Microsoft AI-102 Designing and Implementing a Microsoft Azure AI Solution Exam Practice Test

In user want to change to deploy offline model

1. Change model domain to compact model

2. Retrain compact model

3. Export model

Obtain values to connect your bot to the knowledge base

1.In the QnA Maker site, select your knowledge base.

2.With your knowledge base open, select the SETTINGS tab. Record the value shown for service name. This value is useful for finding your knowledge base of interest when using the QnA Maker portal interface. It ' s not used to connect your bot app to this knowledge base.

3.Scroll down to find Deployment details and record the following values from the Postman sample HTTP request:

4.POST /knowledgebases/ < knowledge-base-id > /generateAnswer

5.Host: < your-host-url >

6.Authorization: EndpointKey < your-endpoint-key >

You need to prepare training data for Azure OpenAI custom model fine-tuning using the OpenAI CLI data preparation tool .

Supported input formats: JSONL (primary).

CLI pre-processing accepted formats: TSV, CSV, JSON.

Not supported: PDF, XML, XLSX directly.

Checking files:

File1.tsv (80 MB) → Supported (CLI can convert TSV to JSONL).

File2.xml (25 MB) → Not supported (XML not accepted).

File3.pdf (50 MB) → Not supported.

File4.xlsx (200 MB) → Not supported directly.

Correct Answer: A. File1.tsv only

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 Text Moderation API of Azure Content Moderator is specifically used for scanning text to detect potentially inappropriate or undesired content.

Key features include:

Adult classification score → Indicates the likelihood that the text contains adult content.

Racy classification score → Indicates the likelihood that the text contains sexually suggestive but not explicit content.

PII detection (personal data) is part of text moderation but is not always required when focusing strictly on enabling content moderation for chatbot text safety.

OCR is for extracting text from images, not for moderating chatbot conversations.

Text classification in this context refers to general machine learning categorization (not part of Content Moderator’s moderation API).

Thus, the two service responses directly tied to text moderation safety are:

Adult classification score

Racy classification score

Correct Answer for Q185: A and E

Microsoft Reference

Azure Content Moderator text moderation overview

The D in ACID stands for Durability .

Durability guarantees that once a transaction is committed, the changes are permanent, even in the case of power loss, crashes, or OS restarts.

Atomicity = all-or-nothing.

Consistency = valid state transitions.

Isolation = transactions run independently without interference.

Correct Answer: A. durability

With Azure OpenAI , you don’t call a raw “model name” directly; you call a deployment of that model. SDKs and REST require:

the resource endpoint ,

the API key (or Entra ID token), and

the deployment name to target the specific model instance.

References

Azure OpenAI quickstart (connect using endpoint, key, and deployment name ).

https://learn.microsoft.com/azure/ai-services/openai/quickstart

Chat/completions—path uses /deployments/{deployment-id}.

https://learn.microsoft.com/azure/ai-services/openai/reference#chat-completions

Let’s carefully analyze the JSON output from the exported LUIS (Language Understanding) model.

The utterance is:

" average amount of rain by month at chicago last year "

The extracted entities are:

Weather.WeatherRange → " average " (positions 0–6)

Weather.WeatherCondition → " rain " (positions 18–21)

Weather.Historic → " by month " (positions 23–30)

We need to determine what text fragment corresponds to Weather.Historic .

Looking at the " startPos " : 23, " endPos " : 30 , these indexes map to the substring " by month " in the utterance.

Why not the other options?

B. chicago → Refers to a location entity, but the entity name is Weather.Location (not shown here).

C. rain → Already captured as Weather.WeatherCondition .

D. location → Would also be “chicago,” but again that’s a different entity type.

So the only correct match for Weather.Historic is " by month " .

Correct Answer: A. by month

Microsoft References

Entities in LUIS

Export LUIS model schema

Labeling utterances with entities

Azure Synapse Analytics (formerly Azure SQL Data Warehouse) uses a massively parallel processing (MPP) architecture for big data and analytics workloads.

