Practice Free AI-102 Designing and Implementing a Microsoft Azure AI Solution Exam Questions Answers With Explanation

The scenario involves storing text files, videos, audio streams, and virtual disk images , along with metadata and a unique ID. This exactly matches Object storage characteristics:

Stores large amounts of unstructured data (blobs).

Each object contains the data itself, metadata, and a unique identifier.

Azure Blob Storage is Microsoft’s object storage solution.

Other options:

Columnar ? Used for analytics workloads, optimized for reading columns (e.g., data warehouses).

Key/Value ? Good for simple lookups but not optimized for storing large multimedia files.

Document ? Stores semi-structured JSON documents, not large binary objects.

Thus, the correct answer is Object storage .

Microsoft Reference: Azure Blob Storage (Object storage) overview

Azure AI Document Intelligence offers pre-built models for common document types like invoices, which extract key fields (e.g., vendor, amount, date) with minimal setup. For custom forms like internal authorization requests, the pre-built general document model handles layout and text extraction, but to minimize development effort, use the pre-built invoice model where applicable and a custom model only if pre-built options fall short. However, the standard recommendation prioritizes pre-built for both to avoid training.

This scenario is about improving the accuracy of QnA Maker responses when user phrasing is different but semantically the same.

Problem

Uploaded document contains: “What is your warranty period?”

Users ask: “How long is the warranty coverage?”

QnA Maker fails because it does not automatically equate these two phrasings.

Both must map to the same answer.

Correct Sequence of Actions

Add alternative phrasing to the question and answer (QnA) pair.

In QnA Maker, you can add synonyms or paraphrases for a question.

Example: Add “How long is the warranty coverage?” as an alternative phrasing to the existing QnA pair.

This ensures the system treats both phrasings as the same question.

Retrain the model.

After modifying the knowledge base with new alternatives, you must retrain the model.

This step incorporates the updates into the QnA Maker service.

Republish the model.

To make the changes live and accessible by the chatbot endpoint, the retrained model must be republished.

Without republishing, the chatbot will still use the old knowledge base.

Actions That Are Not Required

Add a new QnA pair ? Not needed, because we want both questions to map to the same existing answer, not duplicate answers.

Add additional questions to the document ? That’s not required when QnA Maker already supports adding alternative phrasing within the portal or JSON editor.

Final Answer (Sequence)

Add alternative phrasing to the question and answer (QnA) pair.

Retrain the model.

Republish the model.

Microsoft References

QnA Maker – Add alternative phrasing

Train and publish a QnA Maker knowledge base

The system message you provided— “You are an AI assistant that helps people solve mathematical puzzles. Explain your answers as if the request is by a 4-year-old.” —is classic priming (also called system prompt/role prompting). Priming sets the assistant’s persona, objectives, tone, and constraints before any user message, steering all subsequent responses accordingly.

Why the other options are incorrect:

A. Few-shot learning – supplies example Q & A pairs in the prompt to teach the model by example; none were included here.

B. Affordance – relates to interface cues that guide users; not a prompt technique for model behavior.

C. Chain of thought – asks the model to show intermediate reasoning steps; your prompt doesn’t request step-by-step rationale, only a style/tone.

Microsoft Azure AI References

Azure OpenAI – Prompt engineering techniques (role/system prompts, priming).

Azure OpenAI – System message (role) guidance for controlling assistant behavior and tone.

? Provide access to AI1 by using: A bearer token

? Connect to the deployment by using: A deployment name

To ensure only your apps can access the Azure OpenAI resource, authenticate with Microsoft Entra ID (Azure AD) and obtain a Bearer token (ideally via managed identity ). This avoids sharing long-lived API keys and lets you scope access by app identity and Azure RBAC. Microsoft’s guidance recommends token-based auth for Azure OpenAI and shows how to acquire a bearer token/managed identity token for calls

When calling the REST API , Azure OpenAI requires you to specify the deployment name to select which model deployment to use. The REST route includes {deployment-id} (deployment name) in the path, e.g., .../openai/deployments/{deployment-id}/chat/completions. This is explicitly documented as a key difference from OpenAI’s public API, which takes only a model.

Therefore:

Access method: Bearer token (via Entra ID/managed identity)

Deployment selector: Deployment name in the REST path

Microsoft References

Authenticate to Azure OpenAI with Microsoft Entra ID / Managed identity (bearer token): guidance and how-to

Azure OpenAI request header recommends token-based authentication (Bearer).

Azure OpenAI requires deployment name in API calls; REST path uses

Analysis:

GPT-3.5 (A): A generative model; not specialized for embeddings ? ?

Embeddings (B): Purpose-built models (like text-embedding-ada-002) return numeric vector representations of text, ideal for similarity comparisons ? ?

DALL-E (C): Used for image generation, not semantic similarity ? ?

