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 28

Question 270

An acceptance tester needs to validate: Use image interpretation followed by grounded text and speech synthesis. Which observed result proves the intended behavior?

  1. The result contains product boxes in their correct shelf positions
  2. The narrated explanation matches labeled chart features without inventing values
  3. The assistant produces a relevant answer with approved citations
  4. Known spikes score unusually relative to a validated baseline

Explanation

Correct Answer

B:

Multimodal input can describe the graphic and speech output can read that description. The expected acceptance evidence is: The narrated explanation matches labeled chart features without inventing values. This observation directly tests the stated behavior.

Incorrect Answers

A:

Object detectors return classes and spatial boundaries. This observation does not establish the required result: The narrated explanation matches labeled chart features without inventing values.

C:

Generation produces novel text, unlike choosing a fixed class label. This observation does not establish the required result: The narrated explanation matches labeled chart features without inventing values.

D:

Anomaly methods identify deviations rather than assigning preset incident labels. This observation does not establish the required result: The narrated explanation matches labeled chart features without inventing values.

Question 271

The approved design action is: Treat retrieved content as untrusted evidence and prevent it from changing tool authority. Which test result would BEST establish that it worked?

  1. An unsafe self-treatment case triggers a safe refusal or professional referral
  2. A legally ambiguous case offers appropriate caveats and referral
  3. An adversarial page is summarized without executing its prohibited instructions
  4. A changed safety instruction is used after the approved document refresh

Explanation

Correct Answer

C:

Lower-trust text is not a privileged command channel. The expected acceptance evidence is: An adversarial page is summarized without executing its prohibited instructions. 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 adversarial page is summarized without executing its prohibited instructions.

B:

Transparency reduces mistaken reliance on unsupported professional guarantees. This observation does not establish the required result: An adversarial page is summarized without executing its prohibited instructions.

D:

A precise explanation of an obsolete procedure can still be dangerous. This observation does not establish the required result: An adversarial page is summarized without executing its prohibited instructions.

Question 272

Troubleshooting report: The team calls speech-to-text on already written text. Which correction is MOST directly targeted to this failure?

  1. Use a speech synthesis endpoint and compatible output format
  2. Process appropriate video frames and the audio track
  3. Create compatible text embeddings and evaluate vector retrieval
  4. Refresh indexed grounding sources and verify their effective date

Explanation

Correct Answer

A:

This correction directly addresses the reported failure. Speech synthesis generates audio from text rather than transcribing audio. The targeted change is: Use a speech synthesis endpoint and compatible output format.

Incorrect Answers

B:

Different media channels provide complementary scene and dialogue evidence. In the stated case, troubleshooting report: The team calls speech-to-text on already written text; “Process appropriate video frames and the audio track” does not target that failure.

C:

Embeddings encode learned semantic relationships rather than literal string identity. In the stated case, troubleshooting report: The team calls speech-to-text on already written text; “Create compatible text embeddings and evaluate vector retrieval” does not target that failure.

D:

External retrieval updates factual context without retraining model weights. In the stated case, troubleshooting report: The team calls speech-to-text on already written text; “Refresh indexed grounding sources and verify their effective date” does not target that failure.

Question 273

Troubleshooting report: Requests still point to a retired deployment identifier. Which correction is MOST directly targeted to this failure?

  1. Update configuration and run supported compatibility checks
  2. Use managed identity or rotated approved credentials and correct role assignments
  3. Version prompt and deployment settings and retain the actual evaluation output
  4. Add idempotency keys and query the original transaction state

Explanation

Correct Answer

A:

This correction directly addresses the reported failure. A model family label is not necessarily the concrete deployed endpoint. The targeted change is: Update configuration and run supported compatibility checks.

Incorrect Answers

B:

Authentication proves caller identity and authorization grants resource access. In the stated case, troubleshooting report: Requests still point to a retired deployment identifier; “Use managed identity or rotated approved credentials and correct role assignments” does not target that failure.

