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 27

Question 260

Troubleshooting report: The output template forbids any uncertainty language. Which correction is MOST directly targeted to this failure?

  1. Permit calibrated caveats and qualified review paths
  2. Reconcile the provider transaction ID and actual status before confirmation
  3. Separate retrieved data from tool and system instruction channels
  4. Assess both unsafe misses and harmful overblocking

Explanation

Correct Answer

A:

This correction directly addresses the reported failure. Transparency reduces mistaken reliance on unsupported professional guarantees. The targeted change is: Permit calibrated caveats and qualified review paths.

Incorrect Answers

B:

Generated text is not evidence of external side effects. In the stated case, troubleshooting report: The output template forbids any uncertainty language; “Reconcile the provider transaction ID and actual status before confirmation” does not target that failure.

C:

Lower-trust text is not a privileged command channel. In the stated case, troubleshooting report: The output template forbids any uncertainty language; “Separate retrieved data from tool and system instruction channels” does not target that failure.

D:

Safety controls should not overblock permitted assistance. In the stated case, troubleshooting report: The output template forbids any uncertainty language; “Assess both unsafe misses and harmful overblocking” does not target that failure.

Question 261

An Azure AI solution uses the following approved control: Use image editing or segmentation with object-mask preservation. What observable result should pass QA?

  1. Bilingual review confirms equivalent meaning and terminology
  2. The model returns the sign category and any required location
  3. The product shape remains unchanged against the new background
  4. The summary preserves the topic sequence and supported key points

Explanation

Correct Answer

C:

Editing changes selected pixels while leaving product regions intact. The expected acceptance evidence is: The product shape remains unchanged against the new background. This observation directly tests the stated behavior.

Incorrect Answers

A:

Translation transfers meaning across languages rather than changing topic labels. This observation does not establish the required result: The product shape remains unchanged against the new background.

B:

The task chooses a label, with localization needed only when multiple signs appear. This observation does not establish the required result: The product shape remains unchanged against the new background.

D:

Meeting summaries require temporal and spoken source content. This observation does not establish the required result: The product shape remains unchanged against the new background.

Question 262

An AI operations team is testing two distinct capabilities. Failure (1): Only the final prose is logged with no tool execution details. Failure (2): The model forwards raw private records because the tool asked for them. Which TWO specific corrections should the team apply?

  1. Validate case identity and entitlement before each tool call
  2. Filter data at the tool boundary using approved policies
  3. Create and test a simple one-agent tool workflow
  4. Provision proper identity and resource role bindings for the target tenant
  5. Enable privacy-minimized structured agent tracing

Explanation

Correct Answers

B:

Correct for failure (2). Tool descriptions cannot override data-loss-prevention boundaries. It independently corrects the second defect.

E:

Correct for failure (1). Operational observability aids debugging without exposing hidden sensitive prompt data. The proposed correction directly addresses that reported defect.

Incorrect Answers

A:

Structured state prevents subtle substitution of unrelated customer records. The stated pair of failures instead requires “Filter data at the tool boundary using approved policies” and “Enable privacy-minimized structured agent tracing”; this option cannot substitute for both.

C:

A single agent can orchestrate a bounded tool without a multi-agent system. The stated pair of failures instead requires “Filter data at the tool boundary using approved policies” and “Enable privacy-minimized structured agent tracing”; this option cannot substitute for both.

D:

Copying code does not grant access to another tenant's resources. The stated pair of failures instead requires “Filter data at the tool boundary using approved policies” and “Enable privacy-minimized structured agent tracing”; this option cannot substitute for both.

Question 263

A project needs the following AI behavior: A support agent must identify several different objects within a damaged equipment photo. What should the designer implement?

  1. Use object detection that returns labels and bounding boxes
  2. Generate and edit concise factual alt text describing significant content
  3. Provide the actual image to a capable multimodal or OCR pipeline
  4. Select appropriate separate visual features for OCR, tagging and people detection

Explanation

Correct Answer

A:

Object detection reports which objects appear and where they occur. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

B:

Accessible captions should communicate image purpose without invented details. That is not sufficient here: Object detection reports which objects appear and where they occur.

C:

Visual evidence is needed to read sign content. That is not sufficient here: Object detection reports which objects appear and where they occur.

D:

Different visual outputs describe text, scene classes and detected people. That is not sufficient here: Object detection reports which objects appear and where they occur.

Question 264

Troubleshooting report: The app selects a voice ID that is not supported in that locale. Which correction is MOST directly targeted to this failure?

  1. Choose key phrase extraction and evaluate relevance
  2. Ground the summary and penalize unsupported additions
  3. Add a no-speech branch and clarification option
  4. Use a compatible voice from the service's voice catalog

Explanation

Correct Answer

D:

This correction directly addresses the reported failure. Not every voice is supported for all languages. The targeted change is: Use a compatible voice from the service's voice catalog.

Incorrect Answers

A:

Key phrases highlight important concepts expressed in the original text. In the stated case, troubleshooting report: The app selects a voice ID that is not supported in that locale; “Choose key phrase extraction and evaluate relevance” does not target that failure.

B:

Summarization compresses written information rather than changing its facts. In the stated case, troubleshooting report: The app selects a voice ID that is not supported in that locale; “Ground the summary and penalize unsupported additions” does not target that failure.