GPT-4 (D): Powerful LLM, but not efficient for embeddings tasks ? ?

Correct Answer:

B. embeddings

Comprehensive Detailed Explanation

You are deploying Anomaly Detector API as a container. The key requirements are:

Prevent billing and API keys from being exposed in command-line history.

Control access to container images using Azure RBAC.

Step-by-step reasoning:

Pull the Anomaly Detector container image

Microsoft Cognitive Services (including Anomaly Detector) provides container images via Microsoft Container Registry.

You must first pull the container image locally.

Push the image to an Azure container registry (ACR)

To enforce Azure RBAC, you need to use Azure Container Registry instead of Docker Hub.

Storing in ACR ensures authentication and role-based access control to the image.

Distribute a docker run script

Instead of providing credentials directly on the command line (which risks storing API keys in shell history), best practice is to distribute a docker run script (or use environment variables / mounted files) to run the container.

This prevents exposure of sensitive information.

Run the container on local devices or on-premises datacenters

The container now runs securely, with API key and billing info injected at runtime via environment variables or mounted secrets.

Incorrect Options

Create a custom Dockerfile ? Not needed, because Microsoft provides the official Anomaly Detector container image.

Build the image ? Not required; again, the container is prebuilt.

Push the image to Docker Hub ? Does not allow RBAC, so this would not meet the security requirement.

Correct Sequence:

Pull the Anomaly Detector container image.

Push the image to an Azure container registry.

Distribute a docker run script.

Run the container in the local or on-prem environment.

Microsoft References

Use Cognitive Services containers

Azure Container Registry – Authentication and RBAC

Best practices for running Cognitive Services containers

Azure Databricks is an Apache Spark-based analytics platform. ? Yes

Azure Analysis Services is used for transactional workloads. ? No

Azure Data Factory orchestrates data integration workflows. ? Yes

Let’s break down each statement:

Azure Databricks is an Apache Spark-based analytics platform.

True.

Azure Databricks is a fast, easy, and collaborative Apache Spark-based big data analytics platform .

It is designed for data engineering, machine learning, and AI workloads.

Answer: Yes

Azure Analysis Services is used for transactional workloads.

False.

Azure Analysis Services is an analytical service for semantic data modeling, OLAP (online analytical processing), and BI (business intelligence).

Transactional workloads belong to OLTP (online transaction processing) databases such as Azure SQL Database.

Answer: No

Azure Data Factory orchestrates data integration workflows.

True.

Azure Data Factory (ADF) is a cloud-based ETL and data integration service that orchestrates and automates data movement and transformation across on-premises and cloud systems.

Answer: Yes

Detect the incoming language:

Service: Speech to Text

Reasoning: The Speech to Text service can transcribe spoken language into text and includes language identification capabilities, allowing the system to detect whether the caller is speaking French or English.

Respond in the callers ' own language:

Service: Text to Speech

Reasoning: The Text to Speech service can convert text responses into spoken language in the detected language (French or English), enabling the system to respond to callers in their own language.

Ordered Mapping:

Detect the incoming language: Speech to Text

Respond in the callers ' own language: Text to Speech

Box 1: Text Analytics

The Language Detection feature of the Azure Text Analytics REST API evaluates text input for each document and returns language identifiers with a score that indicates the strength of the analysis.

Box 2: Translator

Translator is a cloud-based neural machine translation service that is part of the Azure Cognitive Services family of REST APIs. Translator can be used with any operating system and powers many Microsoft products and services used by thousands of businesses worldwide to perform language translation and other language-related operations.

Correct Sequence:

Add the prebuilt domain ToDo

Add a new application

Train the application

Publish the application

Comprehensive Detailed Explanation

You are building a Conversational Language Understanding (CLU) model for a chatbot that supports task tracking. To minimize development time, the best approach is to use a prebuilt domain (like ToDo) instead of building intents/entities manually.

Steps in correct order:

Add the prebuilt domain ToDo

Prebuilt domains in CLU provide ready-to-use intents and entities for common scenarios (e.g., calendar, email, ToDo).

This significantly reduces development time since you don’t need to create all intents/entities manually.

Add a new application

You must first create a CLU application that uses the prebuilt domain.

The application acts as the container for intents, entities, and utterances.

Train the application

Training is required after adding intents, entities, and utterances (or using a prebuilt domain).

This ensures the model can classify and extract information correctly.

Publish the application

Once trained, you must publish the application so it can be accessed via an endpoint and integrated into the chatbot.

Incorrect Actions in this Case

Add example utterances ? Not required if you’re using the prebuilt ToDo domain (utterances are already included).

Correct Sequence:

Add the prebuilt domain ToDo

Add a new application

Train the application

Publish the application

Microsoft References

Conversational Language Understanding (CLU) overview

Prebuilt domains in CLU (LUIS ToDo example)

Train and publish CLU models

To “recognize geographic locations” in Azure AI Search enrichment, add the EntityRecognitionSkill to your skillset. This built-in cognitive skill performs named entity recognition (NER) and extracts entities such as locations, people, and organizations from text.