Control Node:

Acts as the brain of the system.

Receives queries, optimizes them, and then coordinates execution.

Does not perform large-scale data crunching itself.

Compute Nodes:

Perform the actual parallel execution of queries.

Data is distributed across these nodes.

The MPP engine breaks down queries into smaller tasks and distributes them to compute nodes for fast, parallel execution.

Review of Answer Options:

Distributes processing across compute nodes → Correct. This is the essence of MPP.

Distributes processing across control nodes → Incorrect. The control node does not process queries in parallel; it coordinates them.

Redirects client connections across compute nodes → Incorrect. Client connections go to the control node, not directly to compute nodes.

Redirects client connections across control nodes → Incorrect. There is typically one control node per Synapse instance.

Correct Answer: Distributes processing across compute nodes

Microsoft References

Azure Synapse Analytics architecture

Distributed query processing in Azure Synapse

Azure AI Agent Service projects are created and managed in Azure AI Foundry (ai.azure.com). Foundry provides the “project” workspace where you configure agents, connect tools (like speech and search), and use your Azure OpenAI deployments. Language Studio and Speech Studio are for the classic Language and Speech services respectively, and the Azure portal is used to create resources—not the agent project itself. The Microsoft docs explicitly state you start by creating an Azure AI Foundry project to build agents.

References

What is Azure AI Foundry Agent Service? – “To get started… create an Azure AI Foundry project.” Microsoft Learn

Create a project in Azure AI Foundry (projects organize agents, files, evaluations). Microsoft Learn

Quickstart: Create an agent in Azure AI Foundry Agent Service. Microsoft Learn

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

When an indexer runs in Azure AI Search, it processes each document through a fixed series of stages:

Document cracking

The indexer opens the item (file/row) and extracts text, images, and metadata, creating the root of the enriched document (/document). This is always the first stage.

Field mappings

Next, source fields are mapped directly to target index fields before any transformations . Field mappings occur after document cracking but before skillset execution . Microsoft Learn

Skillset execution

Built-in or custom skills (for example, language and sentiment analysis in your scenario) run and write outputs into the in-memory enriched document. Microsoft Learn

Output field mappings

Because skill outputs live only in memory, you must define output field mappings to route selected nodes of the enriched document into index fields. This step comes after skillset execution.

Push to index

Finally, the indexer hands off raw and enriched content to the search index (the physical data structures) so it becomes searchable—i.e., the content is “pushed to the index.” Microsoft Learn

This order— document cracking → field mappings → skillset execution → output field mappings → push to index —ensures raw content is extracted, direct source-to-index mappings are applied, AI enrichment is performed, enriched outputs are projected to fields, and the completed documents are written into the index.

Microsoft Azure AI References

Indexers in Azure AI Search – Stages of indexing (Document cracking; Field mappings; Skillset execution; Output field mappings, with field mappings occurring before transformations). Microsoft Learn

Skillset concepts (Output of skillset is routed to the index via output field mappings after skillset execution). Microsoft Learn

AI enrichment concepts (Field mappings move raw source content to the index; output field mappings move enriched content; indexing ingests raw and enriched content into the physical index). Microsoft Learn

Annotations & enriched document tree (Enriched document is created during document cracking at /document). Microsoft Learn

The provided code uses the Azure AI Language – Text Analytics client to call the ExtractKeyPhrases method.

The method ExtractKeyPhrases(text) identifies the most important words and phrases in the input text.

The input string is:

" the quick brown fox jumps over the lazy dog "

The ExtractKeyPhrases API removes stopwords such as " the, " " over, " and " jumps " and extracts only meaningful keywords.

Therefore, the returned list of key phrases will include:

" Quick brown fox "

" Lazy dog "

When printed, the console output after " Key phrases: " will show:

Quick brown fox

Lazy dog

This matches option C (Quick brown fox lazy dog), since the key phrases are concatenated as output lines.

