Practice Exams:

Microsoft AI-103 Dumps

EXAM: Developing AI Apps and Agents on Azure
Verified By Experts
Premium Bundle
$49.99

AI-103 Premium ETE File


  • Premium File 187 Questions & Answers
Whats Included:
  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates

Microsoft AI-103 Practice Test 3

Question 20

During a reliability exercise, a cloud developer must accept a change for embeddings for semantic retrieval. The issue to be addressed is: A search service must match paraphrases across internal manuals. Which concrete verification action demonstrates that the relevant behavior works?

  1. Create document and query embeddings using compatible embedding configurations
  2. Compare raw text strings alone while promising semantic vector retrieval
  3. Query paraphrases and verify relevant chunks rank in returned results
  4. Run representative spoken queries and measure transcription and playback

Explanation

Correct Answer

C:

To verify this change, the software consultancy should query paraphrases and verify relevant chunks rank in returned results. Operators still need the implementation to create document and query embeddings using compatible embedding configurations and the preventive requirement to version embedding deployments and reindex after incompatible changes.

Incorrect Answers

A:

The action to create document and query embeddings using compatible embedding configurations can contribute to embeddings for semantic retrieval as a implementation step; here, a search service must match paraphrases across internal manuals. In the software consultancy acceptance review, trying to create document and query embeddings using compatible embedding configurations does not test the exact embeddings for semantic retrieval condition where a search service must match paraphrases across internal manuals; the relevant test is to query paraphrases and verify relevant chunks rank in returned results.

B:

In the software consultancy acceptance review, trying to compare raw text strings alone while promising semantic vector retrieval does not test the exact embeddings for semantic retrieval condition where a search service must match paraphrases across internal manuals; the relevant test is to query paraphrases and verify relevant chunks rank in returned results. Acceptance should instead demonstrate that the service can query paraphrases and verify relevant chunks rank in returned results, using the described failure as the test criterion.

D:

For speech modality selection, it can be sensible to run representative spoken queries and measure transcription and playback if a voice support assistant must handle live speech and produce spoken answers. Without the actual change to query paraphrases and verify relevant chunks rank in returned results, the answer lacks the correct operational scope even if the alternate action is technically valid elsewhere.

Question 21

At a energy cooperative, the site reliability engineer is preparing a service audit. The concrete requirement is: A legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. Which design action BEST addresses the stated need?

  1. Combine keyword retrieval with vector recall and assess exact-identifier matches
  2. Ask for exact clause IDs and paraphrased topics and inspect ranking
  3. Choose a supported context window and retrieval strategy that fits the relevant evidence
  4. Use only nearest-neighbor vectors for regulatory section numbers without validating recall

Explanation

Correct Answer

A:

The proposed implementation is to combine keyword retrieval with vector recall and assess exact-identifier matches; this matches the operational requirement that a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. The team can ask for exact clause IDs and paraphrased topics and inspect ranking. The team should also measure both identifier recall and semantic relevance on a labeled set.

Incorrect Answers

B:

The action to ask for exact clause IDs and paraphrased topics and inspect ranking can contribute to full-text versus vector retrieval as a targeted acceptance check step; here, a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. Choosing to ask for exact clause IDs and paraphrased topics and inspect ranking cannot directly replace the specified action to combine keyword retrieval with vector recall and assess exact-identifier matches when a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary.

C:

Choosing to choose a supported context window and retrieval strategy that fits the relevant evidence would direct effort toward generative model context size, whose characteristic issue is that a document assistant truncates key context when submitting long cases. The relevant corrective or preventive action here remains to combine keyword retrieval with vector recall and assess exact-identifier matches, with its own validation criteria.

D:

Choosing to use only nearest-neighbor vectors for regulatory section numbers without validating recall cannot directly replace the specified action to combine keyword retrieval with vector recall and assess exact-identifier matches when a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. The actual implementation must combine keyword retrieval with vector recall and assess exact-identifier matches, because the stated service condition is that a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary.

Question 22

After a security tabletop exercise at a industrial maintenance team, the team observes this failure: A legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. Which flawed assumption or implementation choice would MOST DIRECTLY explain why the intended outcome was not achieved?

