{"id":22411,"date":"2026-09-25T12:41:58","date_gmt":"2026-09-25T12:41:58","guid":{"rendered":"https:\/\/www.examlabs.com\/certification\/?p=22411"},"modified":"2026-09-25T12:41:58","modified_gmt":"2026-09-25T12:41:58","slug":"microsoft-ai-102-practice-test-questions-and-exam-dumps-part13-q241-260","status":"publish","type":"post","link":"https:\/\/www.examlabs.com\/certification\/microsoft-ai-102-practice-test-questions-and-exam-dumps-part13-q241-260\/","title":{"rendered":"Microsoft AI-102 Practice Test Questions and Exam Dumps Part13 Q241-260"},"content":{"rendered":"

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

 <\/p>\n

Question 241<\/b><\/h3>\n

Which Microsoft Foundry capability helps developers test different prompts and model configurations against the same inputs?<\/b><\/p>\n

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

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

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

    Prompt flow provides a structured way to develop, test, evaluate, and refine generative AI workflows. Developers can connect prompts, models, tools, and other components into a workflow and then test how changes affect application behavior. Using consistent inputs makes it easier to compare different prompt instructions, model configurations, and processing steps. Prompt flow can also support evaluation and tracing activities when building AI applications. It is particularly useful when a simple single prompt is not enough and the application requires multiple processing stages. Developers should test representative scenarios rather than relying only on a few successful examples.<\/span><\/p>\n

    Question 242<\/b><\/h3>\n

    A company needs to extract custom fields from insurance forms that vary slightly between documents. Which solution is appropriate?<\/b><\/p>\n

      \n
    1. Language detection<\/span><\/li>\n
    2. Custom extraction model<\/span><\/li>\n
    3. Speech translation<\/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 in Azure AI Document Intelligence can be trained to identify and extract fields specific to an organization’s documents. This is useful when standard prebuilt models do not provide the required fields or when document layouts and terminology differ from common document types. Training examples should represent the variations expected in production, including different layouts and field values. After training, the model can process similar documents and return structured field information. Developers should evaluate extraction accuracy using documents that were not part of training and determine whether additional examples or model refinement are needed.<\/span><\/p>\n

      Question 243<\/b><\/h3>\n

      Which Azure AI Search configuration is specifically associated with semantic search queries?<\/b><\/p>\n

        \n
      1. Semantic configuration<\/span><\/li>\n
      2. Storage account<\/span><\/li>\n
      3. Speech profile<\/span><\/li>\n
      4. Vision model<\/span><\/li>\n<\/ol>\n

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

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

        A semantic configuration in Azure AI Search defines which searchable fields and related content should be considered when semantic search features are used. It can identify important fields such as a document title, main content, and other relevant information that helps the semantic ranker interpret search results. Proper configuration can improve semantic result quality because the service has clearer information about where meaningful content is located. Developers should ensure that the selected fields contain useful textual information and should test the configuration against representative queries. Semantic configuration is different from the index schema itself, although both contribute to search behavior.<\/span><\/p>\n

        Question 244<\/b><\/h3>\n

        An application needs to generate spoken audio from dynamically generated text. Which Azure AI service should be used?<\/b><\/p>\n

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

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

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

          Text-to-speech converts written text into spoken audio. It is useful for virtual assistants, accessibility applications, navigation systems, educational software, and other solutions that need to communicate information through voice. Developers can configure supported voices and, where applicable, use Speech Synthesis Markup Language to control aspects of speech output. The application should select a voice that matches the target language and intended user experience. Text-to-speech is different from speech-to-text, which converts spoken audio into text. Testing should include representative text, punctuation, numbers, abbreviations, and other content that may affect natural speech generation.<\/span><\/p>\n

          Question 245<\/b><\/h3>\n

          Which Azure AI Search vector algorithm is designed to search all vectors rather than using an approximate nearest-neighbor structure?<\/b><\/p>\n

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

            Correct Answer: 2<\/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 relying on an approximate graph-based search structure. Because it examines the vector space more directly, it can provide an exact similarity result but may require more computational resources as the number of vectors grows. HNSW, by contrast, is designed for efficient approximate nearest-neighbor retrieval. Developers can select an approach based on factors such as dataset size, latency requirements, and retrieval accuracy. Exact search can also be useful when evaluating or benchmarking approximate vector-search configurations.<\/span><\/p>\n

            Question 246<\/b><\/h3>\n

            A chatbot should remember information from earlier turns in the same conversation. What does the application need to maintain?<\/b><\/p>\n