EntityLinkingSkill links already-recognized entities to a knowledge base (e.g., Wikipedia) and is not required just to detect locations.

DocumentExtractionSkill parses file content/structure.

AzureOpenAIEmbeddingSkill creates vector embeddings, not entities.

Microsoft References

Azure AI Search – Entity recognition skill (extracts entities including locations).

Azure AI Search – Entity linking skill (disambiguates entities to a knowledge base; not for initial recognition).

For the category field in Azure Cognitive Search:

searchable ? allows users to search by words in the category field.

facetable ? enables drill-down filtering (facets).

retrievable ? ensures the field appears in search results.

filterable/sortable/key are not required here based on the scenario.

Correct Answer: A

Requirement: The chatbot must classify user input into categories .

The categories must be dynamic and defined at inference time .

Custom text classification (Language service) requires training beforehand and fixed labels.

OpenAI text classification can dynamically handle categories via prompting at inference.

Therefore, Azure OpenAI text classification is correct.

Microsoft Reference:

Classification with Azure OpenAI

This question aligns with Microsoft Responsible AI principles (Fairness, Reliability & Safety, Privacy & Security, Inclusiveness, Transparency, and Accountability).

The scenario:

Sentiment Analysis is used to determine staff bonuses (a financial impact).

Automated sentiment classification can be biased or produce incorrect results.

Decisions that affect people’s financial or legal status should never be made fully automated.

Analysis of the options:

A. Add a human review and approval step ? Correct. This ensures accountability and fairness, reducing harm from potential model bias or errors.

B. Include results even with low confidence ? Wrong. That increases risk of incorrect decisions.

C. Use all surveys including deleted accounts ? Wrong. Violates privacy and compliance.

D. Publish raw survey data ? Wrong. Violates privacy and security principles.

Therefore, the best responsible AI-compliant action is to add human review before making financial decisions.

Correct Answer for Q186: A

Microsoft Reference

Microsoft Responsible AI principles

Responsible use of AI in decision-making

You are securing an Azure Cognitive Search resource with two specific requirements:

Prevent access to Search1 from the internet

The correct way is to use a private endpoint.

A private endpoint connects the Cognitive Search resource directly into your virtual network (VNet), blocking public internet access.

Configuring an IP firewall can restrict access to specific IP ranges but does not fully prevent internet access.

Azure roles and authentication do not address network-level restrictions.

Limit the access of each app to specific queries

This is done via Azure role-based access control (RBAC).

You can assign different Azure roles or API keys with query rights to control what each app can do.

Key authentication alone only provides coarse-grained access (admin/query keys) but does not allow fine-grained per-app restrictions.

RBAC ensures principle of least privilege is enforced.

Correct Answer:

Prevent access from internet ? Create a private endpoint

Limit access to queries ? Use Azure roles

Microsoft References

Secure access to Azure Cognitive Search using private endpoints

Role-based access control in Azure Cognitive Search

To mask (redact) phone numbers and email addresses in text using Azure AI Language, you should call the PII entity recognition operation and then print the redacted text that the service returns. In the SDKs, this corresponds to calling RecognizePiiEntities and reading the RedactedText from the result object. The PII feature automatically detects categories such as PhoneNumber and Email and returns a redacted version of the input (masked) in the redactedText/RedactedText field. This is exactly what the code snippet’s “Masked Text” output is asking for.

Therefore:

Use RecognizePiiEntities to detect and redact PII in the document.

Output RedactedText to print the masked string.

Microsoft References

How-to: Detect and redact PII in text (overview and redaction behavior). Microsoft Learn

Quickstart: Detect Personally Identifiable Information (shows PII detection and redacted output). Microsoft Learn

NET reference: PiiEntityCollection.RedactedText property (“Gets the text of the input document with all PII redacted”). Microsoft Learn

Python client overview: recognize_pii_entities operation for PII detection. Microsoft Learn

Box 1: Adaptive card

Box 2: an image

Customer-managed key (CMK) encryption does not affect throttling.

Instead, you could migrate to a Cognitive Search service that uses a higher tier.

Note: A simple fix to most throttling issues is to throw more resources at the search service (typically replicas for query-based throttling, or partitions for indexing-based throttling). However, increasing replicas or partitions adds cost, which is why it is important to know the reason why throttling is occurring at all.

The given text:

" Our tour guide took us up the Space Needle during our trip to Seattle last week. "

The Language service output shows:

Text

Category

ConfidenceScore

Tour guide

PersonType

0.45

Space Needle

Location

0.38

Trip

Event

0.78

Seattle

Location

0.78

Last week

DateTime

0.80

Why the correct answer is Named Entity Recognition (NER):

NER identifies named entities (people, places, dates, events, organizations, etc.) in text and classifies them into categories with confidence scores.

