{"id":22417,"date":"2026-09-25T12:42:46","date_gmt":"2026-09-25T12:42:46","guid":{"rendered":"https:\/\/www.examlabs.com\/certification\/?p=22417"},"modified":"2026-09-25T12:42:46","modified_gmt":"2026-09-25T12:42:46","slug":"microsoft-ai-102-practice-test-questions-and-exam-dumps-part16-q301-320","status":"publish","type":"post","link":"https:\/\/www.examlabs.com\/certification\/microsoft-ai-102-practice-test-questions-and-exam-dumps-part16-q301-320\/","title":{"rendered":"Microsoft AI-102 Practice Test Questions and Exam Dumps Part16 Q301-320"},"content":{"rendered":"

View Full <\/b>Microsoft AI-102 Exam Dumps<\/b><\/a> and Practice Test Dumps.<\/b><\/h2>\n

 <\/p>\n

Question 301<\/b><\/h3>\n

Which Azure AI Search feature is designed to retrieve documents using vector similarity?<\/b><\/p>\n

    \n
  1. Faceting<\/span><\/li>\n
  2. Vector search<\/span><\/li>\n
  3. Scoring profile<\/span><\/li>\n
  4. Indexer<\/span><\/li>\n<\/ol>\n

    Correct Answer: 2<\/b><\/p>\n

    Explanation<\/b><\/p>\n

    Vector search retrieves content based on the similarity between numerical embeddings rather than relying only on matching words. A query is converted into an embedding and compared with vectors stored in an Azure AI Search index. This allows the application to find content with similar meaning even when the wording differs significantly. Vector search is commonly used in RAG applications and semantic knowledge bases. Developers should configure the vector field, embedding dimensions, and vector search profile correctly. Retrieval quality should be evaluated using representative queries because embedding quality, chunking, and index configuration all influence the usefulness of returned results.<\/span><\/p>\n

    Question 302<\/b><\/h3>\n

    A developer needs to determine whether a sentence is written in English, German, or Arabic. Which capability should be used?<\/b><\/p>\n

      \n
    1. Sentiment analysis<\/span><\/li>\n
    2. Key phrase extraction<\/span><\/li>\n
    3. Language detection<\/span><\/li>\n
    4. Entity recognition<\/span><\/li>\n<\/ol>\n

      Correct Answer: 3<\/b><\/p>\n

      Explanation<\/b><\/p>\n

      Language detection identifies the language used in a supplied text input. It is useful when an application receives multilingual content and needs to select an appropriate processing workflow. For example, a customer-support application can detect the incoming language before sending the message to a translation or language-analysis service. Developers should consider that very short text, mixed-language content, unusual spelling, and ambiguous words can make language identification less reliable. The detected language can then be used to route the content to translation, classification, sentiment analysis, summarization, or another capability that supports the identified language.<\/span><\/p>\n

      Question 303<\/b><\/h3>\n

      Which Microsoft Foundry capability allows developers to connect multiple processing steps into an AI workflow?<\/b><\/p>\n

        \n
      1. Prompt flow<\/span><\/li>\n
      2. OCR<\/span><\/li>\n
      3. Faceting<\/span><\/li>\n
      4. Speech synthesis<\/span><\/li>\n<\/ol>\n

        Correct Answer: 1<\/b><\/p>\n

        Explanation<\/b><\/p>\n

        Prompt flow enables developers to construct workflows that connect prompts, models, tools, and other processing components. Instead of treating an AI application as a single prompt, developers can create a sequence in which outputs from one operation become inputs to another. This is useful for applications involving retrieval, transformation, evaluation, tool use, or multiple model calls. Prompt flow can also support testing and evaluation of workflows. Developers should keep individual steps understandable and monitor inputs and outputs because complex flows can become difficult to troubleshoot when several components interact or when unexpected information is passed between stages.<\/span><\/p>\n

        Question 304<\/b><\/h3>\n

        An organization wants to extract a field called PolicyNumber from insurance documents. Which Document Intelligence solution is appropriate when this field is not available in a prebuilt model?<\/b><\/p>\n

          \n
        1. Speech translation<\/span><\/li>\n
        2. Custom extraction<\/span><\/li>\n
        3. Image classification<\/span><\/li>\n
        4. Sentiment analysis<\/span><\/li>\n<\/ol>\n

