Microsoft AB-100 Exam Dumps, Practice Test Questions

100% Latest & Updated Microsoft AB-100 Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!

Microsoft AB-100  Premium File
$54.99
$49.99

AB-100 Premium File

  • Premium File: 129 Questions & Answers. Last update: Nov 21, 2026
  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates

AB-100 Premium File

Microsoft AB-100  Premium File
  • Premium File: 129 Questions & Answers. Last update: Nov 21, 2026
  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates
$54.99
$49.99

Microsoft AB-100 Practice Test 40

Question 390

A healthcare agent could expose private data via citations that link to accessible public caches. What should the solution architect recommend for this specific need?

  1. Evaluate retrieval poisoning risk and apply source trust/approval controls.
  2. Test inclusive interaction and provide accessible alternative channels.
  3. Audit citation targets and source access controls for unintended disclosure.
  4. Use authoritative retrieval, citation verification and escalation for unsupported legal claims.

Explanation

Correct Answer

C:

Even accurate citations can leak restricted locations or contents.

Incorrect Answers

A:

External content can be altered to influence agent answers and tool choices, whereas even accurate citations can leak restricted locations or contents.

B:

Accessible design is a separate responsible-AI criterion from factual correctness. Even accurate citations can leak restricted locations or contents.

D:

Evidence constraints and review address harmful fabricated citations. Even accurate citations can leak restricted locations or contents.

Question 391

A production agent is grounded on external documents whose authors can change text without notice. What is the most defensible intervention for this case?

  1. Review model provenance, license and integrity before approving deployment.
  2. Use authoritative retrieval, citation verification and escalation for unsupported legal claims.
  3. Threat-model the tool-action path and test prompt injection with server-side authorization.
  4. Evaluate retrieval poisoning risk and apply source trust/approval controls.

Explanation

Correct Answer

D:

External content can be altered to influence agent answers and tool choices.

Incorrect Answers

A:

Third-party model supply-chain checks reduce security and legal uncertainty. External content can be altered to influence agent answers and tool choices.

B:

Evidence constraints and review address harmful fabricated citations, but external content can be altered to influence agent answers and tool choices.

C:

A prompt defense alone is insufficient without transactional access enforcement. External content can be altered to influence agent answers and tool choices.

Question 392

A model occasionally fabricates exact legal references and the organization needs a mitigation plan. Which architecture response most directly addresses the risk?

  1. Use authoritative retrieval, citation verification and escalation for unsupported legal claims.
  2. Threat-model the tool-action path and test prompt injection with server-side authorization.
  3. Review model provenance, license and integrity before approving deployment.
  4. Audit citation targets and source access controls for unintended disclosure.

Explanation

Correct Answer

A:

Evidence constraints and review address harmful fabricated citations.

Incorrect Answers

B:

A prompt defense alone is insufficient without transactional access enforcement. Evidence constraints and review address harmful fabricated citations.

C:

Third-party model supply-chain checks reduce security and legal uncertainty; evidence constraints and review address harmful fabricated citations.

D:

Even accurate citations can leak restricted locations or contents, whereas evidence constraints and review address harmful fabricated citations.

Question 393

A hiring-screening assistant produces very different false-negative rates across candidate groups. Which solution approach best meets the stated requirement?

  1. Revoke the token immediately, contain access and audit recent transactions.
  2. Test inclusive interaction and provide accessible alternative channels.
  3. Evaluate fairness disparities and require accountable review of adverse outcomes.
  4. Clearly disclose AI involvement and identify review or appeal channels.

Explanation

Correct Answer

C:

Group-level error analysis can reveal differential impact hidden by aggregate accuracy.

Incorrect Answers

A:

Credential incident containment requires prompt revocation and impact reconstruction. Group-level error analysis can reveal differential impact hidden by aggregate accuracy.

B:

Accessible design is a separate responsible-AI criterion from factual correctness; group-level error analysis can reveal differential impact hidden by aggregate accuracy.

D:

Transparency supports informed understanding of generated recommendations and decisions. Group-level error analysis can reveal differential impact hidden by aggregate accuracy.

Question 394

Customers mistake automatically drafted decisions for human determinations. Which option most directly closes the documented gap?

  1. Design intelligible refusal explanations and an appropriate human escalation route.
  2. Test inclusive interaction and provide accessible alternative channels.
  3. Require named business owner, technical owner and risk accountability for each agent.
  4. Clearly disclose AI involvement and identify review or appeal channels.

Explanation

Correct Answer

D:

Transparency supports informed understanding of generated recommendations and decisions.

Incorrect Answers

A:

Contestability and clarity improve user agency without removing safety boundaries; transparency supports informed understanding of generated recommendations and decisions.

B:

Accessible design is a separate responsible-AI criterion from factual correctness, whereas transparency supports informed understanding of generated recommendations and decisions.

C:

Ownership aligns agents with responsible control and operating decisions. Transparency supports informed understanding of generated recommendations and decisions.

Question 395

A public services chatbot fails on accessibility tools even though its response accuracy is high. Which approach best satisfies the actual business constraint?

  1. Test inclusive interaction and provide accessible alternative channels.
  2. Design intelligible refusal explanations and an appropriate human escalation route.
  3. Evaluate fairness disparities and require accountable review of adverse outcomes.
  4. Use a narrowly scoped service identity appropriate for the low-risk lookup.

Explanation

Correct Answer

A:

Accessible design is a separate responsible-AI criterion from factual correctness.