In the table, we see categories such as PersonType, Location, Event, DateTime ? these are classic outputs of NER.

Why not the others:

A. Entity Linking ? Identifies entities and links them to a knowledge base (like Wikipedia). Output would include links/IDs rather than just categories.

C. Key Phrase Extraction ? Returns important phrases (“tour guide”, “Space Needle”, “trip”, “Seattle”, “last week”) but does not assign categories like PersonType, Location, DateTime.

D. Sentiment Analysis ? Returns sentiment scores (positive/neutral/negative), not entity categories.

Correct Answer:

B. Named Entity Recognition

Microsoft References

Named Entity Recognition in Azure AI Language

Entity Linking vs Named Entity Recognition

Answer Area

Open the Review suggestions pane.

This step allows you to access suggestions generated by the Azure AI Language service based on user interactions and queries. Reviewing these suggestions helps identify potential improvements or alternative phrasings that can enhance the model ' s understanding and accuracy with minimal effort.

Review and accept the alternative phrases.

After opening the review suggestions pane, you can evaluate and accept alternative phrases or questions that users might have asked. Accepting these updates trains the model to recognize variations in how questions are posed, thereby improving its ability to provide accurate responses without requiring extensive manual retraining.

Enable active learning.

Enabling active learning allows the model to continuously improve by learning from user interactions and feedback over time. This feature automatically suggests refinements based on real-world usage, minimizing development effort while enhancing the chatbot ' s accuracy as it adapts to common questions.

Rationale

These steps are sequenced to progressively improve the chatbot ' s performance with minimal development effort. Starting with reviewing suggestions leverages existing data, accepting alternative phrases refines the model ' s understanding, and enabling active learning ensures ongoing improvement, aligning with the goal of increasing accuracy efficiently.

Upload to: An Azure Storage account

Model type: Custom neural

You have 1,000 scanned hand-written survey responses with no consistent layout. The goal is to extract data while minimizing development effort.

Step 1: Upload location

For Azure AI Document Intelligence (formerly Form Recognizer), training data (images, PDFs) must be placed in a container in an Azure Storage account.

Options like Cosmos DB or Azure Files share are not valid training data sources for Document Intelligence.

Therefore, you upload the survey images to Azure Storage account.

Step 2: Model type

The surveys are handwritten and do not have a consistent layout.

Custom template models work only when the layout is fixed and consistent (e.g., invoices with the same fields in the same place).

Identity document (ID) model is a prebuilt model for passports, driver’s licenses, etc., not for surveys.

Custom neural model is designed to handle unstructured or variable-layout forms, including handwritten content. This best fits the survey scenario.

Thus, the correct configuration is:

Upload to Azure Storage account

Use Custom neural model

Correct Answer:

Upload to: An Azure Storage account

Model type: Custom neural

Microsoft References

Azure AI Document Intelligence – Custom models

Custom neural vs custom template models

Train a custom model with Azure Storage input

The task is to process incoming emails and route them to French or English support teams. This means we need to detect the language of each email.

Let’s break it down:

Which Cognitive Service is needed?

Azure Text Analytics (part of Azure AI Language) provides a Language Detection API.

The Translator service is used for translation, not just detecting the original language. Since the requirement is only to decide whether to send emails to French or English teams, we just need to detect the language.

Endpoint selection

Text Analytics uses regional endpoints such as:

https:// < region > .api.cognitive.microsoft.com/

For example: eastus.api.cognitive.microsoft.com is valid.

api.cognitivemicrosofttranslator.com ? incorrect (this is for Translator service).

portal.azure.com ? Azure portal, not an API endpoint.

API Path selection

To detect language, you call:

/text/analytics/v3.1/languages

/text/analytics/v3.1/entities/recognition/general ? entity recognition, not language.

/translator/text/v3.0/... ? translation, not language detection.

Correct Answer Combination:

eastus.api.cognitive.microsoft.com

/text/analytics/v3.1/languages

Microsoft References

Text Analytics Language Detection API

Text Analytics REST API v3.1

Scenario: Support autocompletion and autosuggestion based on all product name variants.

A: Call a suggester-enabled query, in the form of a Suggestion request or Autocomplete request, using an API. API usage is illustrated in the following call to the Autocomplete REST API.

POST /indexes/myxboxgames/docs/autocomplete?search & api-version=2020-06-30

{

" search " : " minecraf " ,

" suggesterName " : " sg "

}

B: In Azure Cognitive Search, typeahead or " search-as-you-type " is enabled through a suggester. A suggester provides a list of fields that undergo additional tokenization, generating prefix sequences to support matches on partial terms. For example, a suggester that includes a City field with a value for " Seattle " will have prefix combinations of " sea " , " seat " , " seatt " , and " seattl " to support typeahead.