              \n
            1. Conversation state<\/span><\/li>\n
            2. Image metadata<\/span><\/li>\n
            3. OCR output only<\/span><\/li>\n
            4. Search facets<\/span><\/li>\n<\/ol>\n

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

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

              Conversation state contains relevant information from previous interactions so an application can maintain continuity across multiple turns. For example, if a user first identifies a product and then asks about its warranty, the application needs enough context to understand what product the second question refers to. Depending on the architecture, conversation state can include previous messages, selected values, tool results, or other application-specific information. Developers should manage the amount of retained context carefully because excessively large conversations can increase latency and token consumption. Sensitive information should also be handled according to appropriate security and privacy requirements.<\/span><\/p>\n

              Question 247<\/b><\/h3>\n

              Which Azure AI Language capability identifies predefined entity types such as people, organizations, and locations?<\/b><\/p>\n

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

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

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

                Named entity recognition identifies entities in text and categorizes them according to supported entity types. Common examples include people, organizations, locations, dates, and other recognized categories. This capability can support information extraction, document organization, search, analytics, and downstream automation. For example, an application processing news articles can identify organizations and locations without requiring manual tagging. Standard named entity recognition differs from custom named entity recognition, which is intended for specialized categories defined by an organization. Developers should test entity recognition with realistic content because terminology, context, formatting, and language can affect extraction results.<\/span><\/p>\n

                Question 248<\/b><\/h3>\n

                A developer wants to prevent a generative AI application from following malicious instructions hidden inside retrieved documents. Which control is relevant?<\/b><\/p>\n

                  \n
                1. Prompt Shields<\/span><\/li>\n
                2. Faceting<\/span><\/li>\n
                3. Image tagging<\/span><\/li>\n
                4. Speech synthesis<\/span><\/li>\n<\/ol>\n

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

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

                  Prompt Shields can help protect generative AI applications from certain prompt injection and related manipulation attempts. This is particularly relevant to applications that process user-provided content or retrieve external documents because untrusted content may contain instructions intended to influence the model’s behavior. Security controls should be combined with careful prompt design, input validation, access controls, tool restrictions, and output evaluation. Prompt Shields should not be treated as a complete security solution by themselves. Developers need to evaluate the entire application workflow, especially when retrieved content can influence tool calls or access to sensitive enterprise resources.<\/span><\/p>\n

                  Question 249<\/b><\/h3>\n

                  Which Azure AI Vision capability identifies the location of multiple objects within an image?<\/b><\/p>\n

                    \n
                  1. Image classification<\/span><\/li>\n
                  2. Object detection<\/span><\/li>\n
                  3. Image captioning<\/span><\/li>\n
                  4. Text summarization<\/span><\/li>\n<\/ol>\n

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

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

                    Object detection identifies objects in an image and provides information about their locations, typically through bounding regions around detected objects. This makes it suitable for applications that need to know both what objects are present and where they appear. For example, a warehouse application could use object detection to identify packages or equipment in camera images. Image classification has a different purpose because it assigns categories to an image or visual content without necessarily identifying individual object locations. Developers should evaluate detection performance using representative images, object sizes, lighting conditions, and environments expected during actual application use.<\/span><\/p>\n

                    Question 250<\/b><\/h3>\n

                    Which approach is most appropriate for protecting an Azure AI Search index from unauthorized application access?<\/b><\/p>\n

                      \n
                    1. Public anonymous access<\/span><\/li>\n
                    2. Managed identity and authorization<\/span><\/li>\n
                    3. Hard-coded administrator password<\/span><\/li>\n
                    4. Shared credentials in source code<\/span><\/li>\n<\/ol>\n

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

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

                      Managed identity combined with appropriate authorization can provide a controlled authentication mechanism for applications accessing supported Azure resources. Rather than embedding long-lived credentials directly in application code, an Azure-hosted workload can use its assigned identity and receive only the permissions it requires. Administrators should configure access according to least privilege and review permissions regularly. Network controls and service-specific security settings may also be required depending on the architecture. Authentication alone does not grant unrestricted access, so the identity must be assigned appropriate roles or permissions for the operations the application actually needs.<\/span><\/p>\n

                      Question 251<\/b><\/h3>\n

                      A search application must return documents matching an exact product code even when semantic ranking considers other documents relevant. Which retrieval method is important?<\/b><\/p>\n

                        \n
                      1. Keyword search<\/span><\/li>\n
                      2. Image captioning<\/span><\/li>\n
                      3. Speech translation<\/span><\/li>\n
                      4. Entity linking<\/span><\/li>\n<\/ol>\n