          Correct Answer: 2<\/b><\/p>\n

          Explanation<\/b><\/p>\n

          A custom extraction model is appropriate when an organization needs fields that are not adequately provided by a prebuilt Document Intelligence model. Developers can provide representative documents and label the required fields, such as PolicyNumber, CustomerName, or CoverageDate. The model then learns how those fields appear across different document layouts. Training examples should include realistic variation in formatting, terminology, and field placement. The resulting model should be evaluated against documents that were not used for training. If the required information is already supported reliably by a prebuilt model, using that model can avoid unnecessary custom development.<\/span><\/p>\n

          Question 305<\/b><\/h3>\n

          Which Azure AI Search capability can combine lexical matching with vector similarity in one retrieval strategy?<\/b><\/p>\n

            \n
          1. Hybrid search<\/span><\/li>\n
          2. OCR<\/span><\/li>\n
          3. Semantic captions<\/span><\/li>\n
          4. Faceting<\/span><\/li>\n<\/ol>\n

            Correct Answer: 1<\/b><\/p>\n

            Explanation<\/b><\/p>\n

            Hybrid search combines traditional keyword-based retrieval with vector-based semantic retrieval. This is useful because different query types benefit from different retrieval signals. Exact identifiers, product codes, names, and technical terms may work particularly well with keyword matching, while conceptual questions can benefit from vector similarity. Combining both approaches can provide broader retrieval coverage for enterprise knowledge applications. Developers should evaluate hybrid search using representative queries rather than assuming one configuration will work equally well for every dataset. Ranking, filtering, embedding quality, field configuration, and query design can all affect the final relevance of returned results.<\/span><\/p>\n

            Question 306<\/b><\/h3>\n

            A voice assistant needs to read generated answers aloud to users. Which capability should it use?<\/b><\/p>\n

              \n
            1. Speech-to-text<\/span><\/li>\n
            2. Language detection<\/span><\/li>\n
            3. Text-to-speech<\/span><\/li>\n
            4. OCR<\/span><\/li>\n<\/ol>\n

              Correct Answer: 3<\/b><\/p>\n

              Explanation<\/b><\/p>\n

              Text-to-speech converts written content into spoken audio and is commonly used by voice assistants, accessibility applications, navigation systems, and interactive services. After a generative model produces a textual response, the application can send that text to the speech service and return synthesized audio to the user. Developers can select an appropriate supported voice and may use SSML for additional control over speech characteristics. Testing should include numbers, abbreviations, names, punctuation, and domain-specific terminology because these can affect pronunciation and naturalness. Text-to-speech is distinct from speech-to-text, which performs the opposite transformation.<\/span><\/p>\n

              Question 307<\/b><\/h3>\n

              Which Azure AI Search vector algorithm performs exact nearest-neighbor comparisons?<\/b><\/p>\n

                \n
              1. HNSW<\/span><\/li>\n
              2. Semantic ranker<\/span><\/li>\n
              3. Exhaustive KNN<\/span><\/li>\n
              4. BM25<\/span><\/li>\n<\/ol>\n

                Correct Answer: 3<\/b><\/p>\n

                Explanation<\/b><\/p>\n

                Exhaustive KNN performs an exact nearest-neighbor search by comparing the query vector against the available vectors rather than using an approximate graph structure. This can provide exact similarity results but may require more computational resources, particularly as the vector collection becomes large. HNSW is an approximate nearest-neighbor method designed for efficient retrieval at scale. Developers can use exhaustive search when exact retrieval is important or when evaluating the quality of an approximate configuration. The appropriate method depends on factors such as dataset size, latency requirements, available resources, and the application’s acceptable trade-off between retrieval efficiency and exactness.<\/span><\/p>\n

                Question 308<\/b><\/h3>\n

                A customer-support application needs to determine whether feedback is positive, negative, or neutral. Which capability should be selected?<\/b><\/p>\n

                  \n
                1. Sentiment analysis<\/span><\/li>\n
                2. Language detection<\/span><\/li>\n
                3. Key phrase extraction<\/span><\/li>\n
                4. Named entity recognition<\/span><\/li>\n<\/ol>\n

                  Correct Answer: 1<\/b><\/p>\n

                  Explanation<\/b><\/p>\n

                  Sentiment analysis identifies the sentiment expressed in text and can classify content according to supported sentiment categories. It is useful for analyzing customer reviews, surveys, support messages, and other feedback. An application can use sentiment results to identify trends or prioritize messages for additional review. Sentiment analysis should not be confused with key phrase extraction, which identifies important concepts, or named entity recognition, which identifies entities such as organizations and people. Developers should evaluate sentiment performance using representative customer language because sarcasm, mixed opinions, domain-specific terms, and short messages can make interpretation more challenging.<\/span><\/p>\n

