McAfee Secure

Exam Code: AI-901

Exam Name: Microsoft Azure AI Fundamentals

Certification Provider: Microsoft

Corresponding Certification: Microsoft Certified: Azure AI Fundamentals

Microsoft AI-901 Questions & Answers

Study with Up-To-Date REAL Exam Questions and Answers from the ACTUAL Test

166 Questions & Answers with Testing Engine

"Microsoft Azure AI Fundamentals Exam", also known as AI-901 exam, is a Microsoft certification exam.

Pass your tests with the always up-to-date AI-901 Exam Engine. Your AI-901 training materials keep you at the head of the pack!

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Was: $137.49
Now: $124.99

Microsoft AI-901 Practice Test 69

Question 680

A project needs the following AI behavior: An AI agent reports that a refund succeeded even though the payment service returned an error. What should the designer implement?

  1. Test representative document layouts and accessibility-oriented inputs
  2. Assign accountable operational and risk owners with escalation responsibilities
  3. Verify an authoritative transaction receipt before claiming task completion
  4. Provide synchronized transcripts or captions for spoken content

Explanation

Correct Answer

C:

Generated text is not evidence of external side effects. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Inclusion requires robust input handling beyond a single document template. That is not sufficient here: Generated text is not evidence of external side effects.

B:

An AI model cannot be its own accountable incident responder. That is not sufficient here: Generated text is not evidence of external side effects.

D:

Accessible information must not depend on hearing audio. That is not sufficient here: Generated text is not evidence of external side effects.

Question 681

A project needs the following AI behavior: An application needs to detect that a scanned form includes a barcode. What should the designer implement?

  1. Use a video-capable Content Understanding analyzer with timestamps and supported fields
  2. Enable a supported barcode extraction or recognition feature
  3. Use supported locale-aware OCR or language handling with evaluation
  4. Classify or segment the combined PDF into logical documents before extraction

Explanation

Correct Answer

B:

A barcode is a structured visual symbol not ordinary OCR text. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Multimodal extraction draws on visual and spoken evidence over time. That is not sufficient here: A barcode is a structured visual symbol not ordinary OCR text.

C:

Recognition accuracy depends on script and language support. That is not sufficient here: A barcode is a structured visual symbol not ordinary OCR text.

D:

Document boundaries prevent values from unrelated invoices being joined. That is not sufficient here: A barcode is a structured visual symbol not ordinary OCR text.

Question 682

A solution team has selected this control: Use a conventional calculation function for exact arithmetic and let the model explain it. Which observation is the most direct acceptance evidence?

  1. Known arithmetic cases agree exactly with verified calculations
  2. The response includes the requested image artifact and metadata
  3. A user request cannot override the controlling system-level instructions in tests
  4. Holdout examples are assigned valid classes with measured accuracy

Explanation

Correct Answer

A:

Generative models are not a substitute for auditable deterministic computation. The expected acceptance evidence is: Known arithmetic cases agree exactly with verified calculations. This observation directly tests the stated behavior.

Incorrect Answers

B:

Text generation and image generation are distinct model output capabilities. This observation does not establish the required result: Known arithmetic cases agree exactly with verified calculations.

C:

Instruction roles help distinguish application policy from user task text. This observation does not establish the required result: Known arithmetic cases agree exactly with verified calculations.

D:

Classification predicts discrete labels rather than a continuous numeric target. This observation does not establish the required result: Known arithmetic cases agree exactly with verified calculations.

Question 683

A solution team has selected this control: Call the deployed model through an authenticated application backend. Which observation is the most direct acceptance evidence?

  1. The agent selects the expected tool for a supported test query
  2. Browser traffic cannot access the protected model credential
  3. The response contains only permitted fields with correct types
  4. The test output satisfies every defined requirement of the prompt

Explanation

Correct Answer

B:

Backend controls protect credentials, quotas and data policy. The expected acceptance evidence is: Browser traffic cannot access the protected model credential. This observation directly tests the stated behavior.

Incorrect Answers

A:

A single agent can orchestrate a bounded tool without a multi-agent system. This observation does not establish the required result: Browser traffic cannot access the protected model credential.

C:

Model output constraints help downstream parsing but still require validation. This observation does not establish the required result: Browser traffic cannot access the protected model credential.

D:

Specific instructions reduce ambiguity about what constitutes success. This observation does not establish the required result: Browser traffic cannot access the protected model credential.

Question 684

Troubleshooting report: The service predicts a numeric satisfaction score only. Which correction is MOST directly targeted to this failure?

  1. Use key phrase extraction rather than regression
  2. Select visual sign recognition rather than text-only extraction
  3. Add retrieval and grounded citation generation
  4. Embed or semantically index the documents with source provenance

Explanation

Correct Answer

A:

This correction directly addresses the reported failure. Key phrase extraction finds prominent topics without inventing new source material. The targeted change is: Use key phrase extraction rather than regression.

Incorrect Answers

B:

The task chooses a label, with localization needed only when multiple signs appear. In the stated case, troubleshooting report: The service predicts a numeric satisfaction score only; “Select visual sign recognition rather than text-only extraction” does not target that failure.

C:

RAG retrieves current external evidence for a generated response. In the stated case, troubleshooting report: The service predicts a numeric satisfaction score only; “Add retrieval and grounded citation generation” does not target that failure.

D:

Semantic search can retrieve conceptually related passages. In the stated case, troubleshooting report: The service predicts a numeric satisfaction score only; “Embed or semantically index the documents with source provenance” does not target that failure.

Question 685

Troubleshooting report: The app scores fluency as proof of correctness. Which correction is MOST directly targeted to this failure?

