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 13

Question 120

A project needs the following AI behavior: A user requests a sensitive irreversible action through an agent. What should the designer implement?

  1. Require explicit confirmation and backend authorization before tool execution
  2. Write an instruction to identify missing inputs and request clarification before tools
  3. Select and test a compatible deployed model through the Microsoft Foundry portal
  4. Use a versioned evaluation set with stable scoring and test inputs

Explanation

Correct Answer

A:

High-impact side effects need permissions independent of conversational language. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

B:

Prompts can define required preconditions but tools must still enforce them. That is not sufficient here: High-impact side effects need permissions independent of conversational language.

C:

The portal provides model selection, deployment and test interfaces for supported resources. That is not sufficient here: High-impact side effects need permissions independent of conversational language.

D:

Anecdotal success on different prompts cannot establish an improvement. That is not sufficient here: High-impact side effects need permissions independent of conversational language.

Question 121

Troubleshooting report: The application treats temperature zero as a truth guarantee. Which correction is MOST directly targeted to this failure?

  1. Create and test a simple one-agent tool workflow
  2. Add fact checking and grounding instead of relying on sampling
  3. Freeze representative evaluation cases and compare versions
  4. Reinforce role boundaries and enforce secrets outside prompts

Explanation

Correct Answer

B:

This correction directly addresses the reported failure. Sampling parameters influence diversity but do not verify knowledge. The targeted change is: Add fact checking and grounding instead of relying on sampling.

Incorrect Answers

A:

A single agent can orchestrate a bounded tool without a multi-agent system. In the stated case, troubleshooting report: The application treats temperature zero as a truth guarantee; “Create and test a simple one-agent tool workflow” does not target that failure.

C:

Anecdotal success on different prompts cannot establish an improvement. In the stated case, troubleshooting report: The application treats temperature zero as a truth guarantee; “Freeze representative evaluation cases and compare versions” does not target that failure.

D:

Developer and system constraints govern allowed operations over user requests. In the stated case, troubleshooting report: The application treats temperature zero as a truth guarantee; “Reinforce role boundaries and enforce secrets outside prompts” does not target that failure.

Question 122

An Azure-based AI team must satisfy this objective: An analyst wants the main terms from long customer complaints. Which design choice is best?

  1. Call key phrase extraction for supported text inputs
  2. Call a supported text sentiment-analysis feature and interpret returned polarity
  3. Use a supported text-to-speech synthesis deployment
  4. Run the required text analysis operations separately and preserve distinct result types

Explanation

Correct Answer

A:

Key phrases highlight important concepts expressed in the original text. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

B:

Sentiment output classifies attitudes expressed in the submitted text. That is not sufficient here: Key phrases highlight important concepts expressed in the original text.

C:

Speech synthesis generates audio waveforms from supplied text. That is not sufficient here: Key phrases highlight important concepts expressed in the original text.

D:

Sentiment polarity and named entity spans answer different questions. That is not sufficient here: Key phrases highlight important concepts expressed in the original text.

Question 123

An Azure AI Fundamentals solution has this requirement: A phone support service needs to recognize a spoken intent while handling background noise. Which implementation most directly addresses it?

  1. Use noise-aware speech recognition and intent classification with fallback
  2. Use conversational generative AI with grounded retrieval and scoped dialogue state
  3. Use clustering with normalized behavioral features
  4. Use an agent workflow with a permitted ticket tool and an authorization gate

Explanation

Correct Answer

A:

Speech input quality affects subsequent language classification. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

B:

The assistant should incorporate context while remaining connected to reliable sources. That is not sufficient here: Speech input quality affects subsequent language classification.

C:

Clustering discovers groups based on feature similarity. That is not sufficient here: Speech input quality affects subsequent language classification.

D:

Agentic workloads combine reasoning and bounded external actions. That is not sufficient here: Speech input quality affects subsequent language classification.

Question 124

Troubleshooting report: The team sends an image URL to a text-only deployment. Which correction is MOST directly targeted to this failure?

  1. Select a classification output for image-wide categories
  2. Pin compatible deployment settings and maintain versioned evaluations
  3. Use a multimodal-capable deployment and supported message format
  4. Keep external content in lower-trust user/tool data fields

Explanation

Correct Answer

C:

This correction directly addresses the reported failure. Multimodal models can reason over more than one supported input modality. The targeted change is: Use a multimodal-capable deployment and supported message format.

Incorrect Answers

A:

Classification assigns image-wide categories rather than bounding-box coordinates. In the stated case, troubleshooting report: The team sends an image URL to a text-only deployment; “Select a classification output for image-wide categories” does not target that failure.

B:

Reproducibility depends on more than the visible prompt text. In the stated case, troubleshooting report: The team sends an image URL to a text-only deployment; “Pin compatible deployment settings and maintain versioned evaluations” does not target that failure.

D:

Instruction roles help distinguish application policy from user task text. In the stated case, troubleshooting report: The team sends an image URL to a text-only deployment; “Keep external content in lower-trust user/tool data fields” does not target that failure.

Question 125

Troubleshooting report: All text is attributed to the call-center agent. Which correction is MOST directly targeted to this failure?

  1. Apply a confirmation prompt and constrained recognition
  2. Use the correct audio encoding and recognized input format
  3. Enable supported diarization and validate turn assignments
  4. Keep operation outputs separate and map each to its intended field

Explanation

Correct Answer

C:

This correction directly addresses the reported failure. Speaker diarization separates turns and identities at a conversation level. The targeted change is: Enable supported diarization and validate turn assignments.