Correct Answer: C. Quick brown fox lazy dog

Microsoft References

Azure AI Language – Key phrase extraction

TextAnalyticsClient.ExtractKeyPhrases method (Azure.AI.TextAnalytics)

Your app uses the primary admin key for index management. If it’s suspected compromised, the least-downtime rotation is:

Regenerate the secondary admin key to ensure it’s known only to you.

Switch the app to use the secondary key (no downtime—primary still works until you cut it off).

Regenerate the primary key to invalidate the compromised credential.

Query keys are read-only and cannot manage indexes, so switching to a query key would not meet the requirement. Regenerating the primary first would break the app until you update it.

References (Microsoft Azure Cognitive Search)

Admin (primary/secondary) vs. query keys and key rotation; key regeneration invalidates only the selected key.

Guidance to use the alternate admin key while rotating the other to avoid downtime.

When deploying an Azure OpenAI model using the Microsoft.CognitiveServices/accounts/deployments resource in an ARM template, the throughput for a Standard deployment is set on the SKU object. Specifically:

Use the capacity field of the SKU to specify the per-deployment rate.

Provide the integer value that represents the requests per minute (RPM) you want the deployment to handle.

Therefore, to meet the requirement of 600 requests per minute , complete the template as:

" sku " : {

" name " : " Standard " ,

" capacity " : 600

}

This uses a single deployment SKU with a defined capacity. The ARM schema for accounts/deployments explicitly includes sku.capacity for resources that support scale (Azure OpenAI deployments do), and Azure OpenAI quota/rate-limit guidance confirms that rate limits are assigned per deployment.

Microsoft Azure AI Solution References

ARM template—Microsoft.CognitiveServices/accounts/deployments (2023-05-01) : property reference showing sku.capacity. Microsoft Learn

Resource type reference—latest accounts/deployments : confirms SKU usage on deployments. Microsoft Learn

Azure OpenAI quota and rate limits : explains assigning deployment-level limits (RPM/TPM). Microsoft Learn

semi-structured data

The statement is:

" Varying fields for each entity in a JSON document is an example of ______. "

Relational data → Stored in structured tables with fixed schema (e.g., SQL database). Not applicable here.

Structured data → Has a strict schema (columns and types are predefined). JSON allows variation, so not structured.

Semi-structured data → Data that does not require a fixed schema, but still uses markers (tags, keys) to separate elements. JSON and XML are classic examples. ✅

Unstructured data → Data without schema or labeling (e.g., videos, images, free text). Not JSON.

Since JSON documents can have varying fields per entity (flexible schema), this is considered semi-structured data.

Correct Answer: semi-structured data

Microsoft References

Structured, semi-structured, and unstructured data

JSON in Azure

Creating an entity for price is a feature of Conversational Language Understanding (CLU/LUIS), not Question Answering. Custom question answering does not use entities for extraction or intent resolution; it matches user utterances to Q/A pairs using alternates, synonyms, and ranking. Therefore, defining an “entity for price” would not make the chatbot correctly answer “How much does < product > cost?”. The correct fix is to author additional alternate phrasings (as in Q266).

References

Custom question answering capabilities (no entity modeling).

https://learn.microsoft.com/azure/ai-services/language-service/question-answering/overview

CLU entities vs. QnA authoring—separate features and workflows.

https://learn.microsoft.com/azure/ai-services/language-service/conversational-language-understanding/overview

In Custom Question Answering (part of Azure AI Language), if you want to define alternative words or spellings for a specific term (e.g., testing , trials , evaluate ), you must use the alterations property.

Each alteration object contains a list of alternative values for the same meaning.

Example:

{

" alterations " : [

{

" alterations " : [

" testing " ,

" trials " ,

" evaluate "

]

}

]

}

phrases is used for pattern matching in question-answer pairs.

synonyms is used for synonyms within LUIS (Language Understanding), not for QnA alterations.

value is used inside synonyms definitions in some contexts but not here.