  1. Combine keyword retrieval with vector recall and assess exact-identifier matches
  2. Measure both identifier recall and semantic relevance on a labeled set
  3. Ask for exact clause IDs and paraphrased topics and inspect ranking
  4. Assume model pretraining incorporates every future company policy update
  5. Use only nearest-neighbor vectors for regulatory section numbers without validating recall

Explanation

Correct Answer

E:

This design mistake is to use only nearest-neighbor vectors for regulatory section numbers without validating recall; it conflicts with the actual full-text versus vector retrieval behavior needed when a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. The industrial maintenance team should instead combine keyword retrieval with vector recall and assess exact-identifier matches, followed by a check to ask for exact clause IDs and paraphrased topics and inspect ranking.

Incorrect Answers

A:

The alternative to combine keyword retrieval with vector recall and assess exact-identifier matches is a implementation step rather than the faulty assumption being diagnosed. This industrial maintenance team failure-analysis alternative of combine keyword retrieval with vector recall and assess exact-identifier matches does not diagnose the particular full-text versus vector retrieval defect where a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary; the keyed mistake is use only nearest-neighbor vectors for regulatory section numbers without validating recall.

B:

The alternative to measure both identifier recall and semantic relevance on a labeled set is a ongoing preventive rule step rather than the faulty assumption being diagnosed. This industrial maintenance team failure-analysis alternative of measure both identifier recall and semantic relevance on a labeled set does not diagnose the particular full-text versus vector retrieval defect where a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary; the keyed mistake is use only nearest-neighbor vectors for regulatory section numbers without validating recall.

C:

The alternative to ask for exact clause IDs and paraphrased topics and inspect ranking is a targeted acceptance check step rather than the faulty assumption being diagnosed. This industrial maintenance team failure-analysis alternative of ask for exact clause IDs and paraphrased topics and inspect ranking does not diagnose the particular full-text versus vector retrieval defect where a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary; the keyed mistake is use only nearest-neighbor vectors for regulatory section numbers without validating recall.

D:

This approach relates to grounding for policy agents: an agent answers internal policies that change each week. If the team chose it instead, the targeted check to ask for exact clause IDs and paraphrased topics and inspect ranking could still fail because the underlying required action is to use only nearest-neighbor vectors for regulatory section numbers without validating recall.

Question 23

The research university is responding to a support escalation. A agent designer discovers this issue: A legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. What corrective step should the team take to restore the intended functionality most directly?

  1. Add lexical matching for exact identifiers to the retrieval query
  2. Ask for exact clause IDs and paraphrased topics and inspect ranking
  3. Use only nearest-neighbor vectors for regulatory section numbers without validating recall
  4. Reselect an available deployment after checking the regional catalog and quota

Explanation

Correct Answer

A:

The direct repair is to add lexical matching for exact identifiers to the retrieval query. Once the condition is corrected, the research university should ask for exact clause IDs and paraphrased topics and inspect ranking. To reduce recurrence, it should also measure both identifier recall and semantic relevance on a labeled set; those steps provide different recovery and prevention functions.

Incorrect Answers

B:

The action to ask for exact clause IDs and paraphrased topics and inspect ranking can contribute to full-text versus vector retrieval as a targeted acceptance check step; here, a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. During research university remediation, the proposed step to ask for exact clause IDs and paraphrased topics and inspect ranking is not the direct restoration action for full-text versus vector retrieval when a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary; the keyed repair is add lexical matching for exact identifiers to the retrieval query.

C:

During research university remediation, the proposed step to use only nearest-neighbor vectors for regulatory section numbers without validating recall is not the direct restoration action for full-text versus vector retrieval when a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary; the keyed repair is add lexical matching for exact identifiers to the retrieval query. After remediation the team must ask for exact clause IDs and paraphrased topics and inspect ranking; the action in this option does not itself establish that target recovery outcome.

D:

During research university remediation, the proposed step to reselect an available deployment after checking the regional catalog and quota is not the direct restoration action for full-text versus vector retrieval when a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary; the keyed repair is add lexical matching for exact identifiers to the retrieval query. If the team chose it instead, the targeted check to ask for exact clause IDs and paraphrased topics and inspect ranking could still fail because the underlying required action is to add lexical matching for exact identifiers to the retrieval query.

Question 24

A responsible-AI reviewer at a education technology company wants the following issue to be prevented in later deployments: A legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. Which enduring control or engineering guardrail BEST prevents recurrence?