C:

An empty recognition result is not semantically equivalent to the word no. In the stated case, troubleshooting report: The app selects a voice ID that is not supported in that locale; “Add a no-speech branch and clarification option” does not target that failure.

Question 265

An Azure AI solution uses the following approved control: Normalize image orientation before applying vision detection. What observable result should pass QA?

  1. A drawn overlay aligns each box with the corresponding person
  2. The same object is recognized consistently across supported rotations
  3. The returned description matches the selected upper-left object
  4. The thumbnail preserves subject proportions and important content

Explanation

Correct Answer

B:

Rotated inputs can degrade detection or coordinate interpretation. The expected acceptance evidence is: The same object is recognized consistently across supported rotations. This observation directly tests the stated behavior.

Incorrect Answers

A:

Localization metadata is tied to pixel regions, not just labels. This observation does not establish the required result: The same object is recognized consistently across supported rotations.

C:

Specific regions can carry details missed by a whole-image caption. This observation does not establish the required result: The same object is recognized consistently across supported rotations.

D:

Smart cropping differs from simply stretching pixels. This observation does not establish the required result: The same object is recognized consistently across supported rotations.

Question 266

An Azure-based AI team must satisfy this objective: An agent request includes a parameter not accepted by its registered tool. Which design choice is best?

  1. Write an instruction to identify missing inputs and request clarification before tools
  2. Use representative test prompts and inspect execution and tool traces
  3. Insert relevant source passages and require evidence-linked answers
  4. Align the tool argument names and types with the published function schema

Explanation

Correct Answer

D:

Invalid tool parameters break calls even if the intended operation is correct. 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: Invalid tool parameters break calls even if the intended operation is correct.

B:

An answer that sounds plausible may conceal failed tool calls. That is not sufficient here: Invalid tool parameters break calls even if the intended operation is correct.

C:

Grounding constrains the answer to the available text and provenance. That is not sufficient here: Invalid tool parameters break calls even if the intended operation is correct.

Question 267

Troubleshooting report: The app summarizes the entire clip from only the first frame. Which correction is MOST directly targeted to this failure?

  1. Review multiple timestamped frames and flag rapid or unobserved steps instead of inferring them
  2. Improve the product mask and restrict the editable region
  3. Switch to object detection for localized results
  4. Include the earlier reference image and compare grounded features

Explanation

Correct Answer

A:

This correction directly addresses the reported failure. A single still cannot convey action order; sampled video frames may also omit fleeting intermediate steps. The targeted change is: Review multiple timestamped frames and flag rapid or unobserved steps instead of inferring them.

Incorrect Answers

B:

A mask defines which image area may change. In the stated case, troubleshooting report: The app summarizes the entire clip from only the first frame; “Improve the product mask and restrict the editable region” does not target that failure.

C:

Classification predicts image-level categories while detection provides positions. In the stated case, troubleshooting report: The app summarizes the entire clip from only the first frame; “Switch to object detection for localized results” does not target that failure.

D:

Change detection needs both temporal states for a meaningful comparison. In the stated case, troubleshooting report: The app summarizes the entire clip from only the first frame; “Include the earlier reference image and compare grounded features” does not target that failure.

Question 268

Troubleshooting report: The app stores uploaded IDs in a public static image folder. Which correction is MOST directly targeted to this failure?

  1. Map the task to a whole-image classifier
  2. Use detection instead of whole-image classification
  3. Move sensitive images to authenticated private storage
  4. Ground the alt text in the actual picture

Explanation

Correct Answer

C:

This correction directly addresses the reported failure. Images can carry personally identifying information beyond text fields. The targeted change is: Move sensitive images to authenticated private storage.

Incorrect Answers

A:

Image-level classification is distinct from object localization. In the stated case, troubleshooting report: The app stores uploaded IDs in a public static image folder; “Map the task to a whole-image classifier” does not target that failure.

B:

Object detection reports which objects appear and where they occur. In the stated case, troubleshooting report: The app stores uploaded IDs in a public static image folder; “Use detection instead of whole-image classification” does not target that failure.

D:

Accessible captions should communicate image purpose without invented details. In the stated case, troubleshooting report: The app stores uploaded IDs in a public static image folder; “Ground the alt text in the actual picture” does not target that failure.

Question 269

An engineering review documents this requirement: A customer-facing avatar is mistaken for a real employee. Select the most appropriate solution.

  1. Include a verifiable source link and section reference alongside material policy answers
  2. Identify the virtual agent as synthetic and avoid deceptive identity cues
  3. Maintain a governed feature inventory and trace decisions to policy
  4. Include rare but consequential safety edge cases in acceptance tests

Explanation

Correct Answer

B:

Transparency prevents users from falsely assuming they are speaking with a human. This implementation addresses the precise goal rather than an adjacent AI task.

Incorrect Answers

A:

Transparency lets users inspect evidence instead of accepting fluent text. That is not sufficient here: Transparency prevents users from falsely assuming they are speaking with a human.

C:

Accountability requires interpretable reasons for consequential decisions. That is not sufficient here: Transparency prevents users from falsely assuming they are speaking with a human.

D:

A high overall score can conceal severe failure on unusual inputs. That is not sufficient here: Transparency prevents users from falsely assuming they are speaking with a human.

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.

Microsoft Azure AI Fundamentals (AI-901) Practice Tests