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 6

Question 50

A project needs the following AI behavior: An architect wants an AI model to classify a picture as indoors or outdoors. What should the designer implement?

  1. Use regression or forecasting appropriate to time-series data
  2. Use object detection when each product needs a location
  3. Use image classification for the requested category
  4. Use a generative language workload with policy-grounded response constraints

Explanation

Correct Answer

C:

Image classification predicts image-level labels. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Forecasts predict future numeric quantities rather than discrete tags. That is not sufficient here: Image classification predicts image-level labels.

B:

Object detectors return classes and spatial boundaries. That is not sufficient here: Image classification predicts image-level labels.

D:

Generation produces novel text, unlike choosing a fixed class label. That is not sufficient here: Image classification predicts image-level labels.

Question 51

An AI operations team is testing two distinct capabilities. Failure (1): The response says the photo shows writing but has no extracted characters. Failure (2): The app calls a text classification API and expects a picture. Which TWO specific corrections should the team apply?

  1. Define extraction fields and validate values
  2. Use image text extraction and low-confidence human review
  3. Choose a text-to-image generator instead of a text classification API
  4. Use translation and inspect named terms
  5. Use time-series forecasting and numeric error metrics

Explanation

Correct Answers

B:

Correct for failure (1). Text recognition reads characters rather than merely describing visual objects. The proposed correction directly addresses that reported defect.

C:

Correct for failure (2). Text-to-image generation creates new visual content from a description; classification only assigns labels to existing inputs. It independently corrects the second defect.

Incorrect Answers

A:

Document extraction maps unstructured layouts into structured values. In this case the first outcome requires the team to use OCR tuned to the source quality and validate extracted text, and the second requires it to identify a text-to-image generative workload, not an existing-image classifier. This option does not replace those two separate controls.

D:

Translation transfers meaning across languages rather than changing topic labels. In this case the first outcome requires the team to use OCR tuned to the source quality and validate extracted text, and the second requires it to identify a text-to-image generative workload, not an existing-image classifier. This option does not replace those two separate controls.

E:

Forecasts predict future numeric quantities rather than discrete tags. In this case the first outcome requires the team to use OCR tuned to the source quality and validate extracted text, and the second requires it to identify a text-to-image generative workload, not an existing-image classifier. This option does not replace those two separate controls.

Question 52

An Azure-based AI team must satisfy this objective: An inspection solution must locate several defective parts within a photograph. Which design choice is best?

  1. Use the same compatible embedding version and dimensions for stored items and queries
  2. Confirm the model deployment identifier and endpoint resource match
  3. Choose a speech-recognition model or audio-capable multimodal endpoint with a documented audio-input contract
  4. Use object detection with bounding boxes and labels

Explanation

Correct Answer

D:

An appropriate spatial detection model must localize the defect classes, not merely assign an overall picture label.

Incorrect Answers

A:

Incompatible vector spaces can make nearest-neighbor rankings meaningless. That is not sufficient here: An appropriate spatial detection model must localize the defect classes, not merely assign an overall picture label.

B:

Applications call a deployed model resource rather than a generic marketing model name. That is not sufficient here: An appropriate spatial detection model must localize the defect classes, not merely assign an overall picture label.

C:

Model capabilities must include the required audio modality before the client can obtain speech-to-text results. That is not sufficient here: An appropriate spatial detection model must localize the defect classes, not merely assign an overall picture label.

Question 53

Troubleshooting report: Rejected applicants receive an opaque outcome with no case identifier. Which correction is MOST directly targeted to this failure?

  1. Assess both unsafe misses and harmful overblocking
  2. Add auditable explanation records and an appeals workflow
  3. Enforce backend authorization and signed approval regardless of prompt wording
  4. Gate the payment tool with verified role and amount checks

Explanation

Correct Answer

B:

This correction directly addresses the reported failure. Accountability includes traceable responsibility and appropriate contestability. The targeted change is: Add auditable explanation records and an appeals workflow.

