Microsoft AI-103 (Developing AI Apps and Agents on Azure) Practice Test 33

Question 320

A cloud engineering team needs its AI workflow to produce campaign images in a required landscape layout instead of arbitrary dimensions. What technical approach most directly enables that?

  1. Use text-to-image generation for a new visual
  2. Generate a clip from a scripted prompt
  3. Control output aspect ratio and resolution
  4. Use reference-guided image generation

Explanation

Correct Answer

C:

For the stated requirement to produce campaign images in a required landscape layout instead of arbitrary dimensions, supported generation size and aspect-ratio controls affect composition and delivery fit; post-crop alone can cut out important content; this is the mechanism that the system must implement.

Incorrect Answers

A:

For a system whose objective is to create an original illustration from a written scene description without an existing source picture, text-to-image generation synthesizes imagery from prompts; it is suited to new compositions rather than extracting details from stored images; this does not directly meet the required outcome to produce campaign images in a required landscape layout instead of arbitrary dimensions.

B:

For a system whose objective is to create a short visual sequence from a written shot description without any source footage, text-to-video generation produces moving visual content; it is distinct from captioning or summarizing an existing video; this does not directly meet the required outcome to produce campaign images in a required landscape layout instead of arbitrary dimensions.

D:

For a system whose objective is to create visual variants that retain the composition or style cues of an approved reference image, reference-guided generation conditions the output on supplied media as supported by the deployed model, instead of starting from only text; this does not directly meet the required outcome to produce campaign images in a required landscape layout instead of arbitrary dimensions.

Question 321

An Azure AI workload passes basic health probes but has this issue: A generated training clip changes the design of a machine from the reference photograph. Which fix targets it?

  1. Plan discrete shot prompts, generate controlled segments, and assemble the approved sequence
  2. Provide an authorized visual reference through a supported image-generation workflow
  3. Use a supported video-generation model with a time-bounded shot prompt
  4. Supply the authorized reference media through a compatible video-generation workflow

Explanation

Correct Answer

D:

The fault signal is that a generated training clip changes the design of a machine from the reference photograph; the targeted fix works because reference-conditioned video generation uses supplied visual cues when the selected model supports them, rather than relying on words alone, restoring the ability to preserve the recognizable setting of an approved reference while generating a moving scene.

Incorrect Answers

A:

If the failing function were to produce a coherent sequence when a single complex prompt repeatedly loses event ordering, this change could be relevant because dividing the planned clip into constrained shots improves control over actions, transitions and reference continuity, subject to model tooling; the actual symptom is that a generated training clip changes the design of a machine from the reference photograph, which calls for a different repair.

B:

If the failing function were to create visual variants that retain the composition or style cues of an approved reference image, this change could be relevant because reference-guided generation conditions the output on supplied media as supported by the deployed model, instead of starting from only text; the actual symptom is that a generated training clip changes the design of a machine from the reference photograph, which calls for a different repair.

C:

If the failing function were to create a short visual sequence from a written shot description without any source footage, this change could be relevant because text-to-video generation produces moving visual content; it is distinct from captioning or summarizing an existing video; the actual symptom is that a generated training clip changes the design of a machine from the reference photograph, which calls for a different repair.

Question 322

An engineer receives this incident report: A creative team tries to produce new concept art using a captioning endpoint and receives only text descriptions. What should be changed?

  1. Use a supported video-generation model with a time-bounded shot prompt
  2. Call a supported image-generation deployment with a constrained visual prompt
  3. Provide an authorized visual reference through a supported image-generation workflow
  4. Select a supported landscape size or aspect control and compose the prompt for that frame

Explanation

Correct Answer

B:

The fault signal is that a creative team tries to produce new concept art using a captioning endpoint and receives only text descriptions; the targeted fix works because text-to-image generation synthesizes imagery from prompts; it is suited to new compositions rather than extracting details from stored images, restoring the ability to create an original illustration from a written scene description without an existing source picture.

Incorrect Answers

A:

If the failing function were to create a short visual sequence from a written shot description without any source footage, this change could be relevant because text-to-video generation produces moving visual content; it is distinct from captioning or summarizing an existing video; the actual symptom is that a creative team tries to produce new concept art using a captioning endpoint and receives only text descriptions, which calls for a different repair.

