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 40

Question 390

During a operational readiness review, a machine-learning developer must accept a change for multimodal document question answering. The issue to be addressed is: An assistant must reason about a plotted chart and its textual caption. Which concrete verification action demonstrates that the relevant behavior works?

  1. Ask a chart trend question and compare the answer to the actual plot
  2. Use a supported multimodal model and preserve image-text grounding together
  3. Ingest a test vector and run a matching nearest-neighbor query
  4. Extract text alone and promise accurate chart interpretation without vision inputs

Explanation

Correct Answer

A:

To verify this change, the nonprofit organization should ask a chart trend question and compare the answer to the actual plot. Operators still need the implementation to use a supported multimodal model and preserve image-text grounding together and the preventive requirement to require visual evidence citations and accessible descriptions.

Incorrect Answers

B:

The action to use a supported multimodal model and preserve image-text grounding together can contribute to multimodal document question answering as a implementation step; here, an assistant must reason about a plotted chart and its textual caption. In the nonprofit organization acceptance review, trying to use a supported multimodal model and preserve image-text grounding together does not test the exact multimodal document question answering condition where an assistant must reason about a plotted chart and its textual caption; the relevant test is to ask a chart trend question and compare the answer to the actual plot.

C:

For vector index dimensions, it can be sensible to ingest a test vector and run a matching nearest-neighbor query if a search index rejects vectors after the embedding model changes. Without the actual change to ask a chart trend question and compare the answer to the actual plot, the answer lacks the correct operational scope even if the alternate action is technically valid elsewhere.

D:

In the nonprofit organization acceptance review, trying to extract text alone and promise accurate chart interpretation without vision inputs does not test the exact multimodal document question answering condition where an assistant must reason about a plotted chart and its textual caption; the relevant test is to ask a chart trend question and compare the answer to the actual plot. Acceptance should instead demonstrate that the service can ask a chart trend question and compare the answer to the actual plot, using the described failure as the test criterion.

Question 391

At a media platform, the AI application owner is preparing a cost review. The concrete requirement is: A deployed RAG app gives stale product terms after an upstream update. Which design action BEST addresses the stated need?

  1. Monitor ingestion timestamps and refresh embeddings and index data as needed
  2. Combine exact keyword and vector retrieval and evaluate ranking quality
  3. Reindex modified items and invalidate stale retrieval cache entries
  4. Assume an old vector index updates whenever source documents change
  5. Modify a test source and verify searchable revision and answer citations
  6. Enforce maximum source age for time-sensitive answers

Explanation

Correct Answer

A:

The proposed implementation is to monitor ingestion timestamps and refresh embeddings and index data as needed; this matches the operational requirement that a deployed RAG app gives stale product terms after an upstream update. The team can modify a test source and verify searchable revision and answer citations. The team should also enforce maximum source age for time-sensitive answers.

Incorrect Answers

B:

This approach relates to hybrid retrieval for identifiers: users search for part numbers and semantically related troubleshooting symptoms. If the team chose it instead, the targeted check to modify a test source and verify searchable revision and answer citations could still fail because the underlying required action is to monitor ingestion timestamps and refresh embeddings and index data as needed.

C:

The action to reindex modified items and invalidate stale retrieval cache entries can contribute to knowledge-connector freshness as a corrective remediation step; here, a deployed RAG app gives stale product terms after an upstream update. Choosing to reindex modified items and invalidate stale retrieval cache entries cannot directly replace the specified action to monitor ingestion timestamps and refresh embeddings and index data as needed when a deployed RAG app gives stale product terms after an upstream update.

D:

Choosing to assume an old vector index updates whenever source documents change cannot directly replace the specified action to monitor ingestion timestamps and refresh embeddings and index data as needed when a deployed RAG app gives stale product terms after an upstream update. The actual implementation must monitor ingestion timestamps and refresh embeddings and index data as needed, because the stated service condition is that a deployed RAG app gives stale product terms after an upstream update.

E:

The action to modify a test source and verify searchable revision and answer citations can contribute to knowledge-connector freshness as a targeted acceptance check step; here, a deployed RAG app gives stale product terms after an upstream update. Choosing to modify a test source and verify searchable revision and answer citations cannot directly replace the specified action to monitor ingestion timestamps and refresh embeddings and index data as needed when a deployed RAG app gives stale product terms after an upstream update.

