Microsoft AB-100 Dumps
Microsoft AB-100 Practice Test 69
Question 680
During a solution integration review, a security architect must accept a change for multi-agent delegated authority. The issue to be addressed is: A planning agent delegates tasks to an executor agent with more privileges than the original user. Which concrete verification action demonstrates that the relevant behavior works?
Explanation
Correct Answer
To verify this change, the pharmaceutical supplier should delegate a task asking for admin-only records and confirm end-to-end denial. Operators still need the implementation to propagate caller identity purpose and bounded delegation into cross-agent tool policy and the preventive requirement to enforce least-privilege delegation and audit the full chain of authority.
Incorrect Answers
The action to propagate caller identity purpose and bounded delegation into cross-agent tool policy can contribute to multi-agent delegated authority as a implementation step; here, a planning agent delegates tasks to an executor agent with more privileges than the original user. The more directly responsive choice is to delegate a task asking for admin-only records and confirm end-to-end denial.
The assumption to treat delegation as unconditional permission escalation for downstream agents is the exact failure pattern the team needs to avoid. The relevant change is to propagate caller identity purpose and bounded delegation into cross-agent tool policy; a credible check must delegate a task asking for admin-only records and confirm end-to-end denial.
The action to enforce least-privilege delegation and audit the full chain of authority can contribute to multi-agent delegated authority as a ongoing preventive rule step; here, a planning agent delegates tasks to an executor agent with more privileges than the original user. The more directly responsive choice is to delegate a task asking for admin-only records and confirm end-to-end denial.
The option to attempt an unapproved production publish and confirm tenant controls block it would normally serve a different governance of agent publication requirement, where department makers can publish experimental agents directly to the entire tenant without security review. The alternative may be useful in its own context, but it does not directly supply delegate a task asking for admin-only records and confirm end-to-end denial for the stated decision.
The action to revoke overbroad delegated credentials and reissue a scoped task contract can contribute to multi-agent delegated authority as a corrective remediation step; here, a planning agent delegates tasks to an executor agent with more privileges than the original user. The more directly responsive choice is to delegate a task asking for admin-only records and confirm end-to-end denial.
Question 681
At a national charity, the business application owner is preparing a model evaluation review. The concrete requirement is: An AI assistant accidentally includes confidential training examples in an answer. Which design action BEST addresses the stated need?
Explanation
Correct Answer
The proposed implementation is to apply privacy controls output screening and data minimization for model knowledge and retrieval; this matches the operational requirement that an AI assistant accidentally includes confidential training examples in an answer. The team can probe known protected phrases and verify they are not disclosed to unauthorized users. The team should also test regulated output classes and monitor leakage incidents without storing extra secrets.
Incorrect Answers
The action to probe known protected phrases and verify they are not disclosed to unauthorized users can contribute to model output data leakage as a targeted acceptance check step; here, an AI assistant accidentally includes confidential training examples in an answer. The more directly responsive choice is to apply privacy controls output screening and data minimization for model knowledge and retrieval.
The assumption to assume response moderation alone prevents all memorized sensitive-data disclosure is the exact failure pattern the team needs to avoid. The relevant change is to apply privacy controls output screening and data minimization for model knowledge and retrieval; a credible check must probe known protected phrases and verify they are not disclosed to unauthorized users.
Choosing to manage distinct workload identities with expiry review and revocation when agents retire would direct effort toward agent identity lifecycle, whose characteristic issue is that an unattended service principal continues using broad permissions after an agent project is discontinued. The relevant corrective or preventive action here remains to apply privacy controls output screening and data minimization for model knowledge and retrieval, with its own validation criteria.
Question 682
An AI governance lead at a housing association is conducting an agent rollout concerning model output data leakage. The observed problem is: An AI assistant accidentally includes confidential training examples in an answer. Which TWO actions together represent the needed implementation AND a durable preventive safeguard? (Select TWO answers.)
Explanation
Correct Answers
Standardizing the boundary to test regulated output classes and monitor leakage incidents without storing extra secrets is preventive because the same model output data leakage failure could otherwise return: an AI assistant accidentally includes confidential training examples in an answer. For this housing association, the service should also apply privacy controls output screening and data minimization for model knowledge and retrieval, and tests should probe known protected phrases and verify they are not disclosed to unauthorized users so that the ongoing requirement is measurable.
Choosing to apply privacy controls output screening and data minimization for model knowledge and retrieval directly targets the described model output data leakage gap: an AI assistant accidentally includes confidential training examples in an answer. As a lasting boundary, they also need to test regulated output classes and monitor leakage incidents without storing extra secrets, because an implementation by itself does not prove safe operation.