  1. Validate supported modalities and input limits before rollout
  2. Measure both identifier recall and semantic relevance on a labeled set
  3. Add lexical matching for exact identifiers to the retrieval query
  4. Use only nearest-neighbor vectors for regulatory section numbers without validating recall

Explanation

Correct Answer

B:

Standardizing the boundary to measure both identifier recall and semantic relevance on a labeled set is preventive because the same full-text versus vector retrieval failure could otherwise return: a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. For this education technology company, the service should also combine keyword retrieval with vector recall and assess exact-identifier matches, and tests should ask for exact clause IDs and paraphrased topics and inspect ranking so that the ongoing requirement is measurable.

Incorrect Answers

A:

The action to validate supported modalities and input limits before rollout addresses multimodal support selection rather than the decisive full-text versus vector retrieval condition. The education technology company should not infer that fixing one adjacent AI service validates its unrelated workflow; the requested action is to measure both identifier recall and semantic relevance on a labeled set.

C:

The action to add lexical matching for exact identifiers to the retrieval query can contribute to full-text versus vector retrieval as a corrective remediation step; here, a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. For prevention at the education technology company, the measure to add lexical matching for exact identifiers to the retrieval query is weaker or aimed at another issue than the full-text versus vector retrieval guardrail to measure both identifier recall and semantic relevance on a labeled set needed when a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary.

D:

For prevention at the education technology company, the measure to use only nearest-neighbor vectors for regulatory section numbers without validating recall is weaker or aimed at another issue than the full-text versus vector retrieval guardrail to measure both identifier recall and semantic relevance on a labeled set needed when a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. The lasting guardrail for this specific threat is to measure both identifier recall and semantic relevance on a labeled set, not simply to complete an adjacent service task.

Question 25

During a security tabletop exercise, a solution architect must accept a change for full-text versus vector retrieval. The issue to be addressed is: A legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. Which concrete verification action demonstrates that the relevant behavior works?

  1. Compare representative label accuracy and p95 response time against the baseline
  2. Ask for exact clause IDs and paraphrased topics and inspect ranking
  3. Use only nearest-neighbor vectors for regulatory section numbers without validating recall
  4. Combine keyword retrieval with vector recall and assess exact-identifier matches

Explanation

Correct Answer

B:

The targeted acceptance test is to ask for exact clause IDs and paraphrased topics and inspect ranking. The design should first combine keyword retrieval with vector recall and assess exact-identifier matches. A robust pharmaceutical research group deployment also needs to measure both identifier recall and semantic relevance on a labeled set; passing this check provides relevant evidence but does not certify every possible input or security risk.

Incorrect Answers

A:

For small-model latency tradeoff, it can be sensible to compare representative label accuracy and p95 response time against the baseline if an internal classification endpoint has strict CPU budget and bounded label space. Without the actual change to ask for exact clause IDs and paraphrased topics and inspect ranking, the answer lacks the correct operational scope even if the alternate action is technically valid elsewhere.

C:

In the pharmaceutical research group acceptance review, trying to use only nearest-neighbor vectors for regulatory section numbers without validating recall does not test the exact full-text versus vector retrieval condition where a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary; the relevant test is to ask for exact clause IDs and paraphrased topics and inspect ranking. Acceptance should instead demonstrate that the service can ask for exact clause IDs and paraphrased topics and inspect ranking, using the described failure as the test criterion.

D:

The action to combine keyword retrieval with vector recall and assess exact-identifier matches can contribute to full-text versus vector retrieval as a implementation step; here, a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary. In the pharmaceutical research group acceptance review, trying to combine keyword retrieval with vector recall and assess exact-identifier matches does not test the exact full-text versus vector retrieval condition where a legal reference chatbot must retrieve an exact regulation number even when semantic terms vary; the relevant test is to ask for exact clause IDs and paraphrased topics and inspect ranking.

Question 26

At a software consultancy, the machine-learning developer is preparing a design meeting. The concrete requirement is: A service assistant needs frequently changing knowledge across thousands of files. Which design action BEST addresses the stated need?

  1. Embed every document verbatim in a static system prompt indefinitely
  2. Select a multimodal-capable model that accepts both image and text inputs
  3. Use a maintained search index and retrieve relevant passages per request
  4. Update a source paragraph and confirm retrieval reflects the new revision

Explanation

Correct Answer

C:

Choosing to use a maintained search index and retrieve relevant passages per request directly targets the described managed search versus flat prompts gap: a service assistant needs frequently changing knowledge across thousands of files. As a lasting boundary, they also need to schedule source freshness checks and remove stale indexed copies, because an implementation by itself does not prove safe operation.