F:

The action to enforce maximum source age for time-sensitive answers can contribute to knowledge-connector freshness as a ongoing preventive rule step; here, a deployed RAG app gives stale product terms after an upstream update. Choosing to enforce maximum source age for time-sensitive answers cannot directly replace the specified action to monitor ingestion timestamps and refresh embeddings and index data as needed when a deployed RAG app gives stale product terms after an upstream update.

Question 392

After a reliability exercise at a customer support center, the team observes this failure: A deployed RAG app gives stale product terms after an upstream update. Which flawed assumption or implementation choice would MOST DIRECTLY explain why the intended outcome was not achieved?

  1. Assume an old vector index updates whenever source documents change
  2. Extract text alone and promise accurate chart interpretation without vision inputs
  3. Modify a test source and verify searchable revision and answer citations
  4. Monitor ingestion timestamps and refresh embeddings and index data as needed

Explanation

Correct Answer

A:

This design mistake is to assume an old vector index updates whenever source documents change; it conflicts with the actual knowledge-connector freshness behavior needed when a deployed RAG app gives stale product terms after an upstream update. The customer support center should instead monitor ingestion timestamps and refresh embeddings and index data as needed, followed by a check to modify a test source and verify searchable revision and answer citations.

Incorrect Answers

B:

This customer support center failure-analysis alternative of extract text alone and promise accurate chart interpretation without vision inputs does not diagnose the particular knowledge-connector freshness defect where a deployed RAG app gives stale product terms after an upstream update; the keyed mistake is assume an old vector index updates whenever source documents change. The alternative may be useful in its own context, but it does not directly supply assume an old vector index updates whenever source documents change for the stated decision.

C:

The alternative to modify a test source and verify searchable revision and answer citations is a targeted acceptance check step rather than the faulty assumption being diagnosed. This customer support center failure-analysis alternative of modify a test source and verify searchable revision and answer citations does not diagnose the particular knowledge-connector freshness defect where a deployed RAG app gives stale product terms after an upstream update; the keyed mistake is assume an old vector index updates whenever source documents change.

D:

The alternative to monitor ingestion timestamps and refresh embeddings and index data as needed is a implementation step rather than the faulty assumption being diagnosed. This customer support center failure-analysis alternative of monitor ingestion timestamps and refresh embeddings and index data as needed does not diagnose the particular knowledge-connector freshness defect where a deployed RAG app gives stale product terms after an upstream update; the keyed mistake is assume an old vector index updates whenever source documents change.

Question 393

The manufacturing group is responding to a source update. A AI platform engineer discovers this issue: A deployed RAG app gives stale product terms after an upstream update. What corrective step should the team take to restore the intended functionality most directly?

  1. Rebuild a contextual retrieval query from permitted prior turns
  2. Modify a test source and verify searchable revision and answer citations
  3. Reindex modified items and invalidate stale retrieval cache entries
  4. Assume an old vector index updates whenever source documents change

Explanation

Correct Answer

C:

The direct repair is to reindex modified items and invalidate stale retrieval cache entries. Once the condition is corrected, the manufacturing group should modify a test source and verify searchable revision and answer citations. To reduce recurrence, it should also enforce maximum source age for time-sensitive answers; those steps provide different recovery and prevention functions.

Incorrect Answers

A:

This approach relates to multi-turn rag grounding: a user asks a follow-up whose meaning depends on prior context. If the team chose it instead, the targeted check to modify a test source and verify searchable revision and answer citations could still fail because the underlying required action is to reindex modified items and invalidate stale retrieval cache entries.

B:

The action to modify a test source and verify searchable revision and answer citations can contribute to knowledge-connector freshness as a targeted acceptance check step; here, a deployed RAG app gives stale product terms after an upstream update. During manufacturing group remediation, the proposed step to modify a test source and verify searchable revision and answer citations is not the direct restoration action for knowledge-connector freshness when a deployed RAG app gives stale product terms after an upstream update; the keyed repair is reindex modified items and invalidate stale retrieval cache entries.

D:

During manufacturing group remediation, the proposed step to assume an old vector index updates whenever source documents change is not the direct restoration action for knowledge-connector freshness when a deployed RAG app gives stale product terms after an upstream update; the keyed repair is reindex modified items and invalidate stale retrieval cache entries. After remediation the team must modify a test source and verify searchable revision and answer citations; the action in this option does not itself establish that target recovery outcome.

Question 394

A compliance specialist at a legal operations unit wants the following issue to be prevented in later deployments: A deployed RAG app gives stale product terms after an upstream update. Which enduring control or engineering guardrail BEST prevents recurrence?

