Microsoft AI-102 (Designing and Implementing a Microsoft Azure AI Solution) Practice Test 47

Question 460

A trained CLU model is tested in an authoring tool but the live bot still returns an old intent taxonomy. What step may be missing?

  1. Define distinct intents corresponding to the different user actions
  2. Evaluate entity extraction separately from intent classification
  3. Deploy the intended trained model version to the prediction endpoint
  4. Add varied natural utterance phrasings for each supported intent

Explanation

Correct Answer

C:

Training and local tests do not necessarily replace the deployed runtime model. In the service-request virtual assistant, confirm the fix by retesting the affected behavior after deploying the intended trained model version to the prediction endpoint.

Incorrect Answers

A:

Although useful elsewhere in the service-request virtual assistant, cLU intents represent the purpose of an utterance; conflating different actions prevents reliable routing; the present case is different and calls for the team to deploy the intended trained model version to the prediction endpoint.

B:

That change addresses another failure mode: cLU has distinct intent and entity predictions; success at one task does not prove the other works; it does not address this case, for which the relevant intervention is to deploy the intended trained model version to the prediction endpoint.

D:

Linguistic diversity helps a classifier learn user intent beyond superficial wording patterns in its own workflow, but that is a different correction from the present requirement to deploy the intended trained model version to the prediction endpoint. The reported issue would remain.

Question 461

The user responds to a follow-up with “the laptop,” but the bot starts a completely unrelated question. What state is missing?

  1. Replace overly generic alternate questions with intent-specific natural paraphrases
  2. Target the intended knowledge-project deployment identifier in prediction requests
  3. Retain the relevant previous answer or follow-up context for the next turn
  4. Review and correct extracted question-answer pairs from complex source layouts

Explanation

Correct Answer

C:

A follow-up reply must be interpreted in relation to the prior prompt to complete the dialogue. In the employee FAQ chat interface, confirm the fix by retesting the affected behavior after retaining the relevant previous answer or follow-up context for the next turn.

Incorrect Answers

A:

That change addresses another failure mode: very broad phrasing creates ambiguity across multiple answers rather than improving precision; it does not address this case, for which the relevant intervention is to retain the relevant previous answer or follow-up context for the next turn.

B:

Although useful elsewhere in the employee FAQ chat interface, application routing can continue pointing at an older release even when the new one exists; the present case is different and calls for the team to retain the relevant previous answer or follow-up context for the next turn.

D:

Automatic source ingestion can misread structured documents and requires editorial validation in its own workflow, but that is a different correction from the present requirement to retain the relevant previous answer or follow-up context for the next turn. The reported issue would remain.

Question 462

In an employee FAQ chat interface, engineers must handle common greetings and conversational exchanges consistently with the organization's tone. The expected evidence is that greetings receive the configured brief reply. Which implementation is the best match?

  1. Add alternate questions to an answer
  2. Build a multi-turn question-answering conversation
  3. Import curated FAQ source content
  4. Add a custom chit-chat response

Explanation

Correct Answer

D:

Chit-chat entries provide approved responses for short social or conversational inputs within the knowledge base. For this employee FAQ chat interface, the intended evidence is that greetings receive the configured brief reply. They are not a substitute for authentication before disclosing account data.

Incorrect Answers

A:

Alternate questions link multiple formulations to a single answer, improving match coverage without duplicating its content; it may be used to recognize several paraphrases that should return the same approved response in a employee FAQ chat interface, but the current requirement is instead to handle common greetings and conversational exchanges consistently with the organization's tone. That separate capability does not prove that greetings receive the configured brief reply; They do not modify the answer text for every language automatically, so selecting that alternative would not meet the stated acceptance check.

B:

Multi-turn interactions use follow-up prompts and conversational context to refine which answer applies; it may be used to ask clarifying follow-up questions before selecting the final FAQ response in a employee FAQ chat interface, but the current requirement is instead to handle common greetings and conversational exchanges consistently with the organization's tone. That separate capability does not prove that greetings receive the configured brief reply; A one-shot key phrase extraction call cannot conduct the follow-up dialog, so selecting that alternative would not meet the stated acceptance check.