F. Use the default standard Lucene analyzer ( " analyzer " : null) or a language analyzer (for example, " analyzer " : " en.Microsoft " ) on the field.

We need to design a Conversational Language Understanding (CLU) model for smart home commands.

The sample utterances:

“Set all the lights to on.” ? Intent = change setting, Entity = device (lights)

“Turn off the lights in the living room.” ? Intent = change setting, Entities = device (lights) + location (living room)

“What is the current thermostat temperature?” ? Intent = query setting, Entity = device (thermostat)

“Lower the temperature of the thermostat by five degrees.” ? Intent = change setting, Entity = device (thermostat)

Breakdown:

Intents needed:

Change setting intent (D): Covers actions like turning devices on/off, lowering temperature.

Query setting intent (E): Covers questions about current state, like “What is the thermostat temperature?”

Entities needed:

Device entity (F): Identifies devices such as lights and thermostat .

Why not the others:

A. a location intent ? Location (like “living room”) should be modeled as an entity, not an intent.

B. a change setting entity ? Change setting is an action (intent), not an entity.

C. a device intent ? Devices are entities, not intents.

Correct Answer:

D. a change setting intent

E. a query setting intent

F. a device entity

Microsoft References

Intents and entities in CLU

CLU model design best practices

In the Azure portal, the Keys and Endpoint blade for a Cognitive Service resource provides:

API keys (for authentication)

Endpoint URL (the REST interface base URL)

Identity blade relates to managed identities.

Properties blade shows metadata (resource ID, location, subscription ID).

Networking blade controls network access and firewall rules.

Correct Answer: B. Keys and Endpoint

This question is about selecting valid training files for a custom Form Recognizer model in Azure AI Document Intelligence (formerly Form Recognizer).

Supported File Formats for Training

According to Microsoft documentation, Form Recognizer supports the following for training and analysis:

PDF (? 500 pages, ? 50 MB for training)

JPEG/JPG (? 50 MB)

PNG (? 50 MB)

TIFF (? 50 MB)

Unsupported:

MP4 ? Video not supported.

GIF ? Not supported.

Maximum File Size Rules (for training input files):

Images (JPG, PNG, TIFF): ? 50 MB

PDFs: ? 50 MB

Note: 500 pages limit applies to PDFs.

File-by-File Analysis

File1 – PDF – 20 MB

Type supported.

Size within limit ( < 50 MB).

Valid ?

File2 – MP4 – 100 MB

MP4 is not supported.

Invalid ?

File3 – JPG – 20 MB

Type supported.

Size within limit ( < 50 MB).

Valid ?

File4 – PDF – 100 MB

Type supported.

Size exceeds the 50 MB maximum limit.

Invalid ?

File5 – GIF – 1 MB

GIF format not supported.

Invalid ?

File6 – JPG – 40 MB

Type supported.

Size within limit ( < 50 MB).

Valid ?

Correct Answer:

A. File1, C. File3, F. File6

Microsoft References

Azure AI Document Intelligence (Form Recognizer) – Supported file formats

Train a custom model with Azure Form Recognizer

The percentage of false positives is [0].

The value for the number of true positives divided by the total number of true positives and false negatives is [100].

The screenshot shows the Performance tab of a Custom Vision object detection model after training.

Two key performance metrics are in focus:

False Positives (FP):

False positives occur when the model incorrectly predicts an object that isn’t present.

In the provided exhibit, the model shows 0% false positives.

This means it did not incorrectly classify any non-object as an object.

Recall (True Positives ÷ (True Positives + False Negatives)):

This measures how many of the actual objects were correctly detected.

Formula: Recall = TP / (TP + FN).

In the performance report, recall is 100%, meaning the model correctly detected all the actual objects (no false negatives).

Evaluating the answer choices:

False positives: 0%

Recall (TP ÷ (TP + FN)): 100%

Correct Answers:

False positives ? 0

Recall ? 100

Microsoft References

Custom Vision performance metrics

Precision, recall, and accuracy in machine learning

In Power BI, reports are interactive, visual, and support KPIs, drill-through, and drill-down functionality.

Power Apps is for app development, not analytics.

A view is a database object, not a visualization solution.

Dataflow is for ETL/ELT (data preparation), not reporting.

Correct Answer: A

To access multiple Azure AI capabilities (e.g., Speech and Language) using a single endpoint and a single set of credentials, you should provision an Azure AI multi-service resource (called Azure AI service in the portal; formerly “Cognitive Services” multi-service account). This resource provides one endpoint and one key (or Microsoft Entra ID) that authorizes calls across supported services, including Speech and Language. Creating separate “Azure AI Speech” or “Azure AI Language” resources would give you distinct endpoints/keys for each, which does not meet the requirement. Azure AI Content Safety is unrelated.