  1. Preserve tenant session boundaries and limit unnecessary history
  2. Assume an old vector index updates whenever source documents change
  3. Enforce maximum source age for time-sensitive answers
  4. Reindex modified items and invalidate stale retrieval cache entries

Explanation

Correct Answer

C:

Standardizing the boundary to enforce maximum source age for time-sensitive answers is preventive because the same knowledge-connector freshness failure could otherwise return: a deployed RAG app gives stale product terms after an upstream update. For this legal operations unit, the service should also monitor ingestion timestamps and refresh embeddings and index data as needed, and tests should modify a test source and verify searchable revision and answer citations so that the ongoing requirement is measurable.

Incorrect Answers

A:

The action to preserve tenant session boundaries and limit unnecessary history addresses multi-turn rag grounding rather than the decisive knowledge-connector freshness condition. The legal operations unit should not infer that fixing one adjacent AI service validates its unrelated workflow; the requested action is to enforce maximum source age for time-sensitive answers.

B:

For prevention at the legal operations unit, the measure to assume an old vector index updates whenever source documents change is weaker or aimed at another issue than the knowledge-connector freshness guardrail to enforce maximum source age for time-sensitive answers needed when a deployed RAG app gives stale product terms after an upstream update. The lasting guardrail for this specific threat is to enforce maximum source age for time-sensitive answers, not simply to complete an adjacent service task.

D:

The action to reindex modified items and invalidate stale retrieval cache entries can contribute to knowledge-connector freshness as a corrective remediation step; here, a deployed RAG app gives stale product terms after an upstream update. For prevention at the legal operations unit, the measure to reindex modified items and invalidate stale retrieval cache entries is weaker or aimed at another issue than the knowledge-connector freshness guardrail to enforce maximum source age for time-sensitive answers needed when a deployed RAG app gives stale product terms after an upstream update.

Question 395

During a reliability exercise, a product support lead must accept a change for knowledge-connector freshness. The issue to be addressed is: A deployed RAG app gives stale product terms after an upstream update. Which concrete verification action demonstrates that the relevant behavior works?

  1. Submit missing-field and malformed-input cases and confirm predictable handling
  2. Assume an old vector index updates whenever source documents change
  3. Modify a test source and verify searchable revision and answer citations
  4. Monitor ingestion timestamps and refresh embeddings and index data as needed

Explanation

Correct Answer

C:

The targeted acceptance test is to modify a test source and verify searchable revision and answer citations. The design should first monitor ingestion timestamps and refresh embeddings and index data as needed. A robust financial services division deployment also needs to enforce maximum source age for time-sensitive answers; passing this check provides relevant evidence but does not certify every possible input or security risk.

Incorrect Answers

A:

For structured json generation, it can be sensible to submit missing-field and malformed-input cases and confirm predictable handling if a claims workflow requires parseable fields rather than an unconstrained narrative. Without the actual change to modify a test source and verify searchable revision and answer citations, the answer lacks the correct operational scope even if the alternate action is technically valid elsewhere.

B:

In the financial services division acceptance review, trying to assume an old vector index updates whenever source documents change does not test the exact knowledge-connector freshness condition where a deployed RAG app gives stale product terms after an upstream update; the relevant test is to modify a test source and verify searchable revision and answer citations. Acceptance should instead demonstrate that the service can modify a test source and verify searchable revision and answer citations, using the described failure as the test criterion.

D:

The action to monitor ingestion timestamps and refresh embeddings and index data as needed can contribute to knowledge-connector freshness as a implementation step; here, a deployed RAG app gives stale product terms after an upstream update. In the financial services division acceptance review, trying to monitor ingestion timestamps and refresh embeddings and index data as needed does not test the exact knowledge-connector freshness condition where a deployed RAG app gives stale product terms after an upstream update; the relevant test is to modify a test source and verify searchable revision and answer citations.

Question 396

At a nonprofit organization, the data engineer is preparing a service audit. The concrete requirement is: A search-backed agent reveals documents outside the user's team. Which design action BEST addresses the stated need?

  1. Retrieve from all teams and rely on model instructions to hide restricted details
  2. Apply verified per-user filters or secured retrieval identity before synthesis
  3. Run identical queries as two users with different access grants
  4. Deploy a compatible Foundry conversational model and connect the application to its endpoint
  5. Test deny conditions and enforce ACL trimming at retrieval

Explanation

Correct Answer

B:

Choosing to apply verified per-user filters or secured retrieval identity before synthesis directly targets the described rag permission filtering gap: a search-backed agent reveals documents outside the user's team. As a lasting boundary, they also need to test deny conditions and enforce ACL trimming at retrieval, because an implementation by itself does not prove safe operation.