Incorrect Answers

A:

Choosing to embed every document verbatim in a static system prompt indefinitely cannot directly replace the specified action to use a maintained search index and retrieve relevant passages per request when a service assistant needs frequently changing knowledge across thousands of files. The actual implementation must use a maintained search index and retrieve relevant passages per request, because the stated service condition is that a service assistant needs frequently changing knowledge across thousands of files.

B:

Choosing to select a multimodal-capable model that accepts both image and text inputs would direct effort toward multimodal support selection, whose characteristic issue is that an intake app must reason jointly about photographs and textual descriptions. The relevant corrective or preventive action here remains to use a maintained search index and retrieve relevant passages per request, with its own validation criteria.

D:

The action to update a source paragraph and confirm retrieval reflects the new revision can contribute to managed search versus flat prompts as a targeted acceptance check step; here, a service assistant needs frequently changing knowledge across thousands of files. Choosing to update a source paragraph and confirm retrieval reflects the new revision cannot directly replace the specified action to use a maintained search index and retrieve relevant passages per request when a service assistant needs frequently changing knowledge across thousands of files.

Question 27

After a staged service rollout at a regional hospital, the team observes this failure: A service assistant needs frequently changing knowledge across thousands of files. Which flawed assumption or implementation choice would MOST DIRECTLY explain why the intended outcome was not achieved?

  1. Update a source paragraph and confirm retrieval reflects the new revision
  2. Treat a text-only model as though it can inspect image pixels
  3. Embed every document verbatim in a static system prompt indefinitely
  4. Use a maintained search index and retrieve relevant passages per request

Explanation

Correct Answer

C:

The faulty assumption is embed every document verbatim in a static system prompt indefinitely. A functioning alternative would be to use a maintained search index and retrieve relevant passages per request, and the relevant validation would be to update a source paragraph and confirm retrieval reflects the new revision.

Incorrect Answers

A:

The alternative to update a source paragraph and confirm retrieval reflects the new revision is a targeted acceptance check step rather than the faulty assumption being diagnosed. This regional hospital failure-analysis alternative of update a source paragraph and confirm retrieval reflects the new revision does not diagnose the particular managed search versus flat prompts defect where a service assistant needs frequently changing knowledge across thousands of files; the keyed mistake is embed every document verbatim in a static system prompt indefinitely.

B:

The option to treat a text-only model as though it can inspect image pixels would normally serve a different multimodal support selection requirement, where an intake app must reason jointly about photographs and textual descriptions. The alternative may be useful in its own context, but it does not directly supply embed every document verbatim in a static system prompt indefinitely for the stated decision.

D:

The alternative to use a maintained search index and retrieve relevant passages per request is a implementation step rather than the faulty assumption being diagnosed. This regional hospital failure-analysis alternative of use a maintained search index and retrieve relevant passages per request does not diagnose the particular managed search versus flat prompts defect where a service assistant needs frequently changing knowledge across thousands of files; the keyed mistake is embed every document verbatim in a static system prompt indefinitely.

Question 28

The insurance platform is responding to a release acceptance review. A quality assurance lead discovers this issue: A service assistant needs frequently changing knowledge across thousands of files. What corrective step should the team take to restore the intended functionality most directly?

  1. Ingest current documents into the index and reconnect the application retriever
  2. Embed every document verbatim in a static system prompt indefinitely
  3. Re-embed mismatched documents and queries under a consistent representation
  4. Update a source paragraph and confirm retrieval reflects the new revision

Explanation

Correct Answer

A:

The appropriate remediation is to ingest current documents into the index and reconnect the application retriever, because the failure concerns managed search versus flat prompts: a service assistant needs frequently changing knowledge across thousands of files. A meaningful verification should update a source paragraph and confirm retrieval reflects the new revision, and the durable operating rule should schedule source freshness checks and remove stale indexed copies after the immediate fix.

Incorrect Answers

B:

During insurance platform remediation, the proposed step to embed every document verbatim in a static system prompt indefinitely is not the direct restoration action for managed search versus flat prompts when a service assistant needs frequently changing knowledge across thousands of files; the keyed repair is ingest current documents into the index and reconnect the application retriever. After remediation the team must update a source paragraph and confirm retrieval reflects the new revision; the action in this option does not itself establish that target recovery outcome.

C:

This approach relates to embeddings for semantic retrieval: a search service must match paraphrases across internal manuals. If the team chose it instead, the targeted check to update a source paragraph and confirm retrieval reflects the new revision could still fail because the underlying required action is to ingest current documents into the index and reconnect the application retriever.