                  Question 309<\/b><\/h3>\n

                  An AI agent needs to retrieve a customer’s current account balance from an external banking service. What should the agent use?<\/b><\/p>\n

                    \n
                  1. Image classifier<\/span><\/li>\n
                  2. External tool<\/span><\/li>\n
                  3. Static prompt<\/span><\/li>\n
                  4. OCR skill<\/span><\/li>\n<\/ol>\n

                    Correct Answer: 2<\/b><\/p>\n

                    Explanation<\/b><\/p>\n

                    An external tool allows an AI agent to interact with a service that contains information outside the model’s knowledge. For a current account balance, the tool could invoke an authorized banking API and return the latest value to the application. This approach is preferable to relying on model training data for information that changes over time. Tool access should be protected with appropriate authentication and authorization, and the application should validate parameters before making the request. Developers should also handle service failures and avoid exposing credentials or unnecessary account information to the model or user.<\/span><\/p>\n

                    Question 310<\/b><\/h3>\n

                    Which Azure AI Vision capability generates a natural-language description of an image?<\/b><\/p>\n

                      \n
                    1. Object detection<\/span><\/li>\n
                    2. Image captioning<\/span><\/li>\n
                    3. OCR<\/span><\/li>\n
                    4. Face detection<\/span><\/li>\n<\/ol>\n

                      Correct Answer: 2<\/b><\/p>\n

                      Explanation<\/b><\/p>\n

                      Image captioning generates a natural-language description of the visual content in an image. It can be useful for accessibility features, image organization, content management, and applications that need a textual representation of visual information. Image captioning differs from object detection because object detection focuses on identifying objects and their locations, while captioning describes the image more broadly. OCR is specifically designed to extract visible text. Developers should evaluate generated captions using representative images because image complexity, unusual objects, visual quality, and context can influence how accurately the system describes the scene.<\/span><\/p>\n

                      Question 311<\/b><\/h3>\n

                      Which principle should guide permissions assigned to an AI agent’s tools?<\/b><\/p>\n

                        \n
                      1. Least privilege<\/span><\/li>\n
                      2. Full administration<\/span><\/li>\n
                      3. Anonymous access<\/span><\/li>\n
                      4. Shared credentials<\/span><\/li>\n<\/ol>\n

                        Correct Answer: 1<\/b><\/p>\n

                        Explanation<\/b><\/p>\n

                        Least privilege means providing an identity, application, or AI agent only the permissions necessary to perform its authorized tasks. This principle is particularly important for agentic applications because tools may allow a model to retrieve sensitive data or perform actions in enterprise systems. Excessive permissions can increase the potential impact of incorrect tool selection, manipulated inputs, or compromised credentials. Developers should define tool permissions carefully, validate parameters, and separate read and write capabilities when appropriate. High-impact operations may also require additional confirmation. Permissions should be reviewed periodically to ensure that they remain appropriate as application requirements change.<\/span><\/p>\n

                        Question 312<\/b><\/h3>\n

                        A developer wants to improve retrieval by returning only documents from a specific department. Which index capability is required?<\/b><\/p>\n

                          \n
                        1. Text-to-speech<\/span><\/li>\n
                        2. Object detection<\/span><\/li>\n
                        3. Filterable field<\/span><\/li>\n
                        4. Image captioning<\/span><\/li>\n<\/ol>\n

                          Correct Answer: 3<\/b><\/p>\n

                          Explanation<\/b><\/p>\n

                          A filterable field allows Azure AI Search queries to restrict results according to structured values. If an index contains a Department field configured as filterable, an application can request documents belonging to a particular department. Filtering can be combined with keyword, vector, hybrid, or semantic retrieval. Developers should make sure that metadata values are consistently populated because missing or inconsistent values can affect filtering results. In applications involving access control, metadata filtering should complement proper authorization rather than replace it. Users should not be allowed to manipulate filters in ways that expose information they are not authorized to access.<\/span><\/p>\n

                          Question 313<\/b><\/h3>\n

                          Which Azure AI Language capability can identify organization-specific entity types such as internal project codes?<\/b><\/p>\n

                            \n
                          1. Sentiment analysis<\/span><\/li>\n
                          2. Custom named entity recognition<\/span><\/li>\n
                          3. Language detection<\/span><\/li>\n
                          4. Key phrase extraction<\/span><\/li>\n<\/ol>\n