Correct JSON Payload Completion:

The JSON should be:

{

" alterations " : [

{

" alterations " : [

" testing " ,

" trials " ,

" evaluate "

]

}

]

}

Verified Answer:

Left box: " alterations "

Right box: " alterations "

Microsoft References:

Manage alterations in Custom Question Answering (Azure AI Language)

Authoring API reference for alterations

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

The requirement is to extract text from scanned images (OCR at scale).

Computer Vision Image Analysis can describe and tag images but is not optimized for bulk OCR.

Form Recognizer is intended for structured document extraction (invoices, receipts, forms), not raw magazine scans.

Azure Cognitive Service for Language works only with text, not images.

Read API in Computer Vision is specifically designed for extracting text from scanned images, supports multiple languages, and is ideal for this scenario.

Correct Answer: B

 HTTP Method: POST

 Visual Feature: imageType

Comprehensive Detailed Explanation

The question requires configuring an Azure AI Vision API request that determines if an image is clipart or a line drawing.

Step 1 – Determine the HTTP Method

The Analyze Image endpoint ( /analyze ) of Azure AI Vision expects an HTTP POST request.

GET is not valid for analysis with an image body.

PATCH is irrelevant here.

POST is correct because you are submitting image content or a URL for analysis.

You are using Azure Cognitive Search, where access is controlled by:

Admin keys: Provide full read/write control (never given to apps).

Query keys: Provide read-only access, meant for client apps.

Here, the query key has been compromised. You want to minimize downtime and ensure continued read-only access.

Step 1 – Add a new query key

First, generate a new query key from the Azure portal or via REST API.

This allows you to continue providing read-only access to authorized apps.

Step 2 – Change the app to use the new key

Update the application configuration so that it authenticates requests with the new query key.

This ensures uninterrupted service while preparing to retire the compromised key.

Step 3 – Delete the compromised key

Once the app is using the new key successfully, revoke the compromised query key.

This prevents further unauthorized usage.

Why not other options?

Regenerate primary/secondary admin key: Not necessary since only the query key is compromised, and admin keys give full control.

Change app to use secondary admin key: Insecure (admin keys should never be exposed to client apps).

Correct Sequence

Add a new query key.

Change the app to use the new key.

Delete the compromised key.

Microsoft References

Manage admin and query keys in Azure Cognitive Search

Create and use query keys

B: The Text Analytics API ' s Sentiment Analysis feature provides two ways for detecting positive and negative sentiment. If you send a Sentiment Analysis request, the API will return sentiment labels (such as " negative " , " neutral " and " positive " ) and confidence scores at the sentence and document-level.

D: 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.

This capability is useful for content stores that collect arbitrary text, where language is unknown. Reference:

https://docs.microsoft.com/en-us/azure/cognitive-services/text-analytics/how-tos/text-analytics-how-to- sentiment-analysis?tabs=version-3-1

https://docs.microsoft.com/en-us/azure/cognitive-services/text-analytics/how-tos/text-analytics-how-to- language-detection

The scenario:

You are building an in-vehicle application that uses Azure AI Speech for text-to-speech (TTS).

Requirement: Optimize the quality of synthesized voice output.

You must choose the most appropriate SSML (Speech Synthesis Markup Language) attribute.

Option A – style attribute of the mstts:express-as element

Used to apply expressive styles like " cheerful, " " empathetic, " or " angry. "

Helpful for emotion/tone, but not specifically about audio quality for clarity in motor vehicles.

Option B – level attribute of the emphasis element

Controls how strongly certain words are emphasized.

Improves meaning and focus, but does not optimize overall voice quality.

Option C – pitch attribute of the prosody element ✅

The < prosody > element allows adjustments to pitch, rate, and volume.

Adjusting pitch directly affects the naturalness and clarity of the synthesized speech.

Especially useful in noisy environments such as cars, because modifying pitch can make speech stand out better against background noise.

This directly improves perceived quality of voice output.

Option D – effect attribute of the voice element

Used for special voice effects (like whispering).

Not typically used for general clarity or quality improvements.