C:

If the failing function were to create visual variants that retain the composition or style cues of an approved reference image, this change could be relevant because reference-guided generation conditions the output on supplied media as supported by the deployed model, instead of starting from only text; the actual symptom is that a creative team tries to produce new concept art using a captioning endpoint and receives only text descriptions, which calls for a different repair.

D:

If the failing function were to produce campaign images in a required landscape layout instead of arbitrary dimensions, this change could be relevant because supported generation size and aspect-ratio controls affect composition and delivery fit; post-crop alone can cut out important content; the actual symptom is that a creative team tries to produce new concept art using a captioning endpoint and receives only text descriptions, which calls for a different repair.

Question 323

An operator can reproduce this error: A content team asks a video-analysis tool for a new animated demonstration and receives metadata about existing clips. Which targeted repair is justified?

  1. Supply the authorized reference media through a compatible video-generation workflow
  2. Use a supported video-generation model with a time-bounded shot prompt
  3. Define a mask for the poster region and run a supported inpainting edit
  4. Plan discrete shot prompts, generate controlled segments, and assemble the approved sequence

Explanation

Correct Answer

B:

The fault signal is that a content team asks a video-analysis tool for a new animated demonstration and receives metadata about existing clips; the targeted fix works because text-to-video generation produces moving visual content; it is distinct from captioning or summarizing an existing video, restoring the ability to create a short visual sequence from a written shot description without any source footage.

Incorrect Answers

A:

If the failing function were to preserve the recognizable setting of an approved reference while generating a moving scene, this change could be relevant because reference-conditioned video generation uses supplied visual cues when the selected model supports them, rather than relying on words alone; the actual symptom is that a content team asks a video-analysis tool for a new animated demonstration and receives metadata about existing clips, which calls for a different repair.

C:

If the failing function were to replace an unwanted object in a photo while preserving the surrounding approved scene, this change could be relevant because mask-based image editing limits requested changes to selected regions, while untouched content remains the reference context; the actual symptom is that a content team asks a video-analysis tool for a new animated demonstration and receives metadata about existing clips, which calls for a different repair.

D:

If the failing function were to produce a coherent sequence when a single complex prompt repeatedly loses event ordering, this change could be relevant because dividing the planned clip into constrained shots improves control over actions, transitions and reference continuity, subject to model tooling; the actual symptom is that a content team asks a video-analysis tool for a new animated demonstration and receives metadata about existing clips, which calls for a different repair.

Question 324

A product team needs its AI workflow to preserve the recognizable setting of an approved reference while generating a moving scene. What technical approach most directly enables that?

  1. Use targeted video editing for an existing clip
  2. Condition generated video on reference media
  3. Generate a clip from a scripted prompt
  4. Use reference-guided image generation

Explanation

Correct Answer

B:

For the stated requirement to preserve the recognizable setting of an approved reference while generating a moving scene, reference-conditioned video generation uses supplied visual cues when the selected model supports them, rather than relying on words alone; this is the mechanism that the system must implement.

Incorrect Answers

A:

For a system whose objective is to change one visual element within a recorded scene without recreating the whole sequence, a video edit workflow acts on existing media with the model's supported modification controls, instead of generating an unrelated new clip; this does not directly meet the required outcome to preserve the recognizable setting of an approved reference while generating a moving scene.

C:

For a system whose objective is to create a short visual sequence from a written shot description without any source footage, text-to-video generation produces moving visual content; it is distinct from captioning or summarizing an existing video; this does not directly meet the required outcome to preserve the recognizable setting of an approved reference while generating a moving scene.

D:

For a system whose objective is to create visual variants that retain the composition or style cues of an approved reference image, reference-guided generation conditions the output on supplied media as supported by the deployed model, instead of starting from only text; this does not directly meet the required outcome to preserve the recognizable setting of an approved reference while generating a moving scene.

Question 325

The planned deployment includes a mechanism to generate a clip from a scripted prompt. Which outcome should the QA team measure?

  1. Measure delivered pixel dimensions and proportion of preserved key objects
  2. Compare unedited regions, motion continuity and the specific edited element across frames
  3. Score reference fidelity and temporal consistency across the produced clip
  4. Review motion continuity, prompt adherence, and output safety across generated frames

Explanation

Correct Answer

D:

The accepted change is to generate a clip from a scripted prompt; this evaluation checks whether it can create a short visual sequence from a written shot description without any source footage, since text-to-video generation produces moving visual content; it is distinct from captioning or summarizing an existing video.