Incorrect Answers
The alternative to probe known protected phrases and verify they are not disclosed to unauthorized users is a targeted acceptance check step rather than the faulty assumption being diagnosed. The failure in model output data leakage arises from assume response moderation alone prevents all memorized sensitive-data disclosure; the concrete symptom is that an AI assistant accidentally includes confidential training examples in an answer.
This approach relates to retention minimization for agent logs: aI tool traces retain full chat histories indefinitely even though incidents need only minimal evidence. If the team chose it instead, the targeted check to probe known protected phrases and verify they are not disclosed to unauthorized users could still fail because the underlying required action is to apply privacy controls output screening and data minimization for model knowledge and retrieval.
The alternative to remove risky sources and investigate leakage with privacy-preserving evaluation is a corrective remediation step rather than the faulty assumption being diagnosed. The failure in model output data leakage arises from assume response moderation alone prevents all memorized sensitive-data disclosure; the concrete symptom is that an AI assistant accidentally includes confidential training examples in an answer.
Question 683
The retail franchise is responding to a governance board meeting. A privacy engineer discovers this issue: An AI assistant accidentally includes confidential training examples in an answer. What corrective step should the team take to restore the intended functionality most directly?
Explanation
Correct Answer
The direct repair is to remove risky sources and investigate leakage with privacy-preserving evaluation. Once the condition is corrected, the retail franchise should probe known protected phrases and verify they are not disclosed to unauthorized users. To reduce recurrence, it should also test regulated output classes and monitor leakage incidents without storing extra secrets; those steps provide different recovery and prevention functions.
Incorrect Answers
The action to probe known protected phrases and verify they are not disclosed to unauthorized users can contribute to model output data leakage as a targeted acceptance check step; here, an AI assistant accidentally includes confidential training examples in an answer. The more directly responsive choice is to remove risky sources and investigate leakage with privacy-preserving evaluation.
The assumption to assume response moderation alone prevents all memorized sensitive-data disclosure is the exact failure pattern the team needs to avoid. The relevant change is to apply privacy controls output screening and data minimization for model knowledge and retrieval; a credible check must probe known protected phrases and verify they are not disclosed to unauthorized users.
This approach relates to grounding-data row-level access: a user asks an AI assistant for commission data belonging to another sales representative. If the team chose it instead, the targeted check to probe known protected phrases and verify they are not disclosed to unauthorized users could still fail because the underlying required action is to remove risky sources and investigate leakage with privacy-preserving evaluation.
Question 684
A contact center architect at a research institute wants the following issue to be prevented in later deployments: An AI assistant accidentally includes confidential training examples in an answer. Which enduring control or engineering guardrail BEST prevents recurrence?
Explanation
Correct Answer
Standardizing the boundary to test regulated output classes and monitor leakage incidents without storing extra secrets is preventive because the same model output data leakage failure could otherwise return: an AI assistant accidentally includes confidential training examples in an answer. For this research institute, the service should also apply privacy controls output screening and data minimization for model knowledge and retrieval, and tests should probe known protected phrases and verify they are not disclosed to unauthorized users so that the ongoing requirement is measurable.
Incorrect Answers
The action to require traceability and contestability for consequential AI-supported decisions addresses ai accountability and explanation rather than the decisive model output data leakage condition. The research institute should not infer that fixing one adjacent business-system dependency validates its unrelated workflow; the requested action is to test regulated output classes and monitor leakage incidents without storing extra secrets.
The assumption to assume response moderation alone prevents all memorized sensitive-data disclosure is the exact failure pattern the team needs to avoid. The relevant change is to apply privacy controls output screening and data minimization for model knowledge and retrieval; a credible check must probe known protected phrases and verify they are not disclosed to unauthorized users.
The action to remove risky sources and investigate leakage with privacy-preserving evaluation can contribute to model output data leakage as a corrective remediation step; here, an AI assistant accidentally includes confidential training examples in an answer. The more directly responsive choice is to test regulated output classes and monitor leakage incidents without storing extra secrets.
Question 685
During an agent rollout, a product delivery lead must accept a change for model output data leakage. The issue to be addressed is: An AI assistant accidentally includes confidential training examples in an answer. Which concrete verification action demonstrates that the relevant behavior works?
Explanation
Correct Answer
The targeted acceptance test is to probe known protected phrases and verify they are not disclosed to unauthorized users. The design should first apply privacy controls output screening and data minimization for model knowledge and retrieval. A robust hospitality operator deployment also needs to test regulated output classes and monitor leakage incidents without storing extra secrets; passing this check provides relevant evidence but does not certify every possible input or security risk.