C:

Recorded settings support investigation, but generative models can still vary on an apparently identical rerun. In the stated case, troubleshooting report: Requests still point to a retired deployment identifier; “Version prompt and deployment settings and retain the actual evaluation output” does not target that failure.

D:

A timeout does not prove that the original mutation failed. In the stated case, troubleshooting report: Requests still point to a retired deployment identifier; “Add idempotency keys and query the original transaction state” does not target that failure.

Question 274

Troubleshooting report: A text-to-speech service creates more audio instead of transcription. Which correction is MOST directly targeted to this failure?

  1. Switch to continuous recognition with stable stream lifecycle
  2. Send valid audio to a speech recognition endpoint
  3. Enable supported diarization and validate turn assignments
  4. Invoke a text-to-speech voice and compatible audio format

Explanation

Correct Answer

B:

This correction directly addresses the reported failure. Speech recognition converts waveform features to text tokens. The targeted change is: Send valid audio to a speech recognition endpoint.

Incorrect Answers

A:

Long interactions need streaming or segmented recognition over time. In the stated case, troubleshooting report: A text-to-speech service creates more audio instead of transcription; “Switch to continuous recognition with stable stream lifecycle” does not target that failure.

C:

Speaker diarization separates turns and identities at a conversation level. In the stated case, troubleshooting report: A text-to-speech service creates more audio instead of transcription; “Enable supported diarization and validate turn assignments” does not target that failure.

D:

Speech synthesis generates audio waveforms from supplied text. In the stated case, troubleshooting report: A text-to-speech service creates more audio instead of transcription; “Invoke a text-to-speech voice and compatible audio format” does not target that failure.

Question 275

Troubleshooting report: The UI reports complete immediately after submitting a run. Which correction is MOST directly targeted to this failure?

  1. Move system instructions into protected application configuration
  2. Create a supported deployment and confirm resource access
  3. Tighten the output contract and apply post-response validation
  4. Poll or await the documented completion status before reporting success

Explanation

Correct Answer

D:

This correction directly addresses the reported failure. Submission is not the same as successful completion. The targeted change is: Poll or await the documented completion status before reporting success.

Incorrect Answers

A:

Versioned system messages provide a controlled behavioral baseline. In the stated case, troubleshooting report: The UI reports complete immediately after submitting a run; “Move system instructions into protected application configuration” does not target that failure.

B:

The portal provides model selection, deployment and test interfaces for supported resources. In the stated case, troubleshooting report: The UI reports complete immediately after submitting a run; “Create a supported deployment and confirm resource access” does not target that failure.

C:

Explicit examples can improve adherence to bounded output conventions. In the stated case, troubleshooting report: The UI reports complete immediately after submitting a run; “Tighten the output contract and apply post-response validation” does not target that failure.

Question 276

A project needs the following AI behavior: An application needs to recognize whether an image contains a traffic sign. What should the designer implement?

  1. Use an agent workflow with a permitted ticket tool and an authorization gate
  2. Use image editing or segmentation with object-mask preservation
  3. Use a visual classification or detection model appropriate to whether location matters
  4. Use a relevance-ranking or recommendation model evaluated on resolved-case history

Explanation

Correct Answer

C:

The task chooses a label, with localization needed only when multiple signs appear. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Agentic workloads combine reasoning and bounded external actions. That is not sufficient here: The task chooses a label, with localization needed only when multiple signs appear.

B:

Editing changes selected pixels while leaving product regions intact. That is not sufficient here: The task chooses a label, with localization needed only when multiple signs appear.

D:

Recommendations rank candidate items rather than simply generating arbitrary titles. That is not sufficient here: The task chooses a label, with localization needed only when multiple signs appear.

Question 277

Troubleshooting report: The system reports only high/low categories without quantities. Which correction is MOST directly targeted to this failure?