Microsoft References

Azure AI services multi-service resource (single endpoint/credential across services).

https://learn.microsoft.com/azure/ai-services/multi-service-resource

What are Azure AI services (overview of consolidated access model).

https://learn.microsoft.com/azure/ai-services/what-are-ai-services

Explanation

To evaluate a Conversational Language Understanding (CLU) model, you can:

Use the REST authoring API to retrieve the model’s evaluation summary (precision, recall, F1 score).

In Language Studio, use the Model performance section to view evaluation metrics and confusion matrix.

Active Learning (option B) is for improving the model post-deployment by validating unclear utterances, not pre-deployment evaluation.

Log collection in Log Analytics (option D) is for monitoring, not formal evaluation.

Correct Answer: A and C

Comprehensive Detailed Explanation

The question focuses on monitoring a public-facing sales system that processes video and text, ensuring that the system provides equitable results regardless of user location or background.

Two Microsoft Responsible AI principles apply directly here:

Fairness

AI systems should treat all people fairly and avoid bias.

Monitoring ensures that outcomes do not disadvantage people based on location, background, or demographics.

This directly addresses the requirement for equitable results.

Inclusiveness

AI systems should empower everyone and engage people of all backgrounds.

Monitoring for inclusiveness ensures that the system works well across different regions, languages, and cultures, making the system accessible and beneficial for all users.

Why not the other options?

A. Transparency

Important for explaining how AI systems work, but does not directly address equitable results across demographics.

D. Reliability and safety

Focuses on consistent and safe system operation, not equity across users.

E. Privacy and security

Critical for data protection, but not related to equitable treatment of users.

Correct Answer:

B. fairness

C. inclusiveness

Microsoft References

Microsoft Responsible AI Principles

Fairness in AI systems

Inclusiveness in AI systems

According to the Microsoft documentation, Computer Vision is a cloud-based service that provides developers with access to advanced algorithms for processing images and returning information. By uploading an image or specifying an image URL, Computer Vision algorithms can analyze visual content in different ways based on inputs and user choices.

According to the Microsoft documentation, image type detection is one of the features of Computer Vision that can categorize an image as either a photograph or a drawing. You can use the image type detection feature by calling the Analyze Image API with the visualFeatures parameter set to ImageType. The API will return a JSON response with an imageType field that indicates whether the image is a photo or a clipart.

According to the Microsoft documentation, image descriptions is another feature of Computer Vision that can generate a caption for an image. You can use the image descriptions feature by calling the Analyze Image API with the visualFeatures parameter set to Description. The API will return a JSON response with a description field that contains a list of captions for the image, each with a confidence score.

Therefore, by using these two features of Computer Vision, you can achieve your app requirements with minimal development effort. You don’t need to use any other services, such as object detection, content tags, or Custom Vision, which are designed for different purposes.

The requirement is to extract structured information (invoice items, sales amounts, customer details) from scanned invoice images.

Form Recognizer (now part of Azure AI Document Intelligence) is specifically designed for extracting structured data from documents such as invoices, receipts, and business forms.

Custom Vision ? used for image classification or object detection, not for text/data extraction.

Computer Vision ? can extract text with OCR, but it doesn’t provide structured field-level extraction for invoices.

Immersive Reader ? for accessibility (reading content aloud, translating text), not document data extraction.

Correct Answer: D

For a social media messaging app, you need real-time language identification on short text messages. The Azure AI Translator service provides a dedicated Detect API that returns the language and a confidence score for each input text, designed for low-latency scenarios such as chat and in-app communication. This aligns directly with “identify in real time the language used in messages.” Microsoft Azure +2

Why the other options are not the best fit:

Azure AI Speech (A) targets spoken audio (speech-to-text, translation, diarization). It’s not intended for detecting the language of text messages. Azure AI

Azure AI Content Safety (B) focuses on safety/moderation (harmful or unsafe content), not language identification.

Azure AI Language (D) also offers a Language Detection feature, but it’s typically used within broader text analytics workflows. For real-time messaging and potential downstream translation, Translator’s Detect endpoint is the targeted and streamlined choice.

References (Microsoft Learn)

Translator – Detect method (returns detected language + confidence). Microsoft Learn

Translator – Product overview and real-time translation scenarios. Microsoft Azure

Translator – Language support (Auto Language Detection). Microsoft Learn

Azure AI Language – Language Detection (feature overview/how-to) for context.

=========================

Parameter type: stream

Visual features list: VisualFeatureTypes.Description

Result to use: results.Description.Captions[0]

To generate accessible alt text , you should use the caption produced by Azure Computer Vision’s Description feature (it produces a human-readable sentence with a confidence score). Therefore:

The image input should be a stream , because you’re uploading images (not just passing URLs) during product creation and AnalyzeImage…Async supports image streams .

