Category: Artificial Intelligence (AI)

Agentic AI, AI, AI-103, Artificial Intelligence (AI), Azure AI, Generative AI, Microsoft Certification

Configure an application to connect to a Foundry project (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Implement generative AI and agentic solutions (30–35%)
   --> Build generative applications by using Foundry
      --> Configure an application to connect to a Foundry project

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

Azure AI Foundry provides a centralized environment for developing, deploying, and managing AI applications and agentic solutions.

Applications that use generative AI models, agents, retrieval systems, or multimodal capabilities must connect securely and reliably to Foundry projects.

This topic is important for the AI-103: Develop AI Apps and Agents on Azure certification exam.

For the AI-103 exam, you should understand:

  • Azure AI Foundry projects
  • Application connectivity
  • Authentication methods
  • SDK configuration
  • Endpoint configuration
  • Deployment configuration
  • Managed identities
  • API keys
  • Environment variables
  • Network security
  • Role-based access control (RBAC)
  • Connecting to deployed models and agents
  • Configuration management
  • Monitoring and troubleshooting

What Is an Azure AI Foundry Project?

An Azure AI Foundry project is a centralized workspace used to:

  • Manage AI resources
  • Deploy models
  • Configure agents
  • Build workflows
  • Store evaluation assets
  • Monitor AI systems

Projects help organize AI development and operations.

Components of a Foundry Project

A Foundry project may include:

  • Model deployments
  • Agent configurations
  • Prompt flows
  • Evaluation datasets
  • Connections
  • Search resources
  • Storage resources
  • Monitoring tools

Why Applications Need Project Connectivity

Applications connect to Foundry projects to:

  • Access deployed models
  • Invoke agents
  • Perform retrieval operations
  • Execute workflows
  • Use AI services securely

Common Connection Scenarios

Applications commonly connect to:

  • Chat models
  • Embedding models
  • Multimodal models
  • Agent services
  • Prompt flow endpoints
  • Azure AI Search resources

Connection Architecture

Typical connectivity includes:

  1. Application
  2. Authentication layer
  3. Foundry project endpoint
  4. Model or agent deployment

SDK-Based Connectivity

Applications often use SDKs to:

  • Authenticate
  • Send prompts
  • Receive responses
  • Stream outputs
  • Manage workflows

SDKs simplify development.

API-Based Connectivity

Applications may also use:

  • REST APIs
  • HTTP endpoints
  • Direct service requests

Authentication Methods

Applications must authenticate securely.

Common methods include:

  • API keys
  • Managed identities
  • Azure Active Directory (Azure AD)
  • Keyless authentication

API Key Authentication

API keys are:

  • Simple to configure
  • Easy for development and testing

However, they require secure storage.

Managed Identity Authentication

Managed identities provide:

  • Secretless authentication
  • Improved security
  • Automatic credential management

Managed identity is recommended for production workloads.

Azure AD Authentication

Azure AD enables:

  • Enterprise identity management
  • Role-based access
  • Secure authentication workflows

Keyless Authentication

Keyless authentication reduces:

  • Credential exposure
  • Secret management overhead

Secure Credential Storage

Applications should avoid:

  • Hardcoded secrets
  • Plain-text credentials

Credentials should be stored securely.

Environment Variables

Environment variables commonly store:

  • API endpoints
  • Deployment names
  • Keys
  • Configuration settings

Configuration Files

Applications may use:

  • JSON configuration files
  • YAML files
  • Application settings

Endpoint Configuration

Applications must connect to the correct:

  • Foundry endpoint
  • Model deployment endpoint
  • Agent endpoint

Deployment Names

Applications typically reference:

  • Specific deployment names
  • Model identifiers
  • Agent identifiers

Connecting to Model Deployments

Applications may connect to:

  • Chat completion models
  • Embedding models
  • Code models
  • Multimodal models

Connecting to Agent Workflows

Applications may invoke agents that:

  • Use tools
  • Access memory
  • Execute workflows
  • Coordinate tasks

Connecting to Prompt Flows

Applications can invoke:

  • Prompt flow endpoints
  • Orchestrated workflows
  • Multi-step pipelines

Connecting to Azure AI Search

RAG applications often connect to:

  • Azure AI Search
  • Vector indexes
  • Semantic search pipelines

Role-Based Access Control (RBAC)

RBAC controls:

  • Resource permissions
  • Service access
  • Administrative privileges

Least Privilege Principle

Applications should receive:

  • Only required permissions
  • Minimal access rights

Private Networking

Organizations may secure connectivity using:

  • Private endpoints
  • Virtual networks
  • Network isolation

Firewall Configuration

Firewall rules may restrict:

  • Public access
  • Unauthorized IP ranges

Secure Communication

Applications should use:

  • HTTPS
  • Encrypted communication
  • Secure APIs

SDK Initialization

Applications typically initialize:

