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 43

Question 420

Troubleshooting report: The model gives answers without loading approved manuals. Which correction is MOST directly targeted to this failure?

  1. Use layout-aware structured invoice extraction
  2. Detect people and helmets and evaluate spatial associations
  3. Introduce a document classifier and route by type
  4. Add retrieval and grounded citation generation

Explanation

Correct Answer

D:

This correction directly addresses the reported failure. RAG retrieves current external evidence for a generated response. The targeted change is: Add retrieval and grounded citation generation.

Incorrect Answers

A:

Table extraction associates quantities, descriptions and prices by row. In the stated case, troubleshooting report: The model gives answers without loading approved manuals; “Use layout-aware structured invoice extraction” does not target that failure.

B:

Detection relates objects to positions for compliance checks. In the stated case, troubleshooting report: The model gives answers without loading approved manuals; “Detect people and helmets and evaluate spatial associations” does not target that failure.

C:

Classification selects the extraction process matching the document category. In the stated case, troubleshooting report: The model gives answers without loading approved manuals; “Introduce a document classifier and route by type” does not target that failure.

Question 421

Troubleshooting report: The submit action is an unlabeled graphical control. Which correction is MOST directly targeted to this failure?

  1. Propagate typed tool errors and use an explicit unavailable response
  2. Add an accessible name and keyboard activation to the control
  3. Reconcile the provider transaction ID and actual status before confirmation
  4. Exclude obsolete material with approval and timestamp filters

Explanation

Correct Answer

B:

This correction directly addresses the reported failure. Inclusive design permits nonvisual navigation of the same task path. The targeted change is: Add an accessible name and keyboard activation to the control.

Incorrect Answers

A:

A fluent generation cannot substitute for missing authoritative transaction data. In the stated case, troubleshooting report: The submit action is an unlabeled graphical control; “Propagate typed tool errors and use an explicit unavailable response” does not target that failure.

C:

Generated text is not evidence of external side effects. In the stated case, troubleshooting report: The submit action is an unlabeled graphical control; “Reconcile the provider transaction ID and actual status before confirmation” does not target that failure.

D:

A precise explanation of an obsolete procedure can still be dangerous. In the stated case, troubleshooting report: The submit action is an unlabeled graphical control; “Exclude obsolete material with approval and timestamp filters” does not target that failure.

Question 422

An Azure AI solution uses the following approved control: Ask for clarification or abstain rather than acting on inferred intent. What observable result should pass QA?

  1. An unsafe self-treatment case triggers a safe refusal or professional referral
  2. An incident simulation reaches a named owner with clear containment authority
  3. An ambiguous request results in a clarification question and no side effect
  4. A deaf tester receives the complete equivalent content as text

Explanation

Correct Answer

C:

Reliable assistance distinguishes missing intent from authorized action. The expected acceptance evidence is: An ambiguous request results in a clarification question and no side effect. 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: An ambiguous request results in a clarification question and no side effect.

B:

An AI model cannot be its own accountable incident responder. This observation does not establish the required result: An ambiguous request results in a clarification question and no side effect.

D:

Accessible information must not depend on hearing audio. This observation does not establish the required result: An ambiguous request results in a clarification question and no side effect.

Question 423

A solution team has selected this control: Restrict retrieval to effective approved documents and expose source dates. Which observation is the most direct acceptance evidence?

  1. Each policy answer citation resolves to the correct approved paragraph
  2. A changed regulation appears in answers with its approved effective date
  3. Labeled accessible-use examples pass while harmful variants remain blocked
  4. An auditor can map each used input to an approved policy criterion

Explanation

Correct Answer

B:

Truthfulness requires evidence that corresponds to the operative version. The expected acceptance evidence is: A changed regulation appears in answers with its approved effective date. This observation directly tests the stated behavior.

Incorrect Answers

A:

Transparency lets users inspect evidence instead of accepting fluent text. This observation does not establish the required result: A changed regulation appears in answers with its approved effective date.

C:

Safety controls should not overblock permitted assistance. This observation does not establish the required result: A changed regulation appears in answers with its approved effective date.

D:

Accountability requires interpretable reasons for consequential decisions. This observation does not establish the required result: A changed regulation appears in answers with its approved effective date.

Question 424

An Azure-based AI team must satisfy this objective: A user speaks several commands but the bot triggers on partial noisy fragments. Which design choice is best?

  1. Wait for supported final recognition results before high-impact tool calls
  2. Configure a supported Spanish recognition locale and test representative speakers
  3. Use Speech Synthesis Markup Language pauses or breaks where supported
  4. Call key phrase extraction for supported text inputs

Explanation

Correct Answer

A:

Interim hypotheses can change as more audio arrives. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

B:

Recognition models need the appropriate supported language settings. That is not sufficient here: Interim hypotheses can change as more audio arrives.

C:

SSML controls aspects of prosody and timing for synthesized speech. That is not sufficient here: Interim hypotheses can change as more audio arrives.

D:

Key phrases highlight important concepts expressed in the original text. That is not sufficient here: Interim hypotheses can change as more audio arrives.

Question 425

An Azure AI solution uses the following approved control: Choose a supported multilingual model or preprocessing route and test each locale. What observable result should pass QA?

  1. An echo-free test response correctly addresses the supplied question
  2. Repeated events yield at most one order creation under the same operation ID
  3. Labeled examples for target languages meet acceptance thresholds
  4. A smoke test reaches the intended new deployment and parses the output

Explanation

Correct Answer

C:

Language coverage can vary by model and deployment. The expected acceptance evidence is: Labeled examples for target languages meet acceptance thresholds. This observation directly tests the stated behavior.

Incorrect Answers

A:

The inference service needs task content to generate a relevant answer. This observation does not establish the required result: Labeled examples for target languages meet acceptance thresholds.