  1. Use time-series forecasting and numeric error metrics
  2. Use classification outputs for the entire image
  3. Treat absent fields as unknown and request a better image
  4. Use document layout extraction with typed checkbox outputs

Explanation

Correct Answer

A:

This correction directly addresses the reported failure. Forecasts predict future numeric quantities rather than discrete tags. The targeted change is: Use time-series forecasting and numeric error metrics.

Incorrect Answers

B:

Image classification predicts image-level labels. In the stated case, troubleshooting report: The system reports only high/low categories without quantities; “Use classification outputs for the entire image” does not target that failure.

C:

Images of receipts need OCR/layout extraction before typed business values. In the stated case, troubleshooting report: The system reports only high/low categories without quantities; “Treat absent fields as unknown and request a better image” does not target that failure.

D:

Different form elements require more than ordinary running-text OCR. In the stated case, troubleshooting report: The system reports only high/low categories without quantities; “Use document layout extraction with typed checkbox outputs” does not target that failure.

Question 278

An engineering review documents this requirement: A beginner asks why a large language model produces the next sentence word by word. Select the most appropriate solution.

  1. Use retrieval or chunking to supply only relevant passages
  2. Explain token-based next-token prediction conditioned on preceding context
  3. Use regression trained on historical feature/target pairs
  4. Use a supervised classification model with representative labeled cases

Explanation

Correct Answer

B:

A generative language model estimates likely continuations rather than reading a stored answer paragraph. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

A finite context window limits how much material can be processed at once. That is not sufficient here: A generative language model estimates likely continuations rather than reading a stored answer paragraph.

C:

Regression predicts continuous quantities under a defined error metric. That is not sufficient here: A generative language model estimates likely continuations rather than reading a stored answer paragraph.

D:

Classification predicts discrete labels rather than a continuous numeric target. That is not sufficient here: A generative language model estimates likely continuations rather than reading a stored answer paragraph.

Question 279

An acceptance tester needs to validate: Compare model deployment, prompt version, retrieval index and tool configuration. Which observed result proves the intended behavior?

  1. The response cites a completed tool receipt and actual returned data
  2. The difference can be tied to an explicit configuration mismatch
  3. A follow-up pronoun resolves to the correct earlier referent
  4. The agent retrieves an authorized relevant passage and cites it

Explanation

Correct Answer

B:

Environment drift can change behavior without code edits. The expected acceptance evidence is: The difference can be tied to an explicit configuration mismatch. This observation directly tests the stated behavior.

Incorrect Answers

A:

Generated text does not prove external execution. This observation does not establish the required result: The difference can be tied to an explicit configuration mismatch.

C:

Stateful conversations require supplying or retrieving relevant prior context. This observation does not establish the required result: The difference can be tied to an explicit configuration mismatch.

D:

Tool descriptions guide selection, while backend rules enforce allowed sources. This observation does not establish the required result: The difference can be tied to an explicit configuration mismatch.

Question 280

A project needs the following AI behavior: An agent needs to search a trusted FAQ before answering a technical policy question. What should the designer implement?

  1. Use tenant-specific identity, endpoint configuration and permissions
  2. Check supported context and output limits for the chosen deployment
  3. Include a concise output schema with valid and invalid examples where appropriate
  4. Add an approved knowledge/retrieval tool with clear description and access limits

Explanation

Correct Answer

D:

Tool descriptions guide selection, while backend rules enforce allowed sources. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Copying code does not grant access to another tenant's resources. That is not sufficient here: Tool descriptions guide selection, while backend rules enforce allowed sources.

B:

Capabilities and limits vary by model deployment. That is not sufficient here: Tool descriptions guide selection, while backend rules enforce allowed sources.

C:

Explicit examples can improve adherence to bounded output conventions. That is not sufficient here: Tool descriptions guide selection, while backend rules enforce allowed sources.

Microsoft Azure AI Fundamentals (AI-901) Practice Tests