C:

Custom question answering builds a knowledge source from configured documents, URLs, or explicit Q&A pairs; it may be used to seed a question-answering project with approved reference questions and answers in a employee FAQ chat interface, but the current requirement is instead to handle common greetings and conversational exchanges consistently with the organization's tone. That separate capability does not prove that greetings receive the configured brief reply; Source import does not require training a conversational intent classifier first, so selecting that alternative would not meet the stated acceptance check.

Question 463

Editors fix a critical HR answer, but callers continue receiving the outdated text from production. What should they verify?

  1. Refresh the approved FAQ source content and validate the resulting answer pairs
  2. Add approved chit-chat responses for common polite exchanges
  3. Deploy or publish the updated knowledge project before testing the runtime endpoint
  4. Configure a multi-turn follow-up question to collect the missing detail

Explanation

Correct Answer

C:

The live prediction service may use a previously deployed snapshot rather than the latest authoring changes. In the employee FAQ chat interface, confirm the fix by retesting the affected behavior after deploying or publish the updated knowledge project before testing the runtime endpoint.

Incorrect Answers

A:

That change addresses another failure mode: question-answering quality depends on the current curated source knowledge, not only the query model; it does not address this case, for which the relevant intervention is to deploy or publish the updated knowledge project before testing the runtime endpoint.

B:

Although useful elsewhere in the employee FAQ chat interface, chit-chat entries handle conversational phrases that are not substantive policy questions; the present case is different and calls for the team to deploy or publish the updated knowledge project before testing the runtime endpoint.

D:

Multi-turn question answering supports clarification instead of guessing which applicable answer the user wanted in its own workflow, but that is a different correction from the present requirement to deploy or publish the updated knowledge project before testing the runtime endpoint. The reported issue would remain.

Question 464

The solution brief for an employee FAQ chat interface requires the system to seed a question-answering project with approved reference questions and answers. A successful test confirms that the knowledge project contains the expected answer pairs. What should the team use?

  1. Add a custom chit-chat response
  2. Import curated FAQ source content
  3. Add alternate questions to an answer
  4. Publish the question-answering knowledge project

Explanation

Correct Answer

B:

Custom question answering builds a knowledge source from configured documents, URLs, or explicit Q&A pairs. For this employee FAQ chat interface, the intended evidence is that the knowledge project contains the expected answer pairs. Source import does not require training a conversational intent classifier first.

Incorrect Answers

A:

Chit-chat entries provide approved responses for short social or conversational inputs within the knowledge base; it may be used to handle common greetings and conversational exchanges consistently with the organization's tone in a employee FAQ chat interface, but the current requirement is instead to seed a question-answering project with approved reference questions and answers. That separate capability does not prove that the knowledge project contains the expected answer pairs; They are not a substitute for authentication before disclosing account data, so selecting that alternative would not meet the stated acceptance check.

C:

Alternate questions link multiple formulations to a single answer, improving match coverage without duplicating its content; it may be used to recognize several paraphrases that should return the same approved response in a employee FAQ chat interface, but the current requirement is instead to seed a question-answering project with approved reference questions and answers. That separate capability does not prove that the knowledge project contains the expected answer pairs; They do not modify the answer text for every language automatically, so selecting that alternative would not meet the stated acceptance check.

D:

A deployed question-answering project can be queried by client applications using the authorized endpoint and project identifiers; it may be used to make curated FAQ content available to an application at a runtime endpoint in a employee FAQ chat interface, but the current requirement is instead to seed a question-answering project with approved reference questions and answers. That separate capability does not prove that the knowledge project contains the expected answer pairs; Editing content in a draft alone need not make it available to production callers, so selecting that alternative would not meet the stated acceptance check.

Question 465

A team responsible for an employee FAQ chat interface needs to recognize several paraphrases that should return the same approved response; the acceptance check is that distinct user phrasings match one curated answer. Which capability should be selected?

