Microsoft AI-103 Practice Test 75 - Prepaway
Practice Exams:

Pass Microsoft AI-103 Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers!

30 Days Free Updates, Instant Download!

Verified By Experts
Premium File
$89.99 $98.99

AI-103 Premium ETE File


  • Premium File 187 Questions & Answers. Last Update: Nov 15, 2026
Whats Included:
  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates

Microsoft AI-103 Practice Test 75

Question 740

Situation: A retailer processes multilingual refund requests containing local currency formats and missing account fields. Which TWO proposed practices create material safety, correctness, or reliability risks? Select TWO answers.

  1. Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests
  2. Fill missing customer account IDs with plausible values inferred from sentiment
  3. Inspect records that explain the affected process, then check the result described in this case: A retailer processes multilingual refund requests containing local currency formats and missing account fields
  4. Use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text
  5. Use the document sentiment label as a substitute for exact named entities

Explanation

Correct Answer

B:

A retailer processes multilingual refund requests containing local currency formats and missing account fields. “Fill missing customer account IDs with plausible values inferred from sentiment” would violate the requirement. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

E:

A retailer processes multilingual refund requests containing local currency formats and missing account fields. “Use the document sentiment label as a substitute for exact named entities” would violate the requirement. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

Incorrect Answers

A:

A retailer processes multilingual refund requests containing local currency formats and missing account fields. “Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests” is the required outcome of a working control, not an action that weakens it. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Confirm that entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests.

C:

“Inspect records that explain the affected process, then check the result described in this case: A retailer processes multilingual refund requests containing local currency formats” adds a relevant safeguard or verification step rather than weakening the solution. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

D:

“Use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text” is the controlled implementation, not an unsafe workaround. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

Question 741

Architecture case: A compliance team translates contractual notices and must preserve negation and stated dates in structured output. What configuration should the engineer choose for this workload?

  1. Use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text
  2. Apply the proposed setting to all connected workloads before testing the case-specific boundary: a compliance team translates contractual notices and must preserve negation and stated dates in structured output
  3. Treat all currencies and dates as the same locale
  4. Translate legal notices without testing locale-specific negation and date phrasing

Explanation

Correct Answer

A:

Because a compliance team translates contractual notices and must preserve negation and stated dates in structured output, sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests.

Incorrect Answers

B:

“Apply the proposed setting to all connected workloads before testing the case-specific boundary: a compliance team translates contractual notices and must preserve negation” does not implement the control that the facts call for. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Instead, use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text.

C:

A compliance team translates contractual notices and must preserve negation and stated dates in structured output. “Treat all currencies and dates as the same locale” does not implement the control that the facts call for. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Instead, use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text.

D:

A compliance team translates contractual notices and must preserve negation and stated dates in structured output. “Translate legal notices without testing locale-specific negation and date phrasing” would not meet the requirement. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Instead, use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text.

Question 742

Production requirement: A compliance team translates contractual notices and must preserve negation and stated dates in structured output. What technical fact prevents the simplest shortcut from being reliable?

  1. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts
  2. The most powerful or least restricted model is always the safest choice regardless of task conditions
  3. The fact that the team started work is enough to prove correctness for this exact case: a compliance team translates contractual notices and must preserve negation and stated dates in structured output
  4. Successful provisioning by itself proves all security, output, and operational constraints are satisfied

Explanation

Correct Answer

A:

Because a compliance team translates contractual notices and must preserve negation and stated dates in structured output, sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text.

Incorrect Answers

B:

A compliance team translates contractual notices and must preserve negation and stated dates in structured output. “The most powerful or least restricted model is always the safest choice regardless of task conditions” does not fully explain the decisive technical boundary in this case. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. The implementation must use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text.

C:

“The fact that the team started work is enough to prove correctness for this exact case: a compliance team translates contractual notices and must preserve negation and stated dates in structured output” does not fully explain the decisive technical boundary in this case. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. The implementation must use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text.

D:

“Successful provisioning by itself proves all security, output, and operational constraints are satisfied” does not fully explain the decisive technical boundary in this case. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. The implementation must use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text.

Question 743

A team is investigating this task: A compliance team translates contractual notices and must preserve negation and stated dates in structured output. What end-to-end test would distinguish successful operation from an installed configuration?

  1. Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests
  2. The team closes a change ticket without replaying a representative positive and negative test
  3. A project dashboard remains green, although the affected request, output and permissions were not checked
  4. The operator reports success without reproducing the observed behavior: a compliance team translates contractual notices and must preserve negation and stated dates in structured output

Explanation

Correct Answer

A:

A compliance team translates contractual notices and must preserve negation and stated dates in structured output. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests.

