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 1

Question 1

An engineering review documents this requirement: A training site wants to generate a clear summary of a long technical webinar. Select the most appropriate solution.

  1. Process the audio transcript and optionally key video scenes before summarizing
  2. Use a generative language workload with policy-grounded response constraints
  3. Use text embeddings with nearest-neighbor search and business filters
  4. Use key phrase extraction to identify salient text spans

Explanation

Correct Answer

A:

Meeting summaries require temporal and spoken source content. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

B:

Generation produces novel text, unlike choosing a fixed class label. That is not sufficient here: Meeting summaries require temporal and spoken source content.

C:

Semantic vectors capture related concepts while filters enforce eligibility. That is not sufficient here: Meeting summaries require temporal and spoken source content.

D:

Key phrase extraction finds prominent topics without inventing new source material. That is not sufficient here: Meeting summaries require temporal and spoken source content.

Question 2

An AI operations team is testing two distinct capabilities. Failure (1): A global caption describes only the image background. Failure (2): The model receives only the latest image but reports differences. Which TWO specific corrections should the team apply?

  1. Parse outputs by their documented feature-specific schema
  2. Include the earlier reference image and compare grounded features
  3. Crop or request region-focused visual analysis
  4. Constrain the prompt and validate the final artwork
  5. Use image input and OCR or vision grounding

Explanation

Correct Answers

B:

Correct for failure (2). Change detection needs both temporal states for a meaningful comparison. It independently corrects the second defect.

C:

Correct for failure (1). Specific regions can carry details missed by a whole-image caption. The proposed correction directly addresses that reported defect.

Incorrect Answers

A:

Parsing tag or OCR output by schema may normalize an existing result, but it does not focus captioning on the missed upper-left object and cannot compare photographs when the earlier reference image was never supplied.

D:

A prompt constraint can influence generated artwork, but these are understanding failures: a region of the existing image is overlooked, and a comparison lacks its previous image. Artwork-generation controls cannot supply either missing observation.

E:

OCR could read visible lettering in one image, but the stated first failure needs object-focused visual analysis, and the second requires both time-separated images. OCR on the latest image alone cannot establish what changed.

Question 3

A solution team has selected this control: Apply supported image safety filtering and review before distribution. Which observation is the most direct acceptance evidence?

  1. The count matches annotated packages after overlap filtering
  2. The output is a validated outdoor probability or category
  3. The thumbnail preserves subject proportions and important content
  4. Disallowed outputs are blocked in adversarial evaluation

Explanation

Correct Answer

D:

Generative media needs policy controls in addition to prompt instructions. The expected acceptance evidence is: Disallowed outputs are blocked in adversarial evaluation. This observation directly tests the stated behavior.

Incorrect Answers

A:

Object counts require resolving distinct instances, not trusting one scene label. This observation does not establish the required result: Disallowed outputs are blocked in adversarial evaluation.

B:

Image-level classification is distinct from object localization. This observation does not establish the required result: Disallowed outputs are blocked in adversarial evaluation.

C:

Smart cropping differs from simply stretching pixels. This observation does not establish the required result: Disallowed outputs are blocked in adversarial evaluation.

Question 4

An acceptance tester needs to validate: Use an image-classification model with class probabilities. Which observed result proves the intended behavior?

  1. The same brief yields a wider range of acceptable styles without broken constraints
  2. A held-out set receives the correct whole-image classes and empirically calibrated class probabilities
  3. The client succeeds using the deployment configured on the intended resource
  4. A permitted larger output budget allows the complete response

Explanation

Correct Answer

B:

Classification assigns image-wide categories rather than bounding-box coordinates. The expected acceptance evidence is: A held-out set receives the correct whole-image classes and empirically calibrated class probabilities. This observation directly tests the stated behavior.

Incorrect Answers

A:

More varied token sampling can increase creative alternatives. This observation does not establish the required result: A held-out set receives the correct whole-image classes and empirically calibrated class probabilities.

C:

Applications call a deployed model resource rather than a generic marketing model name. This observation does not establish the required result: A held-out set receives the correct whole-image classes and empirically calibrated class probabilities.

D:

Generation stops when an output budget or stop condition is reached. This observation does not establish the required result: A held-out set receives the correct whole-image classes and empirically calibrated class probabilities.

Question 5

An Azure AI solution uses the following approved control: Add an approved knowledge/retrieval tool with clear description and access limits. What observable result should pass QA?

  1. A test run shows correct tool selection, arguments and user-visible response
  2. A simulated timeout yields pending status with a safe retry or escalation
  3. The agent retrieves an authorized relevant passage and cites it
  4. An authenticated test request returns a response from the chosen deployment

Explanation

Correct Answer

C:

Tool descriptions guide selection, while backend rules enforce allowed sources. The expected acceptance evidence is: The agent retrieves an authorized relevant passage and cites it. This observation directly tests the stated behavior.

Incorrect Answers

A:

An answer that sounds plausible may conceal failed tool calls. This observation does not establish the required result: The agent retrieves an authorized relevant passage and cites it.

B:

Timeout handling should not be reported as a successful action. This observation does not establish the required result: The agent retrieves an authorized relevant passage and cites it.

D:

The portal provides model selection, deployment and test interfaces for supported resources. This observation does not establish the required result: The agent retrieves an authorized relevant passage and cites it.