D:

The action to update a source paragraph and confirm retrieval reflects the new revision can contribute to managed search versus flat prompts as a targeted acceptance check step; here, a service assistant needs frequently changing knowledge across thousands of files. During insurance platform remediation, the proposed step to update a source paragraph and confirm retrieval reflects the new revision is not the direct restoration action for managed search versus flat prompts when a service assistant needs frequently changing knowledge across thousands of files; the keyed repair is ingest current documents into the index and reconnect the application retriever.

Question 29

A AI platform engineer at a media platform wants the following issue to be prevented in later deployments: A service assistant needs frequently changing knowledge across thousands of files. Which enduring control or engineering guardrail BEST prevents recurrence?

  1. Ingest current documents into the index and reconnect the application retriever
  2. Schedule source freshness checks and remove stale indexed copies
  3. Embed every document verbatim in a static system prompt indefinitely
  4. Monitor token budgets and require tests near documented limits

Explanation

Correct Answer

B:

The sustained prevention measure is to schedule source freshness checks and remove stale indexed copies. The implementation can still use the separate action to use a maintained search index and retrieve relevant passages per request, but a guardrail constrains future changes. The media platform can measure whether the rule works by arranging to update a source paragraph and confirm retrieval reflects the new revision.

Incorrect Answers

A:

The action to ingest current documents into the index and reconnect the application retriever can contribute to managed search versus flat prompts as a corrective remediation step; here, a service assistant needs frequently changing knowledge across thousands of files. For prevention at the media platform, the measure to ingest current documents into the index and reconnect the application retriever is weaker or aimed at another issue than the managed search versus flat prompts guardrail to schedule source freshness checks and remove stale indexed copies needed when a service assistant needs frequently changing knowledge across thousands of files.

C:

For prevention at the media platform, the measure to embed every document verbatim in a static system prompt indefinitely is weaker or aimed at another issue than the managed search versus flat prompts guardrail to schedule source freshness checks and remove stale indexed copies needed when a service assistant needs frequently changing knowledge across thousands of files. The lasting guardrail for this specific threat is to schedule source freshness checks and remove stale indexed copies, not simply to complete an adjacent service task.

D:

The action to monitor token budgets and require tests near documented limits addresses generative model context size rather than the decisive managed search versus flat prompts condition. The media platform should not infer that fixing one adjacent AI service validates its unrelated workflow; the requested action is to schedule source freshness checks and remove stale indexed copies.

Question 30

During a staged service rollout, a compliance specialist must accept a change for managed search versus flat prompts. The issue to be addressed is: A service assistant needs frequently changing knowledge across thousands of files. Which concrete verification action demonstrates that the relevant behavior works?

  1. Update a source paragraph and confirm retrieval reflects the new revision
  2. Use a maintained search index and retrieve relevant passages per request
  3. Embed every document verbatim in a static system prompt indefinitely
  4. Change one policy and confirm a new answer cites the updated source

Explanation

Correct Answer

A:

To verify this change, the customer support center should update a source paragraph and confirm retrieval reflects the new revision. Operators still need the implementation to use a maintained search index and retrieve relevant passages per request and the preventive requirement to schedule source freshness checks and remove stale indexed copies.

Incorrect Answers

B:

The action to use a maintained search index and retrieve relevant passages per request can contribute to managed search versus flat prompts as a implementation step; here, a service assistant needs frequently changing knowledge across thousands of files. In the customer support center acceptance review, trying to use a maintained search index and retrieve relevant passages per request does not test the exact managed search versus flat prompts condition where a service assistant needs frequently changing knowledge across thousands of files; the relevant test is to update a source paragraph and confirm retrieval reflects the new revision.

C:

In the customer support center acceptance review, trying to embed every document verbatim in a static system prompt indefinitely does not test the exact managed search versus flat prompts condition where a service assistant needs frequently changing knowledge across thousands of files; the relevant test is to update a source paragraph and confirm retrieval reflects the new revision. Acceptance should instead demonstrate that the service can update a source paragraph and confirm retrieval reflects the new revision, using the described failure as the test criterion.

D:

For grounding for policy agents, it can be sensible to change one policy and confirm a new answer cites the updated source if an agent answers internal policies that change each week. Without the actual change to update a source paragraph and confirm retrieval reflects the new revision, the answer lacks the correct operational scope even if the alternate action is technically valid elsewhere.

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