Incorrect Answers

A:

Safety controls should not overblock permitted assistance. In the stated case, troubleshooting report: Rejected applicants receive an opaque outcome with no case identifier; “Assess both unsafe misses and harmful overblocking” does not target that failure.

C:

Accountable high-impact operations need verifiable authorized signatures. In the stated case, troubleshooting report: Rejected applicants receive an opaque outcome with no case identifier; “Enforce backend authorization and signed approval regardless of prompt wording” does not target that failure.

D:

A prompt-level request cannot override a transaction policy. In the stated case, troubleshooting report: Rejected applicants receive an opaque outcome with no case identifier; “Gate the payment tool with verified role and amount checks” does not target that failure.

Question 54

An Azure AI solution uses the following approved control: Include palette and composition constraints in the image prompt and review output. What observable result should pass QA?

  1. Generated samples are checked for the required color and layout
  2. The result includes supported text, tags and person boxes without conflation
  3. The test workload recovers without unbounded retry bursts
  4. The response distinguishes readable values from estimates

Explanation

Correct Answer

A:

Explicit visual guidance can influence generated appearance but is not absolute. The expected acceptance evidence is: Generated samples are checked for the required color and layout. This observation directly tests the stated behavior.

Incorrect Answers

B:

Different visual outputs describe text, scene classes and detected people. This observation does not establish the required result: Generated samples are checked for the required color and layout.

C:

Too many concurrent requests can exceed resource quota. This observation does not establish the required result: Generated samples are checked for the required color and layout.

D:

Visual reasoning is not proof of an exact value when axes are unclear. This observation does not establish the required result: Generated samples are checked for the required color and layout.

Question 55

Troubleshooting report: The code hardcodes a default region from another environment. Which correction is MOST directly targeted to this failure?

  1. Use the correct audio encoding and recognized input format
  2. Select the correct source/target language translation path
  3. Load the deployed speech resource's region and endpoint
  4. Add a no-speech branch and clarification option

Explanation

Correct Answer

C:

This correction directly addresses the reported failure. Region or endpoint mismatch prevents requests reaching the correct resource. The targeted change is: Load the deployed speech resource's region and endpoint.

Incorrect Answers

A:

Speech recognition requires the service to decode audio correctly. In the stated case, troubleshooting report: The code hardcodes a default region from another environment; “Use the correct audio encoding and recognized input format” does not target that failure.

B:

Translation requires semantic equivalence and language direction. In the stated case, troubleshooting report: The code hardcodes a default region from another environment; “Select the correct source/target language translation path” does not target that failure.

D:

An empty recognition result is not semantically equivalent to the word no. In the stated case, troubleshooting report: The code hardcodes a default region from another environment; “Add a no-speech branch and clarification option” does not target that failure.

Question 56

An AI operations team is testing two distinct capabilities. Failure (1): The application merely records the user's speech. Failure (2): The model labels the entire image as a store without locating items. Which TWO specific corrections should the team apply?

  1. Invoke a speech synthesis stage
  2. Use a detection output with bounding boxes
  3. Analyze multiple temporal segments and preserve timecodes
  4. Choose a text-to-image generator instead of a text classification API
  5. Use time-series forecasting and numeric error metrics

Explanation

Correct Answers

A:

Correct for failure (1). Speech synthesis converts supplied text into audible output. The proposed correction directly addresses that reported defect.

B:

Correct for failure (2). Object detectors return classes and spatial boundaries. It independently corrects the second defect.

Incorrect Answers

C:

Video understanding retains order and time, unlike single-image tagging. In this case the first outcome requires the team to use text-to-speech with a supported voice and accessible controls, and the second requires it to use object detection when each product needs a location. This option does not replace those two separate controls.

D:

Text-to-image generation creates new visual content from a description; classification only assigns labels to existing inputs. In this case the first outcome requires the team to use text-to-speech with a supported voice and accessible controls, and the second requires it to use object detection when each product needs a location. This option does not replace those two separate controls.

E:

Forecasts predict future numeric quantities rather than discrete tags. In this case the first outcome requires the team to use text-to-speech with a supported voice and accessible controls, and the second requires it to use object detection when each product needs a location. This option does not replace those two separate controls.

