchronization for device inventory; it would not isolate the actual failure boundary in this scenario. The targeted checks are: Inspect agent reasoning, endpoint startup metrics, device model, background app behavior, and test comparisons.

C:

This evidence primarily tests reporting export permissions and sensitive endpoint data; it would not isolate the actual failure boundary in this scenario. The targeted checks are: Inspect agent reasoning, endpoint startup metrics, device model, background app behavior, and test comparisons.

D:

This evidence primarily tests remediations detection and correction scripts; it would not isolate the actual failure boundary in this scenario. The targeted checks are: Inspect agent reasoning, endpoint startup metrics, device model, background app behavior, and test comparisons.

Question 718

An endpoint analytics agent recommends broad memory upgrades after identifying slow app startups on a subset of PCs. Which acceptance criterion validates the chosen endpoint management solution?

  1. A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.
  2. The resulting report identifies crash-driven reboots and a targeted fix reduces unplanned restarts in the pilot group.
  3. A test recommendation remains advisory until an authorized reviewer approves a bounded policy change.
  4. A simulated enrollment failure cohort produces a timely incident notice and points to the relevant error category.

Explanation

Correct Answer

A:

AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes. A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.

Incorrect Answers

B:

This could validate endpoint restart frequency baseline and trend, but it does not prove the specific requirement has been met. Instead measure: A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.

C:

This could validate security copilot remediation approval workflow, but it does not prove the specific requirement has been met. Instead measure: A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.

D:

This could validate enrollment failures operational alerting, but it does not prove the specific requirement has been met. Instead measure: A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.

Question 719

An endpoint analytics agent recommends broad memory upgrades after identifying slow app startups on a subset of PCs. Which Microsoft endpoint management principle explains the decision?

  1. Graph application permissions govern what service principals may access, independently of script comments or friendly descriptions.
  2. AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes.
  3. Automated endpoint recommendations are hypotheses needing validation, especially when security and connectivity agents are involved.
  4. A compliance dashboard alone does not guarantee timely escalation without alerting and defined operations responsibility.

Explanation

Correct Answer

B:

AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes. Validate the suggested cohort and telemetry before choosing a staged configuration, application, or hardware improvement. The decisive evidence is that a pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.

Incorrect Answers

A:

This concerns graph least-privilege app authorization for reporting, which is a different mechanism than the one causing the stated symptom. The principle that applies here is: AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes.

C:

This concerns endpoint analytics recommended action validation, which is a different mechanism than the one causing the stated symptom. The principle that applies here is: AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes.

D:

This concerns compliance drift alert rule thresholds, which is a different mechanism than the one causing the stated symptom. The principle that applies here is: AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes.

Question 720

An endpoint analytics agent recommends broad memory upgrades after identifying slow app startups on a subset of PCs. Select TWO choices that would create avoidable endpoint security or reliability risk.

  1. A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.
  2. Purchase new memory for every corporate laptop based only on one anomalous example summarized by the agent.
  3. Disable performance measurements after accepting the AI recommendation so no contradictory evidence appears.
  4. Validate the suggested cohort and telemetry before choosing a staged configuration, application, or hardware improvement.
  5. Inspect agent reasoning, endpoint startup metrics, device model, background app behavior, and test comparisons.

Explanation

Correct Answer

B:

“Purchase new memory for every corporate laptop based only on one anomalous example summarized by the agent” would violate the requirement. AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes.

C:

“Disable performance measurements after accepting the AI recommendation so no contradictory evidence appears” would violate the requirement. AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes.

Incorrect Answers

A:

A pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere. This is a bounded acceptance test of the intended outcome, not an unsafe production shortcut.

D:

“Validate the suggested cohort and telemetry before choosing a staged configuration, application, or hardware improvement” is the controlled implementation, not an unsafe workaround. AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes.

E:

“Inspect agent reasoning, endpoint startup metrics, device model, background app behavior, and test comparisons” is the relevant troubleshooting evidence, not a reason to bypass safeguards. AI-assisted performance analysis can highlight patterns but requires representative evidence and human decisions about changes. Confirm that a pilot cohort shows measurable performance improvement without regressions or unnecessary upgrades elsewhere.

Endpoint Administrator (MD-102) Practice Tests