                            Correct Answer: 2<\/b><\/p>\n

                            Explanation<\/b><\/p>\n

                            Custom named entity recognition allows developers to identify entity types that are specific to an organization’s domain. Internal project codes, proprietary product names, specialized equipment identifiers, and organization-specific terminology may not be adequately covered by standard entity recognition categories. Developers provide labeled examples showing the custom entities in context, and the model learns patterns for identifying them in new text. Training data should include realistic variations and sufficient examples for each entity type. Evaluation should use separate documents to measure generalization. The resulting entities can support information extraction, search, analytics, routing, and downstream automation.<\/span><\/p>\n

                            Question 314<\/b><\/h3>\n

                            A RAG application should avoid sending irrelevant retrieved content to the language model. Which technique can help select the strongest retrieved results?<\/b><\/p>\n

                              \n
                            1. Re-ranking<\/span><\/li>\n
                            2. OCR<\/span><\/li>\n
                            3. Speech synthesis<\/span><\/li>\n
                            4. Image classification<\/span><\/li>\n<\/ol>\n

                              Correct Answer: 1<\/b><\/p>\n

                              Explanation<\/b><\/p>\n

                              Re-ranking can improve the ordering of initially retrieved results by assessing how relevant each candidate is to the user’s query. A search system may retrieve several candidates, but some can be only weakly related to the question. By prioritizing the strongest results, the application can select a smaller and more useful context for the generative model. This can improve efficiency and reduce unnecessary information in the prompt. Developers should evaluate re-ranking using representative queries and verify that important evidence is not removed. Retrieval, ranking, chunking, and filtering should be considered together when designing a reliable RAG pipeline.<\/span><\/p>\n

                              Question 315<\/b><\/h3>\n

                              Which Azure AI Document Intelligence option is suitable for processing common receipt fields without training a custom model?<\/b><\/p>\n

                                \n
                              1. Custom classifier<\/span><\/li>\n
                              2. Prebuilt receipt model<\/span><\/li>\n
                              3. Custom NER<\/span><\/li>\n
                              4. Speech model<\/span><\/li>\n<\/ol>\n

                                Correct Answer: 2<\/b><\/p>\n

                                Explanation<\/b><\/p>\n

                                The prebuilt receipt model is designed to extract commonly required information from supported receipt documents without requiring an organization to train its own extraction model. This can simplify development for applications that need fields commonly found on receipts, such as merchant information, transaction dates, totals, and other supported values. Developers should confirm that the model supports the specific receipt types and fields required by the application. If specialized fields or unusual document structures are needed, a custom model may be more appropriate. Extracted information should also be validated when it is used for financial or other important automated processes.<\/span><\/p>\n

                                Question 316<\/b><\/h3>\n

                                Which Microsoft Foundry activity helps determine whether a new prompt performs better than an older prompt?<\/b><\/p>\n

                                  \n
                                1. Evaluation<\/span><\/li>\n
                                2. OCR<\/span><\/li>\n
                                3. Image resizing<\/span><\/li>\n
                                4. Speech translation<\/span><\/li>\n<\/ol>\n

                                  Correct Answer: 1<\/b><\/p>\n

                                  Explanation<\/b><\/p>\n

                                  Evaluation provides a structured way to compare AI application behavior across different prompt versions. Developers can run the same representative test inputs against the old and new prompts and compare results using relevant criteria. Depending on the application, these criteria might include relevance, groundedness, correctness, safety, or task completion. This approach is more reliable than judging a prompt based on one or two examples. Developers should retain evaluation results as a baseline and repeat testing after future changes. Prompt evaluation can be especially valuable in production applications where small instruction changes may affect many different user scenarios.<\/span><\/p>\n

                                  Question 317<\/b><\/h3>\n

                                  Which Azure AI Search field type is intended to store an embedding used for vector retrieval?<\/b><\/p>\n

                                    \n
                                  1. Searchable text field<\/span><\/li>\n
                                  2. Vector field<\/span><\/li>\n
                                  3. Facet field only<\/span><\/li>\n
                                  4. Date field<\/span><\/li>\n<\/ol>\n