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

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

                        Keyword search is particularly useful when exact terms, identifiers, codes, names, or other lexical matches are important. Product codes often contain strings that may not have meaningful semantic relationships, so vector similarity alone may not reliably prioritize the exact identifier. Azure AI Search can combine keyword retrieval with vector search through hybrid search when an application needs both exact matching and semantic similarity. Developers should evaluate queries containing product codes, technical identifiers, names, and natural-language descriptions. Choosing the appropriate retrieval approach depends on the type of information users search for and how that information is represented in the index.<\/span><\/p>\n

                        Question 252<\/b><\/h3>\n

                        A developer wants to extract structured fields from standard identity documents using a model provided by Azure AI Document Intelligence. Which option should be considered first?<\/b><\/p>\n

                          \n
                        1. Prebuilt identity document model<\/span><\/li>\n
                        2. Custom text classification<\/span><\/li>\n
                        3. Speech recognition<\/span><\/li>\n
                        4. Vector search<\/span><\/li>\n<\/ol>\n

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

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

                          A prebuilt identity document model is intended for supported identity documents and can extract structured information without requiring the organization to build and train an extraction model from scratch. This can reduce development effort when the application’s document types match the fields and formats supported by the prebuilt model. Developers should review the model’s supported document types and extracted fields before implementation. If the required document structure or fields are not adequately supported, a custom model may be necessary. Regardless of model type, extracted information should be validated and handled according to the application’s security and privacy requirements.<\/span><\/p>\n

                          Question 253<\/b><\/h3>\n

                          Which evaluation activity checks whether an AI application’s responses remain safe when users submit intentionally harmful prompts?<\/b><\/p>\n

                            \n
                          1. Safety evaluation<\/span><\/li>\n
                          2. OCR evaluation<\/span><\/li>\n
                          3. Image resizing<\/span><\/li>\n
                          4. Language detection<\/span><\/li>\n<\/ol>\n

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

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

                            Safety evaluation tests how an AI application behaves when exposed to harmful, abusive, manipulative, or otherwise unsafe inputs. Developers can create test cases representing relevant risks and assess whether the application follows its safety requirements. For generative AI systems, safety testing can include attempts to bypass instructions, generate prohibited content, expose sensitive information, or misuse available tools. Testing should cover both normal and adversarial scenarios and should be repeated when prompts, models, tools, or application logic change. Safety evaluation is one part of a broader evaluation strategy that can also include quality, groundedness, relevance, and performance.<\/span><\/p>\n

                            Question 254<\/b><\/h3>\n

                            Which Azure AI Search field configuration is required when a field will store vector embeddings for vector search?<\/b><\/p>\n

                              \n
                            1. Vector field configuration<\/span><\/li>\n
                            2. Facet-only configuration<\/span><\/li>\n
                            3. Speech configuration<\/span><\/li>\n
                            4. OCR configuration<\/span><\/li>\n<\/ol>\n

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

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

                              A vector field in an Azure AI Search index is configured to store embedding vectors that can be queried using vector search. The field configuration must be compatible with the dimensions of the embeddings generated by the selected embedding model and with the chosen vector-search settings. Developers also need to configure an appropriate vector search profile or algorithm according to the application’s requirements. If the vector dimensions or field configuration are inconsistent with generated embeddings, indexing or querying can fail. Vector fields can coexist with ordinary searchable fields, allowing applications to implement vector, keyword, or hybrid retrieval strategies.<\/span><\/p>\n

                              Question 255<\/b><\/h3>\n

                              A company wants an AI agent to retrieve current inventory information from an internal system instead of relying on model training data. What should the agent use?<\/b><\/p>\n

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

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

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

                                An external tool allows an AI agent to obtain information from a system outside the language model itself. For current inventory information, the tool could call an approved internal API or service and return the latest available records. This approach helps prevent the application from depending on potentially outdated information contained in model training data. Tool access should be authenticated and authorized, and the application should validate tool inputs and outputs. Developers should also define appropriate failure behavior when the external service is unavailable or returns incomplete information. Sensitive inventory data should be protected through the organization’s security controls.<\/span><\/p>\n

                                Question 256<\/b><\/h3>\n

                                Which Azure AI Language feature can convert text from one supported language into another?<\/b><\/p>\n

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

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

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

                                  Azure AI Translator provides machine translation between supported languages. Applications can use it to translate user messages, documents, interface content, or other text as part of multilingual workflows. Translator can also be integrated with other Azure AI capabilities when an application needs to process translated content further. Developers should identify the source and target languages correctly and evaluate translation quality using domain-specific terminology when the application handles specialized content. For business-critical scenarios, testing should include names, technical vocabulary, abbreviations, formatting, and context-sensitive expressions because translation quality can vary depending on the source material.<\/span><\/p>\n