Incorrect Answers

B:

Contestability and clarity improve user agency without removing safety boundaries. Accessible design is a separate responsible-AI criterion from factual correctness.

C:

Group-level error analysis can reveal differential impact hidden by aggregate accuracy, but accessible design is a separate responsible-AI criterion from factual correctness.

D:

A bounded service credential can fit a shared public lookup when it grants no sensitive rights. Accessible design is a separate responsible-AI criterion from factual correctness.

Question 396

A model refuses legitimate low-risk service questions and users cannot understand or contest the behavior. Which specific remediation should the business adopt?

  1. Evaluate fairness disparities and require accountable review of adverse outcomes.
  2. Design intelligible refusal explanations and an appropriate human escalation route.
  3. Evaluate retrieval poisoning risk and apply source trust/approval controls.
  4. Clearly disclose AI involvement and identify review or appeal channels.

Explanation

Correct Answer

B:

Contestability and clarity improve user agency without removing safety boundaries.

Incorrect Answers

A:

Group-level error analysis can reveal differential impact hidden by aggregate accuracy. Contestability and clarity improve user agency without removing safety boundaries.

C:

External content can be altered to influence agent answers and tool choices. Contestability and clarity improve user agency without removing safety boundaries.

D:

Transparency supports informed understanding of generated recommendations and decisions; contestability and clarity improve user agency without removing safety boundaries.

Question 397

A leaked agent tool token could still perform write operations in a production CRM. Which controlled change is most appropriate for this situation?

  1. Disable the affected capability and verify permissions before restoring service.
  2. Rollback the affected artifact and preserve trace evidence for root-cause analysis.
  3. Validate tool outputs as data against the declared schema and ignore embedded control directives.
  4. Revoke the token immediately, contain access and audit recent transactions.

Explanation

Correct Answer

D:

Credential incident containment requires prompt revocation and impact reconstruction.

Incorrect Answers

A:

Protecting data takes precedence over continued ungoverned availability, but credential incident containment requires prompt revocation and impact reconstruction.

B:

Reversible releases and forensic logs reduce incident duration and support diagnosis. Credential incident containment requires prompt revocation and impact reconstruction.

C:

Tool output has lower trust than the configured execution policy; credential incident containment requires prompt revocation and impact reconstruction.

Question 398

A deployed assistant begins citing restricted records after an access-control configuration change. Which design choice targets the observed problem?

  1. Disable the affected capability and verify permissions before restoring service.
  2. Capture authorized retrieval events with subject, source and purpose metadata.
  3. Rollback the affected artifact and preserve trace evidence for root-cause analysis.
  4. Produce protected residency, configuration and execution audit records.

Explanation

Correct Answer

A:

Protecting data takes precedence over continued ungoverned availability.

Incorrect Answers

B:

Access auditing can demonstrate when protected information was used by the assistant; protecting data takes precedence over continued ungoverned availability.

C:

Reversible releases and forensic logs reduce incident duration and support diagnosis, whereas protecting data takes precedence over continued ungoverned availability.

D:

Compliance evidence must tie the specific transaction to effective controls and versions. Protecting data takes precedence over continued ungoverned availability.

Question 399

An AI model begins producing harmful outputs after a newly installed prompt package. Which concrete solution decision should the team make?

  1. Use authoritative retrieval, citation verification and escalation for unsupported legal claims.
  2. Rollback the affected artifact and preserve trace evidence for root-cause analysis.
  3. Produce protected residency, configuration and execution audit records.
  4. Revoke the token immediately, contain access and audit recent transactions.

Explanation

Correct Answer

B:

Reversible releases and forensic logs reduce incident duration and support diagnosis.

Incorrect Answers

A:

Evidence constraints and review address harmful fabricated citations; reversible releases and forensic logs reduce incident duration and support diagnosis.

C:

Compliance evidence must tie the specific transaction to effective controls and versions. Reversible releases and forensic logs reduce incident duration and support diagnosis.

D:

Credential incident containment requires prompt revocation and impact reconstruction, but reversible releases and forensic logs reduce incident duration and support diagnosis.

Question 400

Regulators request proof of processing location and model versions used in a disputed customer workflow. Which design decision should the architect prioritize?

  1. Apply location and minimization controls to telemetry and audit storage.
  2. Disable the affected capability and verify permissions before restoring service.
  3. Produce protected residency, configuration and execution audit records.
  4. Revoke the token immediately, contain access and audit recent transactions.

Explanation

Correct Answer

C:

Compliance evidence must tie the specific transaction to effective controls and versions.

Incorrect Answers

A:

Observability data may contain protected information and require the same governance review, whereas compliance evidence must tie the specific transaction to effective controls and versions.

B:

Protecting data takes precedence over continued ungoverned availability. Compliance evidence must tie the specific transaction to effective controls and versions.

D:

Credential incident containment requires prompt revocation and impact reconstruction. Compliance evidence must tie the specific transaction to effective controls and versions.

Agentic AI Business Solutions Architect (AB-100) Practice Tests

UP

SPECIAL OFFER: GET 10% OFF

This is ONE TIME OFFER

ExamSnap Discount Offer
Enter Your Email Address to Receive Your 10% Off Discount Code

A confirmation link will be sent to this email address to verify your login. *We value your privacy. We will not rent or sell your email address.

Download Free Demo of VCE Exam Simulator

Experience Avanset VCE Exam Simulator for yourself.

Simply submit your e-mail address below to get started with our interactive software demo of your free trial.

Free Demo Limits: In the demo version you will be able to access only first 5 questions from exam.