  1. Add alternate questions to an answer
  2. Import curated FAQ source content
  3. Publish the question-answering knowledge project
  4. Build a multi-turn question-answering conversation

Explanation

Correct Answer

A:

Alternate questions link multiple formulations to a single answer, improving match coverage without duplicating its content. For this employee FAQ chat interface, the intended evidence is that distinct user phrasings match one curated answer. They do not modify the answer text for every language automatically.

Incorrect Answers

B:

Custom question answering builds a knowledge source from configured documents, URLs, or explicit Q&A pairs; it may be used to seed a question-answering project with approved reference questions and answers in a employee FAQ chat interface, but the current requirement is instead to recognize several paraphrases that should return the same approved response. That separate capability does not prove that distinct user phrasings match one curated answer; Source import does not require training a conversational intent classifier first, so selecting that alternative would not meet the stated acceptance check.

C:

A deployed question-answering project can be queried by client applications using the authorized endpoint and project identifiers; it may be used to make curated FAQ content available to an application at a runtime endpoint in a employee FAQ chat interface, but the current requirement is instead to recognize several paraphrases that should return the same approved response. That separate capability does not prove that distinct user phrasings match one curated answer; Editing content in a draft alone need not make it available to production callers, so selecting that alternative would not meet the stated acceptance check.

D:

Multi-turn interactions use follow-up prompts and conversational context to refine which answer applies; it may be used to ask clarifying follow-up questions before selecting the final FAQ response in a employee FAQ chat interface, but the current requirement is instead to recognize several paraphrases that should return the same approved response. That separate capability does not prove that distinct user phrasings match one curated answer; A one-shot key phrase extraction call cannot conduct the follow-up dialog, so selecting that alternative would not meet the stated acceptance check.

Question 466

A knowledge base imports employee-only HR policies into a public bot. What safeguard is required?

  1. Separate or filter source content according to user authorization before publishing
  2. Review chit-chat matching behavior and remove overly broad conversational triggers
  3. Check the follow-up transition and answer selection conditions for that branch
  4. Group equivalent phrasings under one authoritative answer instead of cloning content entries

Explanation

Correct Answer

A:

Curated content must still obey access scope; successful import is not permission to expose private answers. In the employee FAQ chat interface, confirm the fix by retesting the affected behavior after separating or filter source content according to user authorization before publishing.

Incorrect Answers

B:

Although useful elsewhere in the employee FAQ chat interface, over-generalized greeting entries can compete with real help requests; the present case is different and calls for the team to separate or filter source content according to user authorization before publishing.

C:

That change addresses another failure mode: dialogue paths need a terminal or next-answer transition when their missing parameter is satisfied; it does not address this case, for which the relevant intervention is to separate or filter source content according to user authorization before publishing.

D:

Duplicate answer records increase maintenance risk; one answer with alternates keeps updates consistent in its own workflow, but that is a different correction from the present requirement to separate or filter source content according to user authorization before publishing. The reported issue would remain.

Question 467

A typo-heavy query matches the wrong FAQ after editors add a broad alternate question like “help.” What should change?

  1. Replace overly generic alternate questions with intent-specific natural paraphrases
  2. Target the intended knowledge-project deployment identifier in prediction requests
  3. Review and correct extracted question-answer pairs from complex source layouts
  4. Retain the relevant previous answer or follow-up context for the next turn

Explanation

Correct Answer

A:

Very broad phrasing creates ambiguity across multiple answers rather than improving precision. In the employee FAQ chat interface, confirm the fix by retesting the affected behavior after replacing overly generic alternate questions with intent-specific natural paraphrases.

Incorrect Answers

B:

That change addresses another failure mode: application routing can continue pointing at an older release even when the new one exists; it does not address this case, for which the relevant intervention is to replace overly generic alternate questions with intent-specific natural paraphrases.

C:

Although useful elsewhere in the employee FAQ chat interface, automatic source ingestion can misread structured documents and requires editorial validation; the present case is different and calls for the team to replace overly generic alternate questions with intent-specific natural paraphrases.