Request the Description feature in VisualFeatureTypes so the service returns results.Description.Captions.

Use results.Description.Captions[0] and return the caption text if its confidence is high enough (e.g., > 0.5) to meet the accessibility requirement that all images must have relevant alt text .

Other features (Tags, Objects, Brands) are useful for enrichment but do not directly return natural-language captions suitable for alt text.

Microsoft Azure AI Solution References

Computer Vision (Image Analysis) – Description/Captions and features: Microsoft Docs, Image Analysis – VisualFeatureTypes.Description returns description.captions.

https://learn.microsoft.com/azure/ai-services/computer-vision/concept-image-analysis

SDK usage (analyze image from a stream): Microsoft Docs, Analyze an image by using the Computer Vision client library .

https://learn.microsoft.com/azure/ai-services/computer-vision/how-to/call-analyze-image?tabs=version-3-2#analyze-an-image-from-a-stream

Comprehensive Detailed Explanation

The requirement is to build an Adaptive Card for the chatbot that supports multilingual product display and meets accessibility requirements.

Product Name Display

The product JSON stores localized names like " name " : { " en " : " ... " , " es " : " ... " , " pt " : " ... " } .

To dynamically render based on the current language, you must use name[language].

This ensures the product name shown in the card adapts automatically to the user’s preferred language.

Stock Level Warning

Business rules require warnings when stock is low or out of stock.

This means you display the stock warning only when stockLevel is not ' OK ' .

Therefore, the correct condition is:

" $when " : " ${stockLevel != ' OK ' } "

Image Alt Text for Accessibility

Accessibility requirements specify that all images must have alt text in English, Spanish, and Portuguese.

In the product JSON, " alttext " is also localized:

" altText " : { " en " : " Bicycle " , " es " : " Bicicleta " , " pt " : " Bicicleta " }

To render the correct localized alt text dynamically, use image.altText[language].

Correct Selections:

First blank: name[language]

Second blank: " $when " : " ${stockLevel != ' OK ' } "

Third blank: image.altText[language]

Microsoft References

Adaptive Cards Templating

Adaptive Card $when property

Multilingual data in Azure Cognitive Search and Adaptive Cards

Box 1: api-nam.cognitive.microsofttranslator.com

https://docs.microsoft.com/en-us/azure/cognitive-services/translator/reference/v3-0-reference

Box 2: /translate

Box 1: Azure Blob storage

At the start of the pipeline, you have unstructured text or non-text content (such as images, scanned documents, or JPEG files). Data must exist in an Azure data storage service that can be accessed by an indexer.

Box 2: Computer Vision API

Scenario: Provide users with the ability to search insight gained from the images, manuals, and videos associated with the products.

The Computer Vision Read API is Azure ' s latest OCR technology (learn what ' s new) that extracts printed text (in several languages), handwritten text (English only), digits, and currency symbols from images and multi-page PDF documents.

Box 3: Translator API

Scenario: Product descriptions, transcripts, and all text must be available in English, Spanish, and Portuguese.

Box 4: Azure Files

Scenario: Store all raw insight data that was generated, so the data can be processed later.

Scenario: When the response confidence score is low, ensure that the chatbot can provide other response options to the customers.

When no good match is found by the ranker, the confidence score of 0.0 or " None " is returned and the default response is " No good match found in the KB " . You can override this default response in the bot or application code calling the endpoint. Alternately, you can also set the override response in Azure and this changes the default for all knowledge bases deployed in a particular QnA Maker service.

Choosing Ranker type: By default, QnA Maker searches through questions and answers. If you want to search through questions only, to generate an answer, use the RankerType=QuestionOnly in the POST body of the GenerateAnswer request.

Can search content in different formats, including video

Audio and video insights (multi-channels). When indexing by one channel, partial result for those models will be available.

Keywords extraction: Extracts keywords from speech and visual text.

Named entities extraction: Extracts brands, locations, and people from speech and visual text via natural language processing (NLP).

Topic inference: Makes inference of main topics from transcripts. The 2nd-level IPTC taxonomy is included.

Artifacts: Extracts rich set of " next level of details " artifacts for each of the models.

Sentiment analysis: Identifies positive, negative, and neutral sentiments from speech and visual text.

C: Analyze Receipt - Get Analyze Receipt Result.

Query the status and retrieve the result of an Analyze Receipt operation.

Request URL: https://{endpoint}/formrecognizer/v2.0-preview/prebuilt/receipt/analyzeResults/{resultId}

E: POST {Endpoint}/vision/v3.2/read/analyze

Use this interface to get the result of a Read operation, employing the state-of-the-art Optical Character Recognition (OCR) algorithms optimized for text-heavy documents.

Scenario: Contoso plans to develop a document processing workflow to extract information automatically from PDFs and images of financial documents

The document processing solution must be able to process standardized financial documents that have the following characteristics:

- Contain fewer than 20 pages.