Incorrect Answers
The action to apply privacy controls output screening and data minimization for model knowledge and retrieval can contribute to model output data leakage as a implementation step; here, an AI assistant accidentally includes confidential training examples in an answer. The more directly responsive choice is to probe known protected phrases and verify they are not disclosed to unauthorized users.
The assumption to assume response moderation alone prevents all memorized sensitive-data disclosure is the exact failure pattern the team needs to avoid. The relevant change is to apply privacy controls output screening and data minimization for model knowledge and retrieval; a credible check must probe known protected phrases and verify they are not disclosed to unauthorized users.
For prompt manipulation defense, it can be sensible to inject a forged approval in an invoice and confirm payment authority is not granted if an attacker puts forged approval instructions inside a retrieved supplier invoice. Without the actual change to probe known protected phrases and verify they are not disclosed to unauthorized users, the answer lacks the correct operational scope even if the alternate action is technically valid elsewhere.
Question 686
At a pharmaceutical supplier, the Copilot Studio architect is preparing an AI adoption planning. The concrete requirement is: A business owner accepts residual risk for an agent without specifying duration or compensating measures. Which design action BEST addresses the stated need?
Explanation
Correct Answer
Choosing to document authorized time-bounded exceptions with explicit owner and mitigation evidence directly targets the described agent risk exception governance gap: a business owner accepts residual risk for an agent without specifying duration or compensating measures. As a lasting boundary, they also need to track exception expiry conditions and accountable approval roles, because an implementation by itself does not prove safe operation.
Incorrect Answers
Choosing to restrict outbound destinations payload fields and authorization of network tools would direct effort toward data exfiltration through tools, whose characteristic issue is that a customer-support agent can send unrestricted HTTP requests containing private case content. The relevant corrective or preventive action here remains to document authorized time-bounded exceptions with explicit owner and mitigation evidence, with its own validation criteria.
The assumption to declare every AI risk permanently accepted after an informal project meeting is the exact failure pattern the team needs to avoid. The relevant change is to document authorized time-bounded exceptions with explicit owner and mitigation evidence; a credible check must inspect an expired exception and confirm deployment policy requires review.
The action to inspect an expired exception and confirm deployment policy requires review can contribute to agent risk exception governance as a targeted acceptance check step; here, a business owner accepts residual risk for an agent without specifying duration or compensating measures. The more directly responsive choice is to document authorized time-bounded exceptions with explicit owner and mitigation evidence.
Question 687
After a data access design at a regional hospital group, the team observes this failure: A business owner accepts residual risk for an agent without specifying duration or compensating measures. Which flawed assumption or implementation choice would MOST DIRECTLY explain why the intended outcome was not achieved?
Explanation
Correct Answer
The faulty assumption is to declare every AI risk permanently accepted after an informal project meeting. A functioning alternative would be to document authorized time-bounded exceptions with explicit owner and mitigation evidence, and the relevant validation would be to inspect an expired exception and confirm deployment policy requires review.
Incorrect Answers
For the root-cause diagnosis decision on agent risk exception governance with the option “Share one perpetual tenant administrator identity across all departmental agents”, the option to share one perpetual tenant administrator identity across all departmental agents would normally serve a different agent identity lifecycle requirement, where an unattended service principal continues using broad permissions after an agent project is discontinued. The alternative may be useful in its own context, but it does not directly supply declare every AI risk permanently accepted after an informal project meeting for the stated decision.
The alternative to document authorized time-bounded exceptions with explicit owner and mitigation evidence is a implementation step rather than the faulty assumption being diagnosed. The failure in agent risk exception governance arises from declare every AI risk permanently accepted after an informal project meeting; the concrete symptom is that a business owner accepts residual risk for an agent without specifying duration or compensating measures.
The alternative to inspect an expired exception and confirm deployment policy requires review is a targeted acceptance check step rather than the faulty assumption being diagnosed. The failure in agent risk exception governance arises from declare every AI risk permanently accepted after an informal project meeting; the concrete symptom is that a business owner accepts residual risk for an agent without specifying duration or compensating measures.
Question 688
The insurance provider is responding to a customer case workflow redesign. An AI program manager discovers this issue: A business owner accepts residual risk for an agent without specifying duration or compensating measures. What corrective step should the team take to restore the intended functionality most directly?