Incorrect Answers

A:

Speech recognition uncertainty is higher for confusable numeric strings. In the stated case, troubleshooting report: All text is attributed to the call-center agent; “Apply a confirmation prompt and constrained recognition” does not target that failure.

B:

Speech recognition requires the service to decode audio correctly. In the stated case, troubleshooting report: All text is attributed to the call-center agent; “Use the correct audio encoding and recognized input format” does not target that failure.

D:

Sentiment polarity and named entity spans answer different questions. In the stated case, troubleshooting report: All text is attributed to the call-center agent; “Keep operation outputs separate and map each to its intended field” does not target that failure.

Question 126

An Azure AI Fundamentals solution has this requirement: A customer service workflow needs to pass a case ID between agent steps. Which implementation most directly addresses it?

  1. Use typed state fields and validate identifiers at each tool boundary
  2. Check regional model availability and data handling requirements before deployment
  3. Check supported context and output limits for the chosen deployment
  4. Select and test a compatible deployed model through the Microsoft Foundry portal

Explanation

Correct Answer

A:

Structured state prevents subtle substitution of unrelated customer records. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

B:

Not every model or feature is available in every location. That is not sufficient here: Structured state prevents subtle substitution of unrelated customer records.

C:

Capabilities and limits vary by model deployment. That is not sufficient here: Structured state prevents subtle substitution of unrelated customer records.

D:

The portal provides model selection, deployment and test interfaces for supported resources. That is not sufficient here: Structured state prevents subtle substitution of unrelated customer records.

Question 127

The approved design action is: Choose a crop or resize preserving the requested aspect ratio and subject. Which test result would BEST establish that it worked?

  1. The crop keeps the subject in view at the specified aspect ratio
  2. Generated samples are checked for the required color and layout
  3. Selected candidates differ in composition while meeting required constraints
  4. The thumbnail preserves subject proportions and important content

Explanation

Correct Answer

D:

Smart cropping differs from simply stretching pixels. The expected acceptance evidence is: The thumbnail preserves subject proportions and important content. 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 thumbnail preserves subject proportions and important content.

B:

Explicit visual guidance can influence generated appearance but is not absolute. This observation does not establish the required result: The thumbnail preserves subject proportions and important content.

C:

Image-generation variability can be useful but does not guarantee suitability. This observation does not establish the required result: The thumbnail preserves subject proportions and important content.

Question 128

An Azure AI Fundamentals solution has this requirement: A marketing content generator represents generated endorsements as real customer quotations. Which implementation most directly addresses it?

  1. Enforce the monetary threshold in the backend and require human authorization above it
  2. Require governance evidence with reviewers, test outcomes, and documented exceptions
  3. Mark synthetic examples clearly and require verified consent for attributed statements
  4. Require independent transaction authorization and approval outside prompts

Explanation

Correct Answer

C:

Transparency prohibits presenting invented testimony as genuine. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

A prompt-level request cannot override a transaction policy. That is not sufficient here: Transparency prohibits presenting invented testimony as genuine.

B:

Self-attestation is not independent proof of regulatory controls. That is not sufficient here: Transparency prohibits presenting invented testimony as genuine.

D:

Accountable high-impact operations need verifiable authorized signatures. That is not sufficient here: Transparency prohibits presenting invented testimony as genuine.

Question 129

An Azure AI Fundamentals solution has this requirement: A deployed agent returns a generic answer but never invokes a required tool. Which implementation most directly addresses it?

  1. Write an instruction to identify missing inputs and request clarification before tools
  2. Inspect whether the tool is registered and appropriately described to the agent
  3. Use a versioned evaluation set with stable scoring and test inputs
  4. Limit context to necessary messages and apply safe summaries or retrieval

Explanation

Correct Answer

B:

Tools must be discoverable with valid schemas before selection. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Prompts can define required preconditions but tools must still enforce them. That is not sufficient here: Tools must be discoverable with valid schemas before selection.

C:

Anecdotal success on different prompts cannot establish an improvement. That is not sufficient here: Tools must be discoverable with valid schemas before selection.

D:

Long irrelevant context consumes tokens and may increase latency. That is not sufficient here: Tools must be discoverable with valid schemas before selection.

Question 130

Troubleshooting report: A user can change inventory by entering instructions into a prompt only. Which correction is MOST directly targeted to this failure?

  1. Enforce backend identity and method scopes
  2. Preserve speaker identities and confirm relevant transcript spans
  3. Interpret the actual image first and synthesize the verified description
  4. Provide the supported image input alongside the question

Explanation

Correct Answer

A:

This correction directly addresses the reported failure. Agentic workflows combine information access with controlled actions. The targeted change is: Enforce backend identity and method scopes.

Incorrect Answers

B:

Attribution distinguishes who committed to each action. In the stated case, troubleshooting report: A user can change inventory by entering instructions into a prompt only; “Preserve speaker identities and confirm relevant transcript spans” does not target that failure.

C:

Multimodal input can describe the graphic and speech output can read that description. In the stated case, troubleshooting report: A user can change inventory by entering instructions into a prompt only; “Interpret the actual image first and synthesize the verified description” does not target that failure.

D:

Image-and-text reasoning can connect a question with visible scene content. In the stated case, troubleshooting report: A user can change inventory by entering instructions into a prompt only; “Provide the supported image input alongside the question” does not target that failure.

Microsoft Azure AI Fundamentals (AI-901) Practice Tests