Incorrect Answers

A:

Choosing to retrieve from all teams and rely on model instructions to hide restricted details cannot directly replace the specified action to apply verified per-user filters or secured retrieval identity before synthesis when a search-backed agent reveals documents outside the user's team. The actual implementation must apply verified per-user filters or secured retrieval identity before synthesis, because the stated service condition is that a search-backed agent reveals documents outside the user's team.

C:

The action to run identical queries as two users with different access grants can contribute to rag permission filtering as a targeted acceptance check step; here, a search-backed agent reveals documents outside the user's team. Choosing to run identical queries as two users with different access grants cannot directly replace the specified action to apply verified per-user filters or secured retrieval identity before synthesis when a search-backed agent reveals documents outside the user's team.

D:

The option to deploy a compatible Foundry conversational model and connect the application to its endpoint would normally serve a different model deployment for chat requirement, where a customer portal has no model endpoint capable of supporting conversational completion. The alternative may be useful in its own context, but it does not directly supply apply verified per-user filters or secured retrieval identity before synthesis for the stated decision.

E:

The action to test deny conditions and enforce ACL trimming at retrieval can contribute to rag permission filtering as a ongoing preventive rule step; here, a search-backed agent reveals documents outside the user's team. Choosing to test deny conditions and enforce ACL trimming at retrieval cannot directly replace the specified action to apply verified per-user filters or secured retrieval identity before synthesis when a search-backed agent reveals documents outside the user's team.

Question 397

After a security tabletop exercise at a municipal services team, the team observes this failure: A search-backed agent reveals documents outside the user's team. Which flawed assumption or implementation choice would MOST DIRECTLY explain why the intended outcome was not achieved?

  1. Retrieve from all teams and rely on model instructions to hide restricted details
  2. Assume projects share model deployment names and resource identity interchangeably
  3. Apply verified per-user filters or secured retrieval identity before synthesis
  4. Run identical queries as two users with different access grants

Explanation

Correct Answer

A:

The faulty assumption is retrieve from all teams and rely on model instructions to hide restricted details. A functioning alternative would be to apply verified per-user filters or secured retrieval identity before synthesis, and the relevant validation would be to run identical queries as two users with different access grants.

Incorrect Answers

B:

The option to assume projects share model deployment names and resource identity interchangeably would normally serve a different model connection project scope requirement, where a Python client points to an old Foundry project and silently uses wrong deployment. The alternative may be useful in its own context, but it does not directly supply retrieve from all teams and rely on model instructions to hide restricted details for the stated decision.

C:

The alternative to apply verified per-user filters or secured retrieval identity before synthesis is a implementation step rather than the faulty assumption being diagnosed. This municipal services team failure-analysis alternative of apply verified per-user filters or secured retrieval identity before synthesis does not diagnose the particular rag permission filtering defect where a search-backed agent reveals documents outside the user's team; the keyed mistake is retrieve from all teams and rely on model instructions to hide restricted details.

D:

The alternative to run identical queries as two users with different access grants is a targeted acceptance check step rather than the faulty assumption being diagnosed. This municipal services team failure-analysis alternative of run identical queries as two users with different access grants does not diagnose the particular rag permission filtering defect where a search-backed agent reveals documents outside the user's team; the keyed mistake is retrieve from all teams and rely on model instructions to hide restricted details.

Question 398

The payment services team is responding to a support escalation. A cloud developer discovers this issue: A search-backed agent reveals documents outside the user's team. What corrective step should the team take to restore the intended functionality most directly?

  1. Correct indexing and authorization filters and purge leaked cached snippets
  2. Add lexical matching and rerank merged relevant results
  3. Retrieve from all teams and rely on model instructions to hide restricted details
  4. Run identical queries as two users with different access grants

Explanation

Correct Answer

A:

The appropriate remediation is to correct indexing and authorization filters and purge leaked cached snippets, because the failure concerns rag permission filtering: a search-backed agent reveals documents outside the user's team. A meaningful verification should run identical queries as two users with different access grants, and the durable operating rule should test deny conditions and enforce ACL trimming at retrieval after the immediate fix.

Incorrect Answers

B:

This approach relates to hybrid retrieval for identifiers: users search for part numbers and semantically related troubleshooting symptoms. If the team chose it instead, the targeted check to run identical queries as two users with different access grants could still fail because the underlying required action is to correct indexing and authorization filters and purge leaked cached snippets.