D:

A follow-up reply must be interpreted in relation to the prior prompt to complete the dialogue in its own workflow, but that is a different correction from the present requirement to replace overly generic alternate questions with intent-specific natural paraphrases. The reported issue would remain.

Question 468

The help bot replies “no match” when an employee simply says “thanks.” Which authoring customization could improve usability?

  1. Configure a multi-turn follow-up question to collect the missing detail
  2. Refresh the approved FAQ source content and validate the resulting answer pairs
  3. Add approved chit-chat responses for common polite exchanges
  4. Deploy or publish the updated knowledge project before testing the runtime endpoint

Explanation

Correct Answer

C:

Chit-chat entries handle conversational phrases that are not substantive policy questions. In the employee FAQ chat interface, confirm the fix by retesting the affected behavior after adding approved chit-chat responses for common polite exchanges.

Incorrect Answers

A:

Although useful elsewhere in the employee FAQ chat interface, multi-turn question answering supports clarification instead of guessing which applicable answer the user wanted; the present case is different and calls for the team to add approved chit-chat responses for common polite exchanges.

B:

That change addresses another failure mode: question-answering quality depends on the current curated source knowledge, not only the query model; it does not address this case, for which the relevant intervention is to add approved chit-chat responses for common polite exchanges.

D:

The live prediction service may use a previously deployed snapshot rather than the latest authoring changes in its own workflow, but that is a different correction from the present requirement to add approved chit-chat responses for common polite exchanges. The reported issue would remain.

Question 469

A company configures a playful greeting that conflicts with its regulated support tone. What editorial control is appropriate?

  1. Target the intended knowledge-project deployment identifier in prediction requests
  2. Replace overly generic alternate questions with intent-specific natural paraphrases
  3. Review and correct extracted question-answer pairs from complex source layouts
  4. Review and approve chit-chat text under the organization's communication policy

Explanation

Correct Answer

D:

Predefined or custom conversational responses still represent the organization and require appropriate tone and review. In the employee FAQ chat interface, confirm the fix by retesting the affected behavior after reviewing and approve chit-chat text under the organization's communication policy.

Incorrect Answers

A:

That change addresses another failure mode: application routing can continue pointing at an older release even when the new one exists; it does not address this case, for which the relevant intervention is to review and approve chit-chat text under the organization's communication policy.

B:

Although useful elsewhere in the employee FAQ chat interface, very broad phrasing creates ambiguity across multiple answers rather than improving precision; the present case is different and calls for the team to review and approve chit-chat text under the organization's communication policy.

C:

Automatic source ingestion can misread structured documents and requires editorial validation in its own workflow, but that is a different correction from the present requirement to review and approve chit-chat text under the organization's communication policy. The reported issue would remain.

Question 470

A help-desk release must be audited later. What should the publication workflow record?

  1. Separate or filter source content according to user authorization before publishing
  2. Review chit-chat matching behavior and remove overly broad conversational triggers
  3. Check the follow-up transition and answer selection conditions for that branch
  4. Version the approved FAQ content and the deployment promoted to production

Explanation

Correct Answer

D:

Content and deployment provenance establish which answer set was exposed at a given time. In the employee FAQ chat interface, confirm the fix by retesting the affected behavior after versioning the approved FAQ content and the deployment promoted to production.

Incorrect Answers

A:

That change addresses another failure mode: curated content must still obey access scope; successful import is not permission to expose private answers; it does not address this case, for which the relevant intervention is to version the approved FAQ content and the deployment promoted to production.

B:

Over-generalized greeting entries can compete with real help requests in its own workflow, but that is a different correction from the present requirement to version the approved FAQ content and the deployment promoted to production. The reported issue would remain.

C:

Although useful elsewhere in the employee FAQ chat interface, dialogue paths need a terminal or next-answer transition when their missing parameter is satisfied; the present case is different and calls for the team to version the approved FAQ content and the deployment promoted to production.

Designing and Implementing a Microsoft Azure AI Solution (AI-102) Practice Tests