Correct Answer: C. the pitch attribute of the prosody element

Microsoft References

SSML support in Azure AI Speech

Prosody element in SSML

Your app currently calls the prebuilt Business Card model v2.1. In v2.1, the platform does not provide QR/barcode interpretation. Support for decoding barcodes (including QR codes) was introduced with the v3.x API family as an add-on capability ( ocr.barcode ) that you can enable on analysis requests in API versions 2023-07-31 (GA) and later. Therefore, the first step is to upgrade from v2.1 to v3.0 so your application can use the newer API surface that supports barcode/QR extraction. This meets the “minimize administrative effort” requirement because you keep the same Azure AI Document Intelligence resource, keys, and endpoint—your change is primarily to the API/SDK version and request options.

After upgrading, you have two common implementation patterns:

Continue to use prebuilt Business Card for the contact fields, and in the same flow make an additional call with prebuilt Layout (or General Document) with the ocr.barcode add-on enabled to decode any QR code present on the card. This is the recommended path because the Business Card model itself doesn’t expose add-on features, while the platform’s v3.x analysis APIs do support ocr.barcode on other models. Microsoft Learn

Why the other options are not correct:

A. Deploy a custom model – unnecessary effort; barcode/QR decoding is available out-of-the-box in v3.x via ocr.barcode . Microsoft Learn

B. Implement the read model – Read (OCR) extracts text but does not interpret barcodes/QR codes; you need the barcode add-on introduced in v3.x. Microsoft Learn

D. Implement the contract model – Irrelevant to business cards or QR decoding.

Key Microsoft References

Add-on capabilities (including ocr.barcode with QR support) — available in API 2023-07-31 (GA) and later. Microsoft Learn

Barcode/QR extraction behavior and schema (barcodes collection, kind, value, polygon). Microsoft Learn

Migration guide from v2.1 to v3.x (why upgrade and what changes). Azure Docs

Business Card model documentation and versioning context (v2.1/v3.0/v3.1). Microsoft Learn

SDK/blog announcement of barcode add-on in v3.x (what it returns and when it’s supported). Microsoft for Developers

Code samples showing the barcodes add-on usage. github.com

Comprehensive Detailed Explanation

You are tasked with building a solution using Custom Vision to detect faults in factory components. Custom Vision follows a defined workflow:

Create a project

Before anything else, you must create a Custom Vision project in the Azure AI portal.

This project specifies the type of model you are building: classification (classifying entire images) or object detection (locating items within images).

Since the question is about detecting faults in specific components (a business-specific scenario), it aligns with classification unless explicitly stated otherwise.

Upload and tag images

You must provide training data.

Upload a set of images showing both faulty and non-faulty components.

Tag each image correctly so that the model can learn from labeled examples.

Train the classifier model

Once images are tagged, you run training on the dataset to generate the classifier model.

The model can then be published and used for inference via API calls.

Why not the other options?

Initialize the training dataset: This is not a direct step in Custom Vision; instead, uploading and tagging images inherently prepares the training dataset.

Train the object detection model: Object detection is used if you need bounding boxes around items within an image. Since the question is about recognizing faults in a component (faulty vs non-faulty), a classifier model is appropriate.

Correct Sequence:

Create a project

Upload and tag images

Train the classifier model

Microsoft References

Quickstart: Build an image classification project with the Custom Vision portal

Custom Vision overview

Train and test a Custom Vision model

In the Microsoft Bot Framework SDK, bots communicate using activities (messages, events, conversation updates, etc.).

To respond to events (like conversation started, member added, or custom event activities), you use an ActivityHandler.

The ActivityHandler class provides methods you can override, such as OnMessageActivityAsync , OnMembersAddedAsync , or custom event handlers, to send custom text responses.

Other options:

Adaptive card (A) → used for rich UI responses, not for handling events.

Dialog (C) → used for managing conversation flow, not event handling.

Skill (D) → used to integrate another bot’s functionality, not for handling custom events.

