AI-103, Microsoft Certification

AI-103: Develop AI Apps and Agents on Azure – Practice Exam #2 (30 questions with answers)

30 Practice Questions with Answers and Explanations

Question 1

You are designing an Azure AI solution that must automatically scale during periods of unpredictable traffic.

Which hosting model is MOST appropriate?

A. Dedicated on-premises servers
B. Static virtual machines only
C. Manual scaling groups
D. Serverless compute

Answer

D. Serverless compute

Explanation

Serverless compute services such as Azure Functions automatically scale based on workload demand and reduce operational overhead.

Question 2

You need to generate semantic vector representations for documents used in a RAG solution.

Which service is MOST commonly used?

A. Azure Bastion
B. Azure OpenAI Service
C. Azure Firewall
D. Azure Route Server

Answer

B. Azure OpenAI Service

Explanation

Azure OpenAI Service provides embedding models that generate semantic vectors used in vector search and retrieval systems.

Question 3

You are building an AI application that retrieves enterprise documents using meaning rather than exact keyword matches.

Which capability should you implement?

A. Vector search
B. Blob replication
C. DNS forwarding
D. OCR masking

Answer

A. Vector search

Question 4

MULTIPLE ANSWER — Which components are commonly included in a hybrid search solution? (Choose THREE)

A. Vector retrieval
B. Keyword search
C. Semantic ranking
D. DHCP scopes
E. VPN tunneling

Answer

A. Vector retrieval
B. Keyword search
C. Semantic ranking

Question 5

You need to identify whether customer feedback expresses positive, neutral, or negative opinions.

Which capability should you use?

A. Object detection
B. OCR
C. Sentiment analysis
D. Image tagging

Answer

C. Sentiment analysis

Question 6

MATCHING — Match the capability to the correct scenario.

CapabilityScenario
OCR?
Text summarization?
Named Entity Recognition?

Options:

  • Identify company names in contracts
  • Create concise summaries of reports
  • Extract text from scanned forms

Answer

CapabilityScenario
OCRExtract text from scanned forms
Text summarizationCreate concise summaries of reports
Named Entity RecognitionIdentify company names in contracts

Question 7

You need an AI agent to call APIs and execute workflows dynamically.

Which feature enables this capability?

A. Layout analysis
B. Semantic ranking
C. Function calling
D. Object segmentation

Answer

C. Function calling

Question 8

You are processing scanned PDFs that contain tables, headers, and forms.

Which capability is MOST important for preserving structure?

A. OCR only
B. Layout analysis
C. Translation
D. Speech synthesis

Answer

B. Layout analysis

Question 9

FILL IN THE BLANK

The process of splitting documents into smaller retrievable sections is called __________.

Answer

chunking

Question 10

You are implementing an AI search system that combines:

  • Keyword matching
  • Semantic similarity

What type of search is this?

A. Sequential search
B. OCR search
C. Relational search
D. Hybrid search

Answer

D. Hybrid search

Question 11

You need to extract structured invoice fields such as:

  • Vendor name
  • Total amount
  • Invoice number

Which Azure service should you use?

A. Azure AI Document Intelligence
B. Azure Traffic Manager
C. Azure VPN Gateway
D. Azure DNS

Answer

A. Azure AI Document Intelligence

Question 12

You need to reduce hallucinations in a chatbot that answers questions from enterprise documentation.

Which approach should you implement?

A. Disable embeddings
B. Increase randomness
C. Grounding with RAG
D. Remove retrieval steps

Answer

C. Grounding with RAG

Question 13

HOTSPOT — Select the BEST capability for each requirement.

RequirementCapability
Detect vehicles in images?
Convert speech into text?
Extract key topics from reviews?

Options:

  • Speech-to-text
  • Object detection
  • Key phrase extraction

Answer

RequirementCapability
Detect vehicles in imagesObject detection
Convert speech into textSpeech-to-text
Extract key topics from reviewsKey phrase extraction

Question 14

You are building a multimodal AI system.

Which inputs are examples of multimodal data? (Choose TWO)

A. Images
B. Audio
C. VLAN identifiers
D. Route tables

Answer

A. Images
B. Audio

Question 15

You need to monitor telemetry, logs, and performance metrics for Azure AI APIs.