Incorrect Answers

B:

Because a compliance team translates contractual notices and must preserve negation and stated dates in structured output, “The team closes a change ticket without replaying a representative positive and negative test” alone does not establish all the required behavior. Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

C:

Because a compliance team translates contractual notices and must preserve negation and stated dates in structured output, “A project dashboard remains green, although the affected request, output and permissions were not checked” would not meet the requirement. Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

D:

“The operator reports success without reproducing the observed behavior: a compliance team translates contractual notices and must preserve negation and stated dates in structured output” alone does not establish all the required behavior. Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

Question 744

Situation: A compliance team translates contractual notices and must preserve negation and stated dates in structured output. Which diagnostic investigation best isolates the issue before another change?

  1. Compare input language detection, token/character spans, JSON schema failures, locale formats and human-reviewed translation examples
  2. Read only the broad service-health summary and skip actual workload evidence
  3. Restart every affected component before collecting request or tool traces
  4. Assume the general service is at fault without checking the affected operation in this case: a compliance team translates contractual notices and must preserve negation and stated dates in structured output

Explanation

Correct Answer

A:

A compliance team translates contractual notices and must preserve negation and stated dates in structured output. Compare input language detection, token/character spans, JSON schema failures, locale formats and human-reviewed translation examples. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests.

Incorrect Answers

B:

A compliance team translates contractual notices and must preserve negation and stated dates in structured output. “Read only the broad service-health summary and skip actual workload evidence” would not meet the requirement. Compare input language detection, token/character spans, JSON schema failures, locale formats and human-reviewed translation examples. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

C:

“Restart every affected component before collecting request or tool traces” would not be the most direct first check for this failure. Compare input language detection, token/character spans, JSON schema failures, locale formats and human-reviewed translation examples.

D:

“Assume the general service is at fault without checking the affected operation in this case: a compliance team translates contractual notices and must preserve negation” would not meet the requirement. Compare input language detection, token/character spans, JSON schema failures, locale formats and human-reviewed translation examples. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

Question 745

Architecture case: A compliance team translates contractual notices and must preserve negation and stated dates in structured output. Which TWO shortcuts should be rejected for this scenario? Select TWO answers.

  1. Fill missing customer account IDs with plausible values inferred from sentiment
  2. Inspect records that explain the affected process, then check the result described in this case: A compliance team translates contractual notices and must preserve negation and stated dates in structured output
  3. Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests
  4. Use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text
  5. Use the document sentiment label as a substitute for exact named entities

Explanation

Correct Answer

A:

A compliance team translates contractual notices and must preserve negation and stated dates in structured output. “Fill missing customer account IDs with plausible values inferred from sentiment” would violate the requirement. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

E:

A compliance team translates contractual notices and must preserve negation and stated dates in structured output. “Use the document sentiment label as a substitute for exact named entities” would violate the requirement. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

Incorrect Answers

B:

“Inspect records that explain the affected process, then check the result described in this case: A compliance team translates contractual notices and must preserve negation” adds a relevant safeguard or verification step rather than weakening the solution. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts.

C:

“Entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests” is the required outcome of a working control, not an action that weakens it. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Confirm that entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests.

D:

A compliance team translates contractual notices and must preserve negation and stated dates in structured output. “Use task-specific structured extraction with schema validation, language-aware evaluation and explicit handling of missing or ambiguous text” is the controlled implementation, not an unsafe workaround. Sentiment, named-entity extraction and translation serve different information goals; ambiguous source text needs uncertainty rather than invented structured facts. Confirm that entity fields retain the correct spans and locale, missing values are null, and bilingual tests preserve the meaning of the original requests.

Question 746

Production requirement: A call center requires accurate multilingual call transcripts with speaker labels for quality review. Which next change most directly addresses the reported technical risk?

  1. Apply the proposed setting to all connected workloads before testing the case-specific boundary: a call center requires accurate multilingual call transcripts with speaker labels for quality review
  2. Assume the first speaker’s language applies to every caller in a multilingual meeting
  3. Use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task
  4. Publish raw confidential recordings as model training examples

Explanation

Correct Answer

C:

Use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases.

Incorrect Answers

A:

“Apply the proposed setting to all connected workloads before testing the case-specific boundary: a call center requires accurate multilingual call transcripts with speaker labels for quality review” does not implement the control that the facts call for. Instead, use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task.

B:

“Assume the first speaker’s language applies to every caller in a multilingual meeting” would not meet the requirement. Instead, use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task.

D:

“Publish raw confidential recordings as model training examples” does not implement the control that the facts call for. Instead, use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task.

Question 747

A team is investigating this task: A call center requires accurate multilingual call transcripts with speaker labels for quality review. Which explanation best captures the key behavior of this architecture?