Correct Answer: B

Text Extraction: b) Computer Vision

Computer Vision is a service that can analyze images and extract text from them using optical character recognition (OCR) or read API1. OCR can detect and extract printed or handwritten text from images in various languages and formats, such as PDF, TIFF, or JPEG2. Read API can perform asynchronous batch processing of text from multiple images, and can also handle text in tables, mixed languages, and rotated text3.

Profane Language Detection: c) Content Moderator

Content Moderator is a service that can detect and filter out potentially offensive or unwanted content in text, images, and videos. For text, Content Moderator can identify profane or abusive language, personal data, and custom terms that you define. You can also use Content Moderator to review and moderate the content manually or automatically using workflows.

Analysis:

Azure AI Bot Service (A): Helps build chatbots but does not store grounding data → ❌

Azure SQL (B): Possible storage, but not recommended for vector search grounding → ❌

Azure AI Document Intelligence (C): Extracts structured info, not used directly for grounding → ❌

Azure Blob Storage (D): Required for storing raw files used in retrieval-augmented generation (RAG) → ✅

Azure AI Search (E): Used to index, vectorize, and query data for grounding → ✅

Correct Answer:

D. Azure Blob Storage and E. Azure AI Search

https://learn.microsoft.com/en-us/azure/cognitive-services/language-service/named-entity-recognition/overview Named Entity Recognition (NER) is one of the features offered by Azure Cognitive Service for Language, a collection of machine learning and AI algorithms in the cloud for developing intelligent applications that involve written language. The NER feature can identify and categorize entities in unstructured text. For example: people, places, organizations, and quantities.

To minimize development effort in Azure AI Content Understanding, you should start from a video analyzer template that already targets large libraries and general-purpose metadata extraction. The Media asset management template is designed exactly for this scenario—organizing and extracting structured information from large collections of videos (news, marketing, broadcast, TV episodes, film archives). Templates are fully customizable, so you can add a field such as sentiment per video or per segment without building the schema from scratch. This approach yields the required sentiment analysis while keeping effort low.

“Video shot analysis” is not a Content Understanding template (shot detection is a capability inside the video analyzer), and the Advertising template is tuned for ad-specific fields (brand presence, ad categories), which is unnecessary for a generic video library. Using Media asset management aligns with the library use case and lets you add a sentiment field to the template’s schema.

Microsoft References

Analyzer templates (Video) — lists the available video templates including Media asset management and Advertising and notes that templates are customizable. Microsoft Learn

Video solutions overview — explains prebuilt video analyzer, shot detection (as a capability), and shows how to customize fields (example includes extracting fields such as sentiment per scene ). Microsoft Learn

Data Analyst.

When dealing with data roles in an organization, responsibilities are divided among data analysts, data engineers, and data scientists. Let’s review them:

Data Analyst

Focuses on understanding business requirements and ensuring that the right rules and logic are applied to data so it becomes meaningful for decision-making.

They work closely with stakeholders to define business rules such as data validation rules, aggregation logic, and reporting metrics.

This makes them responsible for identifying which business rules must be applied to data.

Data Engineer

Responsible for designing, building, and maintaining the infrastructure and pipelines that move and process data.

They ensure data availability, reliability, and scalability, but do not usually define business rules.

Data Scientist

Focuses on advanced analytics, predictive modeling, and machine learning.

They use data prepared by engineers and analysts but are not typically the ones who define the business rules that govern how raw data should be interpreted.

Therefore, the correct answer is Data Analyst.

Correct Answer: A data analyst

Microsoft References

Data roles and responsibilities in Azure

What is a Data Analyst?

Responsibilities of Data Engineers vs Data Scientists vs Data Analysts

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

You are analyzing an image with Azure AI Content Safety using the following request payload:

{

" image " : {

" content " : " < base_64_string > "

},

" categories " : [

" Violence "

],

" outputType " : " EightSeverityLevels "

}

Key Points:

Category specified → " Violence "

The Content Safety API will check for violent content in the image.

Test image description → A simple circle.

This is not violent in any way.