Incorrect Answers

A:

This test concerns the ability to produce campaign images in a required landscape layout instead of arbitrary dimensions, for which supported generation size and aspect-ratio controls affect composition and delivery fit; post-crop alone can cut out important content; it does not establish whether the change to generate a clip from a scripted prompt works for the current release.

B:

This test concerns the ability to change one visual element within a recorded scene without recreating the whole sequence, for which a video edit workflow acts on existing media with the model's supported modification controls, instead of generating an unrelated new clip; it does not establish whether the change to generate a clip from a scripted prompt works for the current release.

C:

This test concerns the ability to preserve the recognizable setting of an approved reference while generating a moving scene, for which reference-conditioned video generation uses supplied visual cues when the selected model supports them, rather than relying on words alone; it does not establish whether the change to generate a clip from a scripted prompt works for the current release.

Question 326

Before approving the proposed action to use confidence and source evidence for review, which measurable acceptance check should be required?

  1. Validate extracted field presence and accuracy against labeled inspection images
  2. Measure false approvals and manual-review precision by confidence band
  3. Test schema compliance and classification confusion on representative product images
  4. Measure localization precision and intersection-over-union against annotated regions

Explanation

Correct Answer

B:

The accepted change is to use confidence and source evidence for review; this evaluation checks whether it can route uncertain visual-extraction results to an operator instead of silently accepting them, since where supported, confidence estimates and source references enable risk-based human review rather than assuming every extracted field is correct.

Incorrect Answers

A:

This test concerns the ability to extract selected visual attributes into a repeatable structured schema for an image workflow, for which an analyzer defines content type, desired fields and models, producing structured evidence for downstream processing; it does not establish whether the change to use confidence and source evidence for review works for the current release.

C:

This test concerns the ability to produce a consistent category and identifier from product photos for a downstream system, for which a declared analyzer schema channels model output into specified field names and types; consumers should still validate values; it does not establish whether the change to use confidence and source evidence for review works for the current release.

D:

This test concerns the ability to locate components spatially instead of giving only an image-wide caption, for which region-oriented detection provides object locations or structured references, depending on the model/tool; broad captions do not supply coordinates; it does not establish whether the change to use confidence and source evidence for review works for the current release.

Question 327

Production telemetry reveals this failure: Ad hoc prompts return different JSON keys for every photographed inspection label. Which change most directly corrects it?

  1. Use an appropriate object-localization pipeline with region output and coordinate validation
  2. Configure a reusable image analyzer with explicitly defined output fields
  3. Enable appropriate temporal segmentation and index segment-level descriptions
  4. Record field evidence and confidence signals, then review low-reliability outputs

Explanation

Correct Answer

B:

The fault signal is that ad hoc prompts return different JSON keys for every photographed inspection label; the targeted fix works because an analyzer defines content type, desired fields and models, producing structured evidence for downstream processing, restoring the ability to extract selected visual attributes into a repeatable structured schema for an image workflow.

Incorrect Answers

A:

If the failing function were to locate components spatially instead of giving only an image-wide caption, this change could be relevant because region-oriented detection provides object locations or structured references, depending on the model/tool; broad captions do not supply coordinates; the actual symptom is that ad hoc prompts return different JSON keys for every photographed inspection label, which calls for a different repair.

C:

If the failing function were to classify chapters of a multi-topic video into retrievable units, this change could be relevant because video analyzers can emit scene segments, timestamps and structured or Markdown results that improve retrieval granularity; the actual symptom is that ad hoc prompts return different JSON keys for every photographed inspection label, which calls for a different repair.

D:

If the failing function were to route uncertain visual-extraction results to an operator instead of silently accepting them, this change could be relevant because where supported, confidence estimates and source references enable risk-based human review rather than assuming every extracted field is correct; the actual symptom is that ad hoc prompts return different JSON keys for every photographed inspection label, which calls for a different repair.

Question 328

The acceptance criteria for a Foundry-based application require it to describe the meaningful function and visible content of an interface image for screen-reader users. Which design should be approved?

  1. Generate accessibility-focused alt text
  2. Generate concise captions for an image collection
  3. Use a Content Understanding image analyzer
  4. Compare before-and-after equipment photos using visual evidence

Explanation

Correct Answer

A:

For the stated requirement to describe the meaningful function and visible content of an interface image for screen-reader users, alt text prioritizes purpose and essential information, not decorative speculation or excessive narration; human review matters for accessibility; this is the mechanism that the system must implement.