  1. The fact that the team started work is enough to prove correctness for this exact case: a call center requires accurate multilingual call transcripts with speaker labels for quality review
  2. Successful provisioning by itself proves all security, output, and operational constraints are satisfied
  3. Audio transcription and speech synthesis have different accuracy tests; diarization and time-aligned words cannot be inferred reliably from plain transcript text
  4. The most powerful or least restricted model is always the safest choice regardless of task conditions

Explanation

Correct Answer

C:

Audio transcription and speech synthesis have different accuracy tests; diarization and time-aligned words cannot be inferred reliably from plain transcript text. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases.

Incorrect Answers

A:

“The fact that the team started work is enough to prove correctness for this exact case: a call center requires accurate multilingual call transcripts with speaker labels for quality review” would not be the most direct first check for this failure. Audio transcription and speech synthesis have different accuracy tests; diarization and time-aligned words cannot be inferred reliably from plain transcript text.

B:

“Successful provisioning by itself proves all security, output, and operational constraints are satisfied” would not be the most direct first check for this failure. Audio transcription and speech synthesis have different accuracy tests; diarization and time-aligned words cannot be inferred reliably from plain transcript text.

D:

“The most powerful or least restricted model is always the safest choice regardless of task conditions” would not be the most direct first check for this failure. Audio transcription and speech synthesis have different accuracy tests; diarization and time-aligned words cannot be inferred reliably from plain transcript text.

Question 748

Situation: A call center requires accurate multilingual call transcripts with speaker labels for quality review. Which result would demonstrate that the proposed remedy works end to end?

  1. The operator reports success without reproducing the observed behavior: a call center requires accurate multilingual call transcripts with speaker labels for quality review
  2. A project dashboard remains green, although the affected request, output and permissions were not checked
  3. The team closes a change ticket without replaying a representative positive and negative test
  4. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases

Explanation

Correct Answer

D:

A call center requires accurate multilingual call transcripts with speaker labels for quality review. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases. Use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task.

Incorrect Answers

A:

“The operator reports success without reproducing the observed behavior: a call center requires accurate multilingual call transcripts with speaker labels for quality review” alone does not establish all the required behavior. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases.

B:

Because a call center requires accurate multilingual call transcripts with speaker labels for quality review, “A project dashboard remains green, although the affected request, output and permissions were not checked” would not meet the requirement. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases.

C:

“The team closes a change ticket without replaying a representative positive and negative test” alone does not establish all the required behavior. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases.

Question 749

Architecture case: A call center requires accurate multilingual call transcripts with speaker labels for quality review. What evidence should the engineer correlate to verify the actual failure cause?

  1. Inspect sample rates, channel mapping, selected language, word timestamps, speaker turns, model errors and redacted audio/log retention
  2. Assume the general service is at fault without checking the affected operation in this case: a call center requires accurate multilingual call transcripts with speaker labels for quality review
  3. Read only the broad service-health summary and skip actual workload evidence
  4. Restart every affected component before collecting request or tool traces

Explanation

Correct Answer

A:

Because a call center requires accurate multilingual call transcripts with speaker labels for quality review, words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases. Use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task.

Incorrect Answers

B:

“Assume the general service is at fault without checking the affected operation in this case: a call center requires accurate multilingual call transcripts with speaker labels for quality review” would not meet the requirement. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases.

C:

“Read only the broad service-health summary and skip actual workload evidence” would not meet the requirement. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases.

D:

“Restart every affected component before collecting request or tool traces” alone does not establish all the required behavior. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases.

Question 750

Production requirement: A call center requires accurate multilingual call transcripts with speaker labels for quality review. Which TWO decisions would undermine the required system boundary? Select TWO answers.

  1. Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases
  2. Inspect records that explain the affected process, then check the result described in this case: A call center requires accurate multilingual call transcripts with speaker labels for quality review
  3. Use text translation output as proof of who spoke each sentence in an audio recording
  4. Use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task
  5. Delete all timing markers before a task that requires subtitle synchronization

Explanation

Correct Answer

C:

“Use text translation output as proof of who spoke each sentence in an audio recording” would violate the requirement. The appropriate control instead is to use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task.

E:

“Delete all timing markers before a task that requires subtitle synchronization” would violate the requirement. The appropriate control instead is to use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task.

Incorrect Answers

A:

“Words, speakers and event timestamps match checked audio samples, and synthetic speech reproduces authorized text without leaking restricted phrases” is the required outcome of a working control, not an action that weakens it.

B:

“Inspect records that explain the affected process, then check the result described in this case: A call center requires accurate multilingual call transcripts with speaker labels for quality review” adds a relevant safeguard or verification step rather than weakening the solution.

D:

“Use appropriate speech-to-text or text-to-speech components with language detection, timing, speaker separation and privacy controls for the task” is the controlled implementation, not an unsafe workaround.

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