Question 57

Troubleshooting report: The system confuses a generated tool name with executed tool output. Which correction is MOST directly targeted to this failure?

  1. Validate the tool call and observe an actual execution receipt
  2. Apply context selection and retain necessary evidence only
  3. Set the API request to the correct deployment name and endpoint
  4. Verify claims against trusted evidence and permit abstention

Explanation

Correct Answer

A:

This correction directly addresses the reported failure. Agent workflows can reason about tool selection but tools still need authorization. The targeted change is: Validate the tool call and observe an actual execution receipt.

Incorrect Answers

B:

Processing extra tokens increases cost and can affect response time. In the stated case, troubleshooting report: The system confuses a generated tool name with executed tool output; “Apply context selection and retain necessary evidence only” does not target that failure.

C:

Applications call a deployed model resource rather than a generic marketing model name. In the stated case, troubleshooting report: The system confuses a generated tool name with executed tool output; “Set the API request to the correct deployment name and endpoint” does not target that failure.

D:

Language-model likelihood does not independently establish factual truth. In the stated case, troubleshooting report: The system confuses a generated tool name with executed tool output; “Verify claims against trusted evidence and permit abstention” does not target that failure.

Question 58

An Azure AI Fundamentals solution has this requirement: An education app wants to identify passages containing personal names. Which implementation most directly addresses it?

  1. Use text-to-speech with a supported voice and accessible controls
  2. Use conversational generative AI with grounded retrieval and scoped dialogue state
  3. Use semantic document indexing and retrieval rather than only keyword equality
  4. Use PII-aware entity recognition with appropriate redaction

Explanation

Correct Answer

D:

Entity extraction locates sensitive spans for privacy handling. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Speech synthesis converts supplied text into audible output. That is not sufficient here: Entity extraction locates sensitive spans for privacy handling.

B:

The assistant should incorporate context while remaining connected to reliable sources. That is not sufficient here: Entity extraction locates sensitive spans for privacy handling.

C:

Semantic search can retrieve conceptually related passages. That is not sufficient here: Entity extraction locates sensitive spans for privacy handling.

Question 59

The approved design action is: Provide both images to an appropriate multimodal comparison workflow. Which test result would BEST establish that it worked?

  1. The crop keeps the subject in view at the specified aspect ratio
  2. A drawn overlay aligns each box with the corresponding person
  3. The response identifies changes actually visible between the two images
  4. The summary preserves the order of observed demonstration steps and identifies gaps or uncertain actions

Explanation

Correct Answer

C:

Change detection needs both temporal states for a meaningful comparison. The expected acceptance evidence is: The response identifies changes actually visible between the two images. This observation directly tests the stated behavior.

Incorrect Answers

A:

Smart cropping attempts to retain salient regions within target dimensions. This observation does not establish the required result: The response identifies changes actually visible between the two images.

B:

Localization metadata is tied to pixel regions, not just labels. This observation does not establish the required result: The response identifies changes actually visible between the two images.

D:

A single still cannot convey action order; sampled video frames may also omit fleeting intermediate steps. This observation does not establish the required result: The response identifies changes actually visible between the two images.

Question 60

An Azure-based AI team must satisfy this objective: A generative assistant for public services omits captions for audio explanations. Which design choice is best?

  1. Evaluate safety again after tool and retrieval results are incorporated
  2. Provide synchronized transcripts or captions for spoken content
  3. Apply retention and deletion controls subject to recorded legal exceptions
  4. Assign accountable operational and risk owners with escalation responsibilities

Explanation

Correct Answer

B:

Accessible information must not depend on hearing audio. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Threats can enter later stages even when the original user request was harmless. That is not sufficient here: Accessible information must not depend on hearing audio.

C:

Accountability requires evidence of actual data lifecycle actions. That is not sufficient here: Accessible information must not depend on hearing audio.

D:

An AI model cannot be its own accountable incident responder. That is not sufficient here: Accessible information must not depend on hearing audio.

Microsoft Azure AI Fundamentals (AI-901) Practice Tests