Incorrect Answers

B:

For a system whose objective is to create short searchable descriptions of individual images without long interpretation, captioning summarizes salient visual content for discovery; detailed descriptions and object coordinates are different output requirements; this does not directly meet the required outcome to describe the meaningful function and visible content of an interface image for screen-reader users.

C:

For a system whose objective is to extract selected visual attributes into a repeatable structured schema for an image workflow, an analyzer defines content type, desired fields and models, producing structured evidence for downstream processing; this does not directly meet the required outcome to describe the meaningful function and visible content of an interface image for screen-reader users.

D:

For a system whose objective is to identify changed switch positions between a baseline photograph and a later inspection photograph, a multimodal request can include both clearly labeled images and ask for observed differences; single-image captions do not establish what changed across time; this does not directly meet the required outcome to describe the meaningful function and visible content of an interface image for screen-reader users.

Question 329

A solution architect is deciding how to explain a demonstration whose narration and visible operations carry complementary facts. Which Azure AI pattern addresses that task?

  1. Analyze a specific video time segment
  2. Extract searchable key-frame descriptions
  3. Generate accessibility-focused alt text
  4. Combine speech transcript with video frames

Explanation

Correct Answer

D:

For the stated requirement to explain a demonstration whose narration and visible operations carry complementary facts, multimodal video understanding aligns spoken words with visual events; ignoring either modality can lose essential meaning; this is the mechanism that the system must implement.

Incorrect Answers

A:

For a system whose objective is to answer what action occurred between defined timestamps in a long maintenance recording, temporal segmentation narrows the evidence to the relevant interval and reduces confusion between similar events; this does not directly meet the required outcome to explain a demonstration whose narration and visible operations carry complementary facts.

B:

For a system whose objective is to make a long training video discoverable by scenes and visible equipment, frame or scene metadata indexes visual evidence that ordinary speech transcripts miss; sampling cadence should match the content; this does not directly meet the required outcome to explain a demonstration whose narration and visible operations carry complementary facts.

C:

For a system whose objective is to describe the meaningful function and visible content of an interface image for screen-reader users, alt text prioritizes purpose and essential information, not decorative speculation or excessive narration; human review matters for accessibility; this does not directly meet the required outcome to explain a demonstration whose narration and visible operations carry complementary facts.

Question 330

What test should engineers run after they compare before-and-after equipment photos using visual evidence, rather than relying solely on a successful API response?

  1. Measure false approvals and manual-review precision by confidence band
  2. Check action accuracy and timestamp alignment against annotated segments
  3. Score correct changes, missed changes, and invented differences against annotated image pairs
  4. Measure caption factuality and search usefulness on human-labeled sample images
  5. Review whether users can understand the image's essential message without seeing it

Explanation

Correct Answer

C:

The accepted change is to compare before-and-after equipment photos using visual evidence; this evaluation checks whether it can identify changed switch positions between a baseline photograph and a later inspection photograph, since a multimodal request can include both clearly labeled images and ask for observed differences; single-image captions do not establish what changed across time.

Incorrect Answers

A:

This test concerns the ability to route uncertain visual-extraction results to an operator instead of silently accepting them, for which where supported, confidence estimates and source references enable risk-based human review rather than assuming every extracted field is correct; it does not establish whether the change to compare before-and-after equipment photos using visual evidence works for the current release.

B:

This test concerns the ability to answer what action occurred between defined timestamps in a long maintenance recording, for which temporal segmentation narrows the evidence to the relevant interval and reduces confusion between similar events; it does not establish whether the change to compare before-and-after equipment photos using visual evidence works for the current release.

D:

This test concerns the ability to create short searchable descriptions of individual images without long interpretation, for which captioning summarizes salient visual content for discovery; detailed descriptions and object coordinates are different output requirements; it does not establish whether the change to compare before-and-after equipment photos using visual evidence works for the current release.

E:

This test concerns the ability to describe the meaningful function and visible content of an interface image for screen-reader users, for which alt text prioritizes purpose and essential information, not decorative speculation or excessive narration; human review matters for accessibility; it does not establish whether the change to compare before-and-after equipment photos using visual evidence works for the current release.

Developing AI Apps and Agents on Azure (AI-103) Practice Tests