B:

A timeout does not prove that the original mutation failed. This observation does not establish the required result: Labeled examples for target languages meet acceptance thresholds.

D:

A model family label is not necessarily the concrete deployed endpoint. This observation does not establish the required result: Labeled examples for target languages meet acceptance thresholds.

Question 426

The approved design action is: Limit automatic language identification to supported locales and provide escalation. Which test result would BEST establish that it worked?

  1. Silent input yields a repeat or timeout response rather than a denial
  2. The translated request preserves numbers and intent
  3. Raw call recordings expire while authorized transcripts remain per policy
  4. Unsupported inputs yield a language selection or human support path

Explanation

Correct Answer

D:

Unrecognized languages must not be silently interpreted as a known one. The expected acceptance evidence is: Unsupported inputs yield a language selection or human support path. This observation directly tests the stated behavior.

Incorrect Answers

A:

An empty recognition result is not semantically equivalent to the word no. This observation does not establish the required result: Unsupported inputs yield a language selection or human support path.

B:

Translation requires semantic equivalence and language direction. This observation does not establish the required result: Unsupported inputs yield a language selection or human support path.

C:

Speech workloads still process personal data and require controlled lifecycle. This observation does not establish the required result: Unsupported inputs yield a language selection or human support path.

Question 427

An acceptance tester needs to validate: Apply confidence-based review instead of forcing every field into a number. Which observed result proves the intended behavior?

  1. Ambiguous handwritten values are flagged and linked to source image
  2. A reply links its claim to a transcript time or relevant video segment
  3. A test invoice yields correct merchant, date and total with source anchors
  4. A question links to the correct contract section and page

Explanation

Correct Answer

A:

Low-quality inputs can make extraction genuinely uncertain. The expected acceptance evidence is: Ambiguous handwritten values are flagged and linked to source image. This observation directly tests the stated behavior.

Incorrect Answers

B:

Multimodal grounding should preserve how the answer maps to source media. This observation does not establish the required result: Ambiguous handwritten values are flagged and linked to source image.

C:

Document extraction turns unstructured pages into validated structured fields. This observation does not establish the required result: Ambiguous handwritten values are flagged and linked to source image.

D:

Structured layout and provenance let users verify a quoted clause. This observation does not establish the required result: Ambiguous handwritten values are flagged and linked to source image.

Question 428

A solution team has selected this control: Preserve the higher-priority application constraints in the message hierarchy. Which observation is the most direct acceptance evidence?

  1. Conflict tests keep disallowed actions blocked
  2. Injection attempts fail to access protected secrets or tools
  3. The answer mentions relevant visual details unique to the uploaded image
  4. Lookup questions invoke read-only tools and refund tests remain gated

Explanation

Correct Answer

A:

Developer and system constraints govern allowed operations over user requests. The expected acceptance evidence is: Conflict tests keep disallowed actions blocked. This observation directly tests the stated behavior.

Incorrect Answers

B:

A user cannot use text to elevate permissions beyond the application policy. This observation does not establish the required result: Conflict tests keep disallowed actions blocked.

C:

The model must receive the actual visual content to reason about it. This observation does not establish the required result: Conflict tests keep disallowed actions blocked.

D:

Clear tool contracts reduce confusion between read and write operations. This observation does not establish the required result: Conflict tests keep disallowed actions blocked.

Question 429

Troubleshooting report: A single image-level classification is mistaken for a box count. Which correction is MOST directly targeted to this failure?

  1. Choose a text-to-image generator instead of a text classification API
  2. Require a user-approved booking and payment action
  3. Introduce a document classifier and route by type
  4. Count distinct detections after overlap filtering

Explanation

Correct Answer

D:

This correction directly addresses the reported failure. Instance detections support counts where one image contains several objects. The targeted change is: Count distinct detections after overlap filtering.

Incorrect Answers

A:

Text-to-image generation creates new visual content from a description; classification only assigns labels to existing inputs. In the stated case, troubleshooting report: A single image-level classification is mistaken for a box count; “Choose a text-to-image generator instead of a text classification API” does not target that failure.

B:

An agent may propose actions but payment must remain permission-controlled. In the stated case, troubleshooting report: A single image-level classification is mistaken for a box count; “Require a user-approved booking and payment action” does not target that failure.

C:

Classification selects the extraction process matching the document category. In the stated case, troubleshooting report: A single image-level classification is mistaken for a box count; “Introduce a document classifier and route by type” does not target that failure.

Question 430

A solution team has selected this control: Exclude protected proxies from eligibility logic and test the remaining permitted factors. Which observation is the most direct acceptance evidence?

  1. Exceptional high-impact cases are tested with explicit pass criteria
  2. Missing dosage context triggers abstention and a review route
  3. Equally qualified synthetic applicants receive the same recommendation across protected attributes
  4. A legally ambiguous case offers appropriate caveats and referral

Explanation

Correct Answer

C:

Legitimate credit criteria cannot be replaced by sensitive identity traits. The expected acceptance evidence is: Equally qualified synthetic applicants receive the same recommendation across protected attributes. This observation directly tests the stated behavior.

Incorrect Answers

A:

A high overall score can conceal severe failure on unusual inputs. This observation does not establish the required result: Equally qualified synthetic applicants receive the same recommendation across protected attributes.

B:

Safety-critical recommendations require verified constraints outside open-ended generation. This observation does not establish the required result: Equally qualified synthetic applicants receive the same recommendation across protected attributes.

D:

Transparency reduces mistaken reliance on unsupported professional guarantees. This observation does not establish the required result: Equally qualified synthetic applicants receive the same recommendation across protected attributes.

Microsoft Azure AI Fundamentals (AI-901) Practice Tests