Output type → " EightSeverityLevels "

This setting returns a floating-point severity score from 0.0 (no severity) up to 7.0 (highest severity).

Expected result:

Since the image is only a circle (non-violent), the violence severity level will be at the lowest possible value: 0.0

Correct Answer: B. 0.0

Microsoft References

Azure AI Content Safety – Image moderation

Content Safety severity levels

The output will include the following words: our and included. → No

The output will include the following words: Paris, Eiffel, and Tower. → Yes

The function will output all the key phrases from the input string to the console. → No

Comprehensive Detailed Explanation

Let’s examine the function:

static void GetKeyWords ( TextAnalyticsClient textAnalyticsClient, string text)

{

var response = textAnalyticsClient.RecognizeEntities(text);

Console.WriteLine( " Key words: " );

foreach (CategorizedEntity entity in response.Value)

{

Console.WriteLine( $ " {entity.Text} " );

}

}

The function uses RecognizeEntities from the Azure Text Analytics SDK.

RecognizeEntities detects and categorizes named entities in text (such as locations, people, organizations, dates, etc.).

It does not extract all words or key phrases.

For extracting key phrases, you would use ExtractKeyPhrases , not RecognizeEntities .

Evaluating the statements:

" The output will include the following words: our and included. "

These are common words, not entities. They will not appear.

Answer: No

" The output will include the following words: Paris, Eiffel, and Tower. "

These are named entities (location, landmark). They will be recognized and returned.

Answer: Yes

" The function will output all the key phrases from the input string to the console. "

No. The function extracts entities, not key phrases. Key phrases would require ExtractKeyPhrases .

Answer: No

Comprehensive Detailed Explanation

You are working with a custom question answering project in Azure AI Language (formerly QnA Maker). The requirement is to enable the chatbot to engage in multi-turn conversations.

Multi-turn conversation: This occurs when a bot needs to ask clarifying questions or present follow-up options to guide the user toward the right answer.

Now let’s evaluate the options:

A. Add alternate questions

Alternate questions help improve query matching by allowing different phrasings of the same question.

This does not enable multi-turn dialogue.

B. Enable chit-chat

Chit-chat adds conversational personality (e.g., greetings, small talk).

Useful for user engagement but not for structured multi-turn dialogues.

C. Add follow-up prompts ✅

Follow-up prompts allow you to define next-step questions after an answer.

They are the key mechanism in custom question answering to enable multi-turn conversation flows.

Example: If the answer is about " subscriptions, " the follow-up prompts might include " How to upgrade? " or " How to cancel? "

D. Enable active learning

Active learning helps improve accuracy by suggesting alternative phrasing for questions based on user feedback.

This is for model refinement, not multi-turn conversation.

Thus, the correct way to configure the project for multi-turn conversations is to add follow-up prompts.

Correct Answer: C. Add follow-up prompts

Microsoft References

Custom question answering in Azure AI Language – Multi-turn conversations

Custom question answering overview

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

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

Box 2: /translate

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: 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.

Box 1: QnA Maker Editor

Scenario: Provide all employees with the ability to edit Q & As.

The QnA Maker Editor (read/write) has the following permissions:

Create KB API

Update KB API

Replace KB API

Replace Alterations

" Train API " [in

new service model v5]

Box 2: Contributor

Scenario: Only senior managers must be able to publish updates.

Contributor permission: All except ability to add new members to roles

Scenario: All videos must have transcripts that are associated to the video and included in product descriptions.

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

Step 1: Upload the video to blob storage

Given a video or audio file, the file is first dropped into a Blob Storage. T

Step 2: Index the video by using the Video Indexer API.

When a video is indexed, Video Indexer produces the JSON content that contains details of the specified video insights. The insights include: transcripts, OCRs, faces, topics, blocks, etc.

Step 3: Extract the transcript from the Video Indexer API.

Step 4: Translate the transcript by using the Translator API.

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.

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.

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.

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.

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.

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.

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.

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.

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

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

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.