                                  Question 257<\/b><\/h3>\n

                                  A developer is designing an AI Search index for a knowledge base. Which field is most appropriate for storing the main document text used during retrieval?<\/b><\/p>\n

                                    \n
                                  1. Searchable text field<\/span><\/li>\n
                                  2. Audio field<\/span><\/li>\n
                                  3. Image-only field<\/span><\/li>\n
                                  4. Password field<\/span><\/li>\n<\/ol>\n

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

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

                                    A searchable text field is appropriate for storing document content that users need to retrieve through search queries. Azure AI Search can use searchable fields for lexical retrieval and, depending on the application design, the same content can also contribute to semantic or vector-based workflows after appropriate processing. The index schema should clearly define which fields are searchable, retrievable, filterable, sortable, or facetable according to application needs. Developers should also consider document chunking when creating a RAG system because storing excessively large sections as individual retrieval units can reduce retrieval precision and increase unnecessary context.<\/span><\/p>\n

                                    Question 258<\/b><\/h3>\n

                                    Which Speech feature allows developers to control pronunciation, pauses, emphasis, or speaking style in synthesized speech?<\/b><\/p>\n

                                      \n
                                    1. SSML<\/span><\/li>\n
                                    2. OCR<\/span><\/li>\n
                                    3. HNSW<\/span><\/li>\n
                                    4. JSON filtering<\/span><\/li>\n<\/ol>\n

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

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

                                      Speech Synthesis Markup Language, or SSML, allows applications to provide structured instructions that influence how synthesized speech is produced. Depending on supported features, developers can control aspects such as pronunciation, pauses, emphasis, speaking rate, pitch, and voice selection. This can improve the naturalness and clarity of generated speech, particularly when applications contain names, technical terms, numbers, or content that requires specific delivery. Developers should ensure that the SSML structure is valid and compatible with the selected speech service and voice. Testing with realistic application content helps identify pronunciation and timing issues before deployment.<\/span><\/p>\n

                                      Question 259<\/b><\/h3>\n

                                      A knowledge-mining solution needs to enrich documents with OCR, entity recognition, and other processing steps during indexing. Which Azure AI Search component supports this workflow?<\/b><\/p>\n

                                        \n
                                      1. Skillset<\/span><\/li>\n
                                      2. Facet<\/span><\/li>\n
                                      3. Vector query<\/span><\/li>\n
                                      4. Scoring profile<\/span><\/li>\n<\/ol>\n

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

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

                                        A skillset in Azure AI Search defines a sequence of enrichment skills that can process content during an indexing pipeline. Skills can perform operations such as OCR, language detection, entity recognition, key phrase extraction, and other supported enrichment tasks. This allows unstructured documents to be transformed into additional searchable or structured information. The enriched output can then be mapped into fields within the search index or used by other parts of the knowledge-mining workflow. Developers should design the skillset according to the source documents and desired outputs and verify that each enrichment step provides useful information for downstream search scenarios.<\/span><\/p>\n

                                        Question 260<\/b><\/h3>\n

                                        A generative AI application needs to record model inputs, outputs, tool calls, and intermediate steps to investigate an unexpected response. Which capability is useful?<\/b><\/p>\n

                                          \n
                                        1. Tracing<\/span><\/li>\n
                                        2. Faceting<\/span><\/li>\n
                                        3. Image classification<\/span><\/li>\n
                                        4. Language detection<\/span><\/li>\n<\/ol>\n

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

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

                                          Tracing helps developers inspect the sequence of operations performed during an AI workflow. Depending on the application architecture, traces can provide visibility into prompts, model calls, tool invocations, intermediate processing, retrieved information, and resulting outputs. This information can help identify problems such as incorrect tool parameters, poor retrieval, unexpected model behavior, or failures between workflow components. Tracing is especially valuable in multi-step agentic applications where the final response may depend on several operations. Developers should ensure that trace data is handled securely because logs can contain user inputs, retrieved content, or other sensitive information.<\/span><\/p>\n

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

                                          View Full Microsoft AI-102 Exam Dumps and Practice Test Dumps.   Question 241 Which Microsoft Foundry capability helps developers test different prompts and model configurations against the same inputs? Prompt flow OCR Face detection Speech synthesis Correct Answer: 1 Explanation Prompt flow provides a structured way to develop, test, evaluate, and refine generative AI workflows. […]<\/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\/22411"}],"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=22411"}],"version-history":[{"count":1,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/22411\/revisions"}],"predecessor-version":[{"id":22412,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/22411\/revisions\/22412"}],"wp:attachment":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/media?parent=22411"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/categories?post=22411"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/tags?post=22411"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}