Explanation
Correct Answer
The appropriate remediation is to reopen undocumented risks and obtain a scoped signed decision, because the failure concerns agent risk exception governance: a business owner accepts residual risk for an agent without specifying duration or compensating measures. A meaningful verification should inspect an expired exception and confirm deployment policy requires review, and the durable operating rule should track exception expiry conditions and accountable approval roles after the immediate fix.
Incorrect Answers
The action to inspect an expired exception and confirm deployment policy requires review can contribute to agent risk exception governance as a targeted acceptance check step; here, a business owner accepts residual risk for an agent without specifying duration or compensating measures. The more directly responsive choice is to reopen undocumented risks and obtain a scoped signed decision.
This approach relates to responsible ai review of accessibility: an AI-powered customer portal works for typed English input but fails users relying on accessibility features. If the team chose it instead, the targeted check to inspect an expired exception and confirm deployment policy requires review could still fail because the underlying required action is to reopen undocumented risks and obtain a scoped signed decision.
The assumption to declare every AI risk permanently accepted after an informal project meeting is the exact failure pattern the team needs to avoid. The relevant change is to document authorized time-bounded exceptions with explicit owner and mitigation evidence; a credible check must inspect an expired exception and confirm deployment policy requires review.
Question 689
A business systems designer at a software publisher wants the following issue to be prevented in later deployments: A business owner accepts residual risk for an agent without specifying duration or compensating measures. Which enduring control or engineering guardrail BEST prevents recurrence?
Explanation
Correct Answer
The sustained prevention measure is to track exception expiry conditions and accountable approval roles. The implementation can still use the separate action to document authorized time-bounded exceptions with explicit owner and mitigation evidence, but a guardrail constrains future changes. The software publisher can measure whether the rule works by arranging to inspect an expired exception and confirm deployment policy requires review.
Incorrect Answers
The action to require use-case risk assessment and periodic responsible AI evaluation addresses responsible ai harm assessment rather than the decisive agent risk exception governance condition. The software publisher should not infer that fixing one adjacent business-system dependency validates its unrelated workflow; the requested action is to track exception expiry conditions and accountable approval roles.
The action to reopen undocumented risks and obtain a scoped signed decision can contribute to agent risk exception governance as a corrective remediation step; here, a business owner accepts residual risk for an agent without specifying duration or compensating measures. The more directly responsive choice is to track exception expiry conditions and accountable approval roles.
The assumption to declare every AI risk permanently accepted after an informal project meeting is the exact failure pattern the team needs to avoid. The relevant change is to document authorized time-bounded exceptions with explicit owner and mitigation evidence; a credible check must inspect an expired exception and confirm deployment policy requires review.
Question 690
During a data access design, an integration engineer must accept a change for agent risk exception governance. The issue to be addressed is: A business owner accepts residual risk for an agent without specifying duration or compensating measures. Which concrete verification action demonstrates that the relevant behavior works?
Explanation
Correct Answer
To verify this change, the government services agency should inspect an expired exception and confirm deployment policy requires review. Operators still need the implementation to document authorized time-bounded exceptions with explicit owner and mitigation evidence and the preventive requirement to track exception expiry conditions and accountable approval roles.
Incorrect Answers
The action to document authorized time-bounded exceptions with explicit owner and mitigation evidence can contribute to agent risk exception governance as a implementation step; here, a business owner accepts residual risk for an agent without specifying duration or compensating measures. The more directly responsive choice is to inspect an expired exception and confirm deployment policy requires review.
The assumption to declare every AI risk permanently accepted after an informal project meeting is the exact failure pattern the team needs to avoid. The relevant change is to document authorized time-bounded exceptions with explicit owner and mitigation evidence; a credible check must inspect an expired exception and confirm deployment policy requires review.
For model output data leakage, it can be sensible to probe known protected phrases and verify they are not disclosed to unauthorized users if an AI assistant accidentally includes confidential training examples in an answer. Without the actual change to inspect an expired exception and confirm deployment policy requires review, the answer lacks the correct operational scope even if the alternate action is technically valid elsewhere.
Agentic AI Business Solutions Architect (AB-100) Practice Tests
- Microsoft AB-100 Practice Test 67 - 10 Questions (Last Reviewed 12-Nov-2026)
- Microsoft AB-100 Practice Test 68 - 10 Questions (Last Reviewed 11-Nov-2026)
- Microsoft AB-100 Practice Test 69 - 10 Questions (Last Reviewed 10-Nov-2026)
- Microsoft AB-100 Practice Test 70 - 10 Questions (Last Reviewed 09-Nov-2026)