C:

During payment services team remediation, the proposed step to retrieve from all teams and rely on model instructions to hide restricted details is not the direct restoration action for rag permission filtering when a search-backed agent reveals documents outside the user's team; the keyed repair is correct indexing and authorization filters and purge leaked cached snippets. After remediation the team must run identical queries as two users with different access grants; the action in this option does not itself establish that target recovery outcome.

D:

The action to run identical queries as two users with different access grants can contribute to rag permission filtering as a targeted acceptance check step; here, a search-backed agent reveals documents outside the user's team. During payment services team remediation, the proposed step to run identical queries as two users with different access grants is not the direct restoration action for rag permission filtering when a search-backed agent reveals documents outside the user's team; the keyed repair is correct indexing and authorization filters and purge leaked cached snippets.

Question 399

A site reliability engineer at a transport authority wants the following issue to be prevented in later deployments: A search-backed agent reveals documents outside the user's team. Which enduring control or engineering guardrail BEST prevents recurrence?

  1. Retrieve from all teams and rely on model instructions to hide restricted details
  2. Correct indexing and authorization filters and purge leaked cached snippets
  3. Test deny conditions and enforce ACL trimming at retrieval
  4. Preserve tenant session boundaries and limit unnecessary history

Explanation

Correct Answer

C:

The sustained prevention measure is to test deny conditions and enforce ACL trimming at retrieval. The implementation can still use the separate action to apply verified per-user filters or secured retrieval identity before synthesis, but a guardrail constrains future changes. The transport authority can measure whether the rule works by arranging to run identical queries as two users with different access grants.

Incorrect Answers

A:

For prevention at the transport authority, the measure to retrieve from all teams and rely on model instructions to hide restricted details is weaker or aimed at another issue than the rag permission filtering guardrail to test deny conditions and enforce ACL trimming at retrieval needed when a search-backed agent reveals documents outside the user's team. The lasting guardrail for this specific threat is to test deny conditions and enforce ACL trimming at retrieval, not simply to complete an adjacent service task.

B:

The action to correct indexing and authorization filters and purge leaked cached snippets can contribute to rag permission filtering as a corrective remediation step; here, a search-backed agent reveals documents outside the user's team. For prevention at the transport authority, the measure to correct indexing and authorization filters and purge leaked cached snippets is weaker or aimed at another issue than the rag permission filtering guardrail to test deny conditions and enforce ACL trimming at retrieval needed when a search-backed agent reveals documents outside the user's team.

D:

The action to preserve tenant session boundaries and limit unnecessary history addresses multi-turn rag grounding rather than the decisive rag permission filtering condition. The transport authority should not infer that fixing one adjacent AI service validates its unrelated workflow; the requested action is to test deny conditions and enforce ACL trimming at retrieval.

Question 400

During a security tabletop exercise, a security engineer must accept a change for rag permission filtering. The issue to be addressed is: A search-backed agent reveals documents outside the user's team. Which concrete verification action demonstrates that the relevant behavior works?

  1. Apply verified per-user filters or secured retrieval identity before synthesis
  2. Retrieve from all teams and rely on model instructions to hide restricted details
  3. Run identical queries as two users with different access grants
  4. Compare labeled query relevance before and after semantic reranking

Explanation

Correct Answer

C:

To verify this change, the large online retailer should run identical queries as two users with different access grants. Operators still need the implementation to apply verified per-user filters or secured retrieval identity before synthesis and the preventive requirement to test deny conditions and enforce ACL trimming at retrieval.

Incorrect Answers

A:

The action to apply verified per-user filters or secured retrieval identity before synthesis can contribute to rag permission filtering as a implementation step; here, a search-backed agent reveals documents outside the user's team. In the large online retailer acceptance review, trying to apply verified per-user filters or secured retrieval identity before synthesis does not test the exact rag permission filtering condition where a search-backed agent reveals documents outside the user's team; the relevant test is to run identical queries as two users with different access grants.

B:

In the large online retailer acceptance review, trying to retrieve from all teams and rely on model instructions to hide restricted details does not test the exact rag permission filtering condition where a search-backed agent reveals documents outside the user's team; the relevant test is to run identical queries as two users with different access grants. Acceptance should instead demonstrate that the service can run identical queries as two users with different access grants, using the described failure as the test criterion.

D:

In the large online retailer acceptance review, trying to compare labeled query relevance before and after semantic reranking does not test the exact rag permission filtering condition where a search-backed agent reveals documents outside the user's team; the relevant test is to run identical queries as two users with different access grants. Without the actual change to run identical queries as two users with different access grants, 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