Which Azure service should you use?

A. Azure Backup
B. Azure Monitor
C. Azure Front Door
D. Azure Bastion

Answer

B. Azure Monitor

Question 16

MULTIPLE ANSWER — Which techniques improve retrieval quality in RAG systems? (Choose THREE)

A. Semantic chunking
B. Metadata enrichment
C. Disabling semantic ranking
D. Removing indexes
E. Chunk overlap

Answer

A. Semantic chunking
B. Metadata enrichment
E. Chunk overlap

Question 17

You need to authenticate Azure resources securely without storing secrets in application code.

Which feature should you implement?

A. Shared admin passwords
B. Anonymous access
C. Managed identities
D. Public storage containers

Answer

C. Managed identities

Question 18

You are building a solution that analyzes handwritten forms.

Which capability is MOST important?

A. OCR with handwriting recognition
B. Image resizing
C. Object tracking
D. Speech synthesis

Answer

A. OCR with handwriting recognition

Question 19

DRAG AND DROP — Match the Azure service to its primary purpose.

ServicePurpose
Azure AI Search?
Azure AI Vision?
Azure Logic Apps?

Options:

  • Workflow orchestration
  • Image analysis
  • Search and retrieval

Answer

ServicePurpose
Azure AI SearchSearch and retrieval
Azure AI VisionImage analysis
Azure Logic AppsWorkflow orchestration

Question 20

You need to create concise summaries of long legal contracts.

Which AI capability should you implement?

A. Translation
B. OCR
C. Text summarization
D. Face analysis

Answer

C. Text summarization

Question 21

You need to preserve headings, tables, and reading order for downstream AI reasoning.

Which output format is BEST suited?

A. Binary serialization
B. Markdown
C. JPEG
D. TIFF

Answer

B. Markdown

Question 22

You need an AI system to identify products, locations, and organizations in support tickets.

Which capability should you use?

A. OCR
B. Translation
C. Image segmentation
D. Named Entity Recognition

Answer

D. Named Entity Recognition

Question 23

MULTIPLE ANSWER — Which are benefits of vector embeddings? (Choose THREE)

A. Semantic similarity matching
B. Contextual retrieval
C. Natural language understanding
D. DNS optimization
E. Network segmentation

Answer

A. Semantic similarity matching
B. Contextual retrieval
C. Natural language understanding

Question 24

You need to orchestrate AI workflows between Azure services and third-party systems using a low-code approach.

Which service should you use?

A. Azure Load Balancer
B. Azure Logic Apps
C. Azure Firewall
D. Azure Site Recovery

Answer

B. Azure Logic Apps

Question 25

FILL IN THE BLANK

The process of supplying trusted external context to an LLM is called __________.

Answer

grounding

Question 26

You need an AI solution that retrieves only documents users are authorized to access.

Which approach is MOST appropriate?

A. Shared global administrator access
B. Anonymous indexes
C. Security trimming with RBAC
D. Public blob storage

Answer

C. Security trimming with RBAC

Question 27

You are building a computer vision solution that identifies multiple objects within warehouse images.

Which capability should you use?

A. OCR
B. Object detection
C. Translation
D. Sentiment analysis

Answer

B. Object detection

Question 28

You need to improve search precision by storing additional context such as:

  • Department
  • Region
  • Document type

Which technique should you implement?

A. Metadata enrichment
B. Token suppression
C. DNS replication
D. Embedding deletion

Answer

A. Metadata enrichment

Question 29

You need to build a chatbot that retrieves current enterprise information before generating responses.

Which architecture is MOST appropriate?

A. Static FAQ architecture
B. RAG architecture
C. Traditional ETL architecture
D. Batch processing architecture

Answer

B. RAG architecture

Question 30

You are implementing enterprise AI governance and want to restrict access to sensitive retrieval indexes.

Which practice BEST improves security?

A. Disable authentication
B. Public anonymous access
C. RBAC enforcement
D. Shared credentials across teams

Answer

C. RBAC enforcement

Explanation

Role-based access control (RBAC) ensures that:

Enterprise governance policies are enforced

Users access only authorized resources

AI systems retrieve approved content

Go to the AI-103 Exam Prep Hub main page

Leave a comment