Question 6

A solution team has selected this control: Test representative document layouts and accessibility-oriented inputs. Which observation is the most direct acceptance evidence?

  1. Equivalent skills are extracted correctly from multiple accessible résumé formats
  2. Scanner-specific evaluation identifies the weak cohort and blocks unsupported rollout
  3. Matched applications differing only in names receive consistent decisions
  4. Labeled accessible-use examples pass while harmful variants remain blocked

Explanation

Correct Answer

A:

Inclusion requires robust input handling beyond a single document template. The expected acceptance evidence is: Equivalent skills are extracted correctly from multiple accessible résumé formats. This observation directly tests the stated behavior.

Incorrect Answers

B:

Reliability requires detecting distribution shifts hidden in aggregate accuracy. This observation does not establish the required result: Equivalent skills are extracted correctly from multiple accessible résumé formats.

C:

Fairness checks compare materially equivalent cases rather than raw group averages alone. This observation does not establish the required result: Equivalent skills are extracted correctly from multiple accessible résumé formats.

D:

Safety controls should not overblock permitted assistance. This observation does not establish the required result: Equivalent skills are extracted correctly from multiple accessible résumé formats.

Question 7

An Azure-based AI team must satisfy this objective: A chat service receives an API authentication failure. Which design choice is best?

  1. Handle tool timeout as a typed error and preserve partial workflow state
  2. Capture allowed tool calls, response statuses and source references in traces
  3. Archive the actual evaluated response together with the prompt, model deployment, input and generation settings
  4. Check the selected credential, identity scope and target resource permission

Explanation

Correct Answer

D:

Authentication proves caller identity and authorization grants resource access. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Timeout handling should not be reported as a successful action. That is not sufficient here: Authentication proves caller identity and authorization grants resource access.

B:

Operational observability aids debugging without exposing hidden sensitive prompt data. That is not sufficient here: Authentication proves caller identity and authorization grants resource access.

C:

Recorded settings support investigation, but generative models can still vary on an apparently identical rerun. That is not sufficient here: Authentication proves caller identity and authorization grants resource access.

Question 8

An engineering review documents this requirement: An app needs multiple turns to reference the same user's active order. Select the most appropriate solution.

  1. Place the current request in the supported user message field
  2. Handle tool timeout as a typed error and preserve partial workflow state
  3. Maintain bounded authenticated conversation state with order context
  4. Implement a network timeout/retry strategy that avoids duplicate tool effects

Explanation

Correct Answer

C:

Conversation continuity should preserve identifiers without crossing user sessions. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

The inference service needs task content to generate a relevant answer. That is not sufficient here: Conversation continuity should preserve identifiers without crossing user sessions.

B:

Timeout handling should not be reported as a successful action. That is not sufficient here: Conversation continuity should preserve identifiers without crossing user sessions.

D:

Client retries should distinguish read requests from side-effecting tool calls. That is not sufficient here: Conversation continuity should preserve identifiers without crossing user sessions.

Question 9

Troubleshooting report: The model invents a resolution absent from the ticket. Which correction is MOST directly targeted to this failure?

  1. Set the supported Spanish locale and verify accuracy
  2. Ground the summary and penalize unsupported additions
  3. Use the sentiment endpoint and validate its supported language
  4. Use explicit language confirmation or a support fallback

Explanation

Correct Answer

B:

This correction directly addresses the reported failure. Summarization compresses written information rather than changing its facts. The targeted change is: Ground the summary and penalize unsupported additions.

Incorrect Answers

A:

Recognition models need the appropriate supported language settings. In the stated case, troubleshooting report: The model invents a resolution absent from the ticket; “Set the supported Spanish locale and verify accuracy” does not target that failure.

C:

Sentiment output classifies attitudes expressed in the submitted text. In the stated case, troubleshooting report: The model invents a resolution absent from the ticket; “Use the sentiment endpoint and validate its supported language” does not target that failure.

D:

Unrecognized languages must not be silently interpreted as a known one. In the stated case, troubleshooting report: The model invents a resolution absent from the ticket; “Use explicit language confirmation or a support fallback” does not target that failure.

Question 10

An Azure AI solution uses the following approved control: Pass relevant conversation context or a verified summarized state within token limits. What observable result should pass QA?

  1. A follow-up pronoun resolves to the correct earlier referent
  2. An out-of-scope request is handled under the approved restrictions
  3. An echo-free test response correctly addresses the supplied question
  4. Execution traces show the intended tool invoked with valid arguments

Explanation

Correct Answer

A:

Stateful conversations require supplying or retrieving relevant prior context. The expected acceptance evidence is: A follow-up pronoun resolves to the correct earlier referent. This observation directly tests the stated behavior.

Incorrect Answers

B:

High-priority instructions should describe approved behavior and escalation rules. This observation does not establish the required result: A follow-up pronoun resolves to the correct earlier referent.

C:

The inference service needs task content to generate a relevant answer. This observation does not establish the required result: A follow-up pronoun resolves to the correct earlier referent.

D:

Tools must be discoverable with valid schemas before selection. This observation does not establish the required result: A follow-up pronoun resolves to the correct earlier referent.

Microsoft Azure AI Fundamentals (AI-901) Practice Tests