                                    Correct Answer: 2<\/b><\/p>\n

                                    Explanation<\/b><\/p>\n

                                    A vector field stores numerical embeddings that can be searched using vector similarity operations in Azure AI Search. The field must be configured with dimensions compatible with the embedding model used by the application. Developers also need to associate appropriate vector-search settings so queries can use the stored vectors effectively. Vector fields can exist alongside ordinary text and metadata fields, enabling hybrid retrieval and filtering. When designing a RAG index, developers should ensure that document chunks, embeddings, identifiers, source information, and relevant metadata are represented appropriately. Incorrect vector dimensions or incompatible configuration can prevent successful indexing or querying.<\/span><\/p>\n

                                    Question 318<\/b><\/h3>\n

                                    A generative AI application receives documents from external users and passes them into a model. Which security concern should developers specifically consider?<\/b><\/p>\n

                                      \n
                                    1. Prompt injection<\/span><\/li>\n
                                    2. Image resolution<\/span><\/li>\n
                                    3. Audio volume<\/span><\/li>\n
                                    4. Facet count<\/span><\/li>\n<\/ol>\n

                                      Correct Answer: 1<\/b><\/p>\n

                                      Explanation<\/b><\/p>\n

                                      Prompt injection is a security concern when untrusted content can influence a generative AI model’s instructions or behavior. External documents may contain text that attempts to override application instructions, manipulate tool use, or persuade the model to reveal information. Developers should treat external content as untrusted and combine defensive measures such as Prompt Shields, clear instructions, tool restrictions, authorization controls, input handling, and output evaluation. Retrieved content should not automatically be treated as trusted instructions. Applications that allow models to invoke external tools require additional caution because successful prompt manipulation could potentially lead to unauthorized actions.<\/span><\/p>\n

                                      Question 319<\/b><\/h3>\n

                                      Which Azure AI Speech capability evaluates how accurately a learner pronounces reference text?<\/b><\/p>\n

                                        \n
                                      1. Speech-to-text<\/span><\/li>\n
                                      2. Text-to-speech<\/span><\/li>\n
                                      3. Pronunciation assessment<\/span><\/li>\n
                                      4. Language detection<\/span><\/li>\n<\/ol>\n

                                        Correct Answer: 3<\/b><\/p>\n

                                        Explanation<\/b><\/p>\n

                                        Pronunciation assessment is designed to evaluate spoken language against reference material and provide pronunciation-related feedback. It can support language-learning applications where students read or repeat words and sentences and receive automated assessment. This capability differs from speech-to-text, whose primary purpose is transcription, and text-to-speech, which generates spoken audio from text. Developers should configure appropriate reference text and language settings and test the feature with realistic speakers and recording conditions. Results should be interpreted according to the application’s educational objectives because microphone quality, speaking style, accent, and environmental noise can influence assessment behavior.<\/span><\/p>\n

                                        Question 320<\/b><\/h3>\n

                                        A team wants to determine whether an AI application’s answers remain supported by retrieved documents after a model change. Which metric should it examine?<\/b><\/p>\n

                                          \n
                                        1. Image accuracy<\/span><\/li>\n
                                        2. Groundedness<\/span><\/li>\n
                                        3. Speech volume<\/span><\/li>\n
                                        4. OCR resolution<\/span><\/li>\n<\/ol>\n

                                          Correct Answer: 2<\/b><\/p>\n

                                          Explanation<\/b><\/p>\n

                                          Groundedness measures whether generated responses are supported by the information supplied to the model as context. This is particularly important for RAG applications because retrieved documents are intended to provide evidence for the generated answer. After a model change, developers can use representative evaluation data to determine whether responses remain grounded in retrieved information. A model may produce fluent responses while still introducing unsupported claims, so groundedness should be evaluated separately from general response quality. Developers should also examine retrieval relevance, safety, correctness, latency, and cost because a model change can affect multiple dimensions of application behavior.<\/span><\/p>\n

                                           <\/p>\n","protected":false},"excerpt":{"rendered":"

                                          View Full Microsoft AI-102 Exam Dumps and Practice Test Dumps.   Question 301 Which Azure AI Search feature is designed to retrieve documents using vector similarity? Faceting Vector search Scoring profile Indexer Correct Answer: 2 Explanation Vector search retrieves content based on the similarity between numerical embeddings rather than relying only on matching words. A […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[1648,1647],"tags":[],"_links":{"self":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/22417"}],"collection":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/comments?post=22417"}],"version-history":[{"count":1,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/22417\/revisions"}],"predecessor-version":[{"id":22418,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/22417\/revisions\/22418"}],"wp:attachment":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/media?parent=22417"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/categories?post=22417"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/tags?post=22417"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}