- Be formatted as PDF or JPEG files.

- Have a distinct standard for each office.

*The document processing solution must be able to extract tables and text from the financial documents.

The document processing solution must be able to extract information from receipt images.

Box 1: " categories " : [ " Locations " , " Persons " , " Organizations " ],

Locations, Persons, Organizations are in the outputs.

Scenario: Contoso plans to develop a searchable knowledgebase of all the intellectual property

Note: The categories parameter is an array of categories that should be extracted. Possible category types: " Person " , " Location " , " Organization " , " Quantity " , " Datetime " , " URL " , " Email " . If no category is provided, all types are returned.

Box 2: { " name " : " entities " }

The include wikis, so should include entities in the outputs.

Note: entities is an array of complex types that contains rich information about the entities extracted from text, with the following fields

name (the actual entity name. This represents a " normalized " form)

wikipediaId

wikipediaLanguage

wikipediaUrl (a link to Wikipedia page for the entity)

etc.

The requirements state:

All AI solution responses must have a confidence score ? 70%.

If the response confidence score is < 70%, the response must be improved by human input.

Members of the Consultant-Bookkeeper group must be able to process financial documents, which includes performing manual review when the AI confidence is below threshold.

Looking at the provided JSON:

" fields " : {

" ReceiptType " : {

" type " : " string " ,

" valueString " : " Itemized " ,

" confidence " : 0.672

},

" MerchantName " : {

" type " : " string " ,

" valueString " : " Tailwind " ,

" confidence " : 0.913

}

}

ReceiptType.confidence = 0.672 ? this is below 0.7.

MerchantName.confidence = 0.913 ? this is above 0.7.

Therefore, the correct condition for triggering manual review is:

documentResults.fields. < field > .confidence < 0.7

Now, evaluating the options:

A. documentResults.docType == " prebuilt:receipt "

Always true for receipts, does not check confidence. Not correct.

B. documentResults.fields.*.confidence < 0.7

This is the correct general expression: trigger manual review whenever any field confidence is below 0.7.

C. documentResults.fields.ReceiptType.confidence > 0.7

This would bypass manual review when ReceiptType has high confidence. The requirement is to trigger review when confidence < 0.7, so this is the opposite.

D. documentResults.fields.MerchantName.confidence < 0.7

This only checks one field (MerchantName). In the JSON, MerchantName has confidence 0.913 ( > 0.7), so this condition would not trigger, but ReceiptType clearly needs review. Too narrow.

Correct Answer: B. documentResults.fields.*.confidence < 0.7

Microsoft References

Azure AI Document Intelligence – Confidence scores

Human-in-the-loop for Document Intelligence

Within a search service, synonym maps are a global resource that associate equivalent terms, expanding the scope of a query without the user having to actually provide the term. For example, assuming " dog " , " canine " , and " puppy " are mapped synonyms, a query on " canine " will match on a document containing " dog " .

Create synonyms: A synonym map is an asset that can be created once and used by many indexes.

Step 1: Create a new project and load sample documents

Create a new project. Projects store your configurations and settings.

Step 2: Label the sample documents

When you create or open a project, the main tag editor window opens.

Step 3: Train a custom model.

Finally, train a custom model.

The wiki contains text in English, French and Portuguese.

Scenario: All planned projects must support English, French, and Portuguese.

The Document Extraction skill extracts content from a file within the enrichment pipeline. This allows you to take advantage of the document extraction step that normally happens before the skillset execution with files that may be generated by other skills.

Note: The Translator Text API will be used to determine the from language. The Language detection skill is not required.

Box 1: Cognitive Service User

Ensure that the members of a group named Management-Accountants can approve the FAQs.

Approve=publish.

Cognitive Service User (read/write/publish): API permissions: All access to Cognitive Services resource except for ability to:

1. Add new members to roles.

2. Create new resources.

Box 2: Cognitive Services QnA Maker Editor

Ensure that the members of a group named Consultant-Accountants can create and amend the FAQs.

QnA Maker Editor: API permissions:

1. Create KB API

2. Update KB API

3. Replace KB API

4. Replace Alterations

5. " Train API " [in new service model v5]

Box 3: Cognitive Services QnA Maker Read

Ensure that the members of a group named the Agent-CustomerServices can browse the FAQs.

QnA Maker Read: API Permissions:

1. Download KB API

2. List KBs for user API

3. Get Knowledge base details

4. Download Alterations

Generate Answer

Scenario: All planned projects must support English, French, and Portuguese.

If a bot uses multiple LUIS models and QnA Maker knowledge bases (knowledge bases), you can use the Dispatch tool to determine which LUIS model or QnA Maker knowledge base best matches the user input. The dispatch tool does this by creating a single LUIS app to route user input to the correct model.