  1. Apply context selection and retain necessary evidence only
  2. Use a multimodal-capable deployment and supported message format
  3. Process appropriate video frames and the audio track
  4. Verify claims against trusted evidence and permit abstention

Explanation

Correct Answer

D:

This correction directly addresses the reported failure. Language-model likelihood does not independently establish factual truth. The targeted change is: Verify claims against trusted evidence and permit abstention.

Incorrect Answers

A:

Processing extra tokens increases cost and can affect response time. In the stated case, troubleshooting report: The app scores fluency as proof of correctness; “Apply context selection and retain necessary evidence only” does not target that failure.

B:

Multimodal models can reason over more than one supported input modality. In the stated case, troubleshooting report: The app scores fluency as proof of correctness; “Use a multimodal-capable deployment and supported message format” does not target that failure.

C:

Different media channels provide complementary scene and dialogue evidence. In the stated case, troubleshooting report: The app scores fluency as proof of correctness; “Process appropriate video frames and the audio track” does not target that failure.

Question 686

An Azure AI Fundamentals solution has this requirement: A medical imaging model performs well overall but fails for one scanner manufacturer. Which implementation most directly addresses it?

  1. Record decision reasons and give a human review route for important outcomes
  2. Evaluate safety again after tool and retrieval results are incorporated
  3. Maintain a governed feature inventory and trace decisions to policy
  4. Report subgroup performance by device and acquisition setting

Explanation

Correct Answer

D:

Reliability requires detecting distribution shifts hidden in aggregate accuracy. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Accountability includes traceable responsibility and appropriate contestability. That is not sufficient here: Reliability requires detecting distribution shifts hidden in aggregate accuracy.

B:

Threats can enter later stages even when the original user request was harmless. That is not sufficient here: Reliability requires detecting distribution shifts hidden in aggregate accuracy.

C:

Accountability requires interpretable reasons for consequential decisions. That is not sufficient here: Reliability requires detecting distribution shifts hidden in aggregate accuracy.

Question 687

A solution team has selected this control: Use array-typed field extraction rather than one scalar field. Which observation is the most direct acceptance evidence?

  1. Ambiguous totals are flagged for manual review with source crops
  2. Extracted line items reconcile to visible row amounts
  3. Each product is returned as a separate item with its quantity
  4. Total mismatches trigger a review rather than automatic posting

Explanation

Correct Answer

C:

Repeated records require structured lists with item-level associations. The expected acceptance evidence is: Each product is returned as a separate item with its quantity. This observation directly tests the stated behavior.

Incorrect Answers

A:

A low-confidence amount must not be silently treated as definitive. This observation does not establish the required result: Each product is returned as a separate item with its quantity.

B:

Cell relationships are necessary to check item totals. This observation does not establish the required result: Each product is returned as a separate item with its quantity.

D:

Independent business rules catch plausible extraction errors. This observation does not establish the required result: Each product is returned as a separate item with its quantity.

Question 688

An acceptance tester needs to validate: Represent uncertainty explicitly and refer to approved teaching sources. Which observed result proves the intended behavior?

  1. An unsafe self-treatment case triggers a safe refusal or professional referral
  2. An adversarial page is summarized without executing its prohibited instructions
  3. Language-specific test cases produce equivalent safe outcomes and escalation options
  4. Out-of-scope questions trigger qualified language and reference suggestions

Explanation

Correct Answer

D:

Honesty requires acknowledging when a claim cannot be substantiated. The expected acceptance evidence is: Out-of-scope questions trigger qualified language and reference suggestions. This observation directly tests the stated behavior.

Incorrect Answers

A:

Safety mechanisms should reduce unsupported high-impact recommendations. This observation does not establish the required result: Out-of-scope questions trigger qualified language and reference suggestions.

B:

Lower-trust text is not a privileged command channel. This observation does not establish the required result: Out-of-scope questions trigger qualified language and reference suggestions.

C:

Inclusion requires equivalent safety and meaning across supported locales. This observation does not establish the required result: Out-of-scope questions trigger qualified language and reference suggestions.

Question 689

An Azure AI Fundamentals solution has this requirement: A security team wants to prevent exfiltration of identity document images via extraction logs. Which implementation most directly addresses it?

  1. Run supported image text extraction with field schema and source location
  2. Extract currency code and monetary amount as separate validated fields
  3. Classify or segment the combined PDF into logical documents before extraction
  4. Minimize extracted PII and restrict diagnostic payload storage

Explanation

Correct Answer

D:

Inference outputs and telemetry can contain protected identity information. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Visual information extraction can map pixels to structured identifiers. That is not sufficient here: Inference outputs and telemetry can contain protected identity information.

B:

A currency symbol or amount alone may not establish financial meaning. That is not sufficient here: Inference outputs and telemetry can contain protected identity information.

C:

Document boundaries prevent values from unrelated invoices being joined. That is not sufficient here: Inference outputs and telemetry can contain protected identity information.

Question 690

An Azure-based AI team must satisfy this objective: A shop wants to automatically replace a green product background with white. Which design choice is best?

  1. Use machine translation with locale-specific quality review
  2. Use clustering with normalized behavioral features
  3. Use image editing or segmentation with object-mask preservation
  4. Use text classification with labeled intent examples

Explanation

Correct Answer

C:

Editing changes selected pixels while leaving product regions intact. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Translation transfers meaning across languages rather than changing topic labels. That is not sufficient here: Editing changes selected pixels while leaving product regions intact.

B:

Clustering discovers groups based on feature similarity. That is not sufficient here: Editing changes selected pixels while leaving product regions intact.

D:

Classification maps text to one of several predefined categories. That is not sufficient here: Editing changes selected pixels while leaving product regions intact.

Microsoft Azure AI Fundamentals (AI-901) Practice Tests