  • Client objects
  • Authentication providers
  • Connection settings

Client Configuration

Client configuration may include:

  • Endpoint URLs
  • API versions
  • Deployment names
  • Authentication credentials

Streaming Configuration

Applications may enable:

  • Streaming responses
  • Incremental output rendering

Retry Policies

Applications should implement:

  • Retry logic
  • Exponential backoff
  • Timeout handling

Error Handling

Applications should handle:

  • Authentication failures
  • Network issues
  • Rate limits
  • Invalid requests

Logging and Monitoring

Applications should log:

  • Requests
  • Responses
  • Failures
  • Latency metrics

Observability

Observability helps organizations:

  • Monitor usage
  • Diagnose issues
  • Improve reliability

Application Scalability

Applications should support:

  • High concurrency
  • Distributed workloads
  • Elastic scaling

Cost Considerations

Connection design impacts:

  • Token usage
  • API consumption
  • Search operations
  • Infrastructure costs

CI/CD Integration

Connection settings may be managed through:

  • Deployment pipelines
  • Infrastructure as code
  • Environment promotion

Development vs Production Environments

Organizations often separate:

  • Development
  • Testing
  • Staging
  • Production

Each environment may use different:

  • Endpoints
  • Credentials
  • Policies

Multi-Region Connectivity

Global applications may connect to:

  • Multiple regional deployments
  • Regional failover systems

High Availability

Applications should support:

  • Redundant deployments
  • Failover strategies
  • Resilient architecture

Governance Considerations

Organizations may enforce:

  • Access policies
  • Security baselines
  • Audit logging
  • Compliance requirements

Troubleshooting Connectivity Issues

Common issues include:

  • Invalid credentials
  • Incorrect endpoints
  • Missing RBAC permissions
  • Network restrictions
  • Deployment mismatches

Performance Optimization

Organizations should optimize:

  • Connection reuse
  • Latency
  • Request batching
  • Streaming efficiency

Real-World Scenario

Scenario: Enterprise AI Assistant

Requirements:

  • Secure authentication
  • RAG integration
  • Agent orchestration
  • Enterprise access control

Recommended Approach:

  • Managed identity
  • RBAC
  • Private networking
  • Azure AI Search integration
  • SDK-based connectivity

Common AI-103 Exam Tips

Understand Authentication Options

Know when to use:

  • API keys
  • Managed identities
  • Azure AD

Understand Endpoint Configuration

Know:

  • Deployment names
  • Service endpoints
  • Agent endpoints

Learn RBAC Concepts

Understand:

  • Least privilege
  • Role assignments
  • Secure access management

Understand Networking Concepts

Know:

  • Private endpoints
  • Firewalls
  • Secure connectivity

Learn Application Integration Concepts

Understand:

  • SDK initialization
  • Client configuration
  • Retry logic
  • Monitoring

Summary

Connecting applications to Azure AI Foundry projects is a foundational skill for AI-103.

For the exam, you should understand:

  • Foundry projects
  • Application connectivity
  • SDK integration
  • API integration
  • Authentication methods
  • Managed identities
  • RBAC
  • Deployment configuration
  • Endpoint management
  • Networking security
  • Logging and monitoring
  • Scalability and reliability

These skills are essential for building secure, scalable enterprise AI applications on Azure.

Practice Exam Questions

Question 1

What is the purpose of an Azure AI Foundry project?

A. Replace Azure subscriptions
B. Centrally manage AI resources, deployments, and workflows
C. Eliminate authentication
D. Replace APIs entirely

Answer

B. Centrally manage AI resources, deployments, and workflows

Explanation

Foundry projects organize AI development and operational assets.

Question 2

Which authentication method is recommended for production Azure workloads?

A. Hardcoded credentials
B. Managed identity
C. Shared public keys
D. Anonymous access

Answer

B. Managed identity

Explanation

Managed identities improve security by avoiding embedded secrets.

Question 3

What is a primary advantage of SDKs?

A. They eliminate APIs completely
B. They simplify application development and integration
C. They remove all authentication requirements
D. They prevent monitoring

Answer

B. They simplify application development and integration

Explanation

SDKs provide abstractions that simplify connectivity and workflow development.

Question 4

Why should applications use environment variables?

A. To increase GPU performance
B. To securely manage configuration values
C. To eliminate authentication
D. To disable RBAC

Answer

B. To securely manage configuration values

Explanation

Environment variables help manage endpoints and credentials securely.

Question 5

What does RBAC primarily control?

A. Token compression
B. Permissions and access to resources
C. Model quantization
D. Network bandwidth

Answer

B. Permissions and access to resources

Explanation

RBAC enforces authorization policies.

Question 6

Why are private endpoints used?

A. To increase hallucinations
B. To improve network security and isolate traffic
C. To disable monitoring
D. To reduce embedding dimensions

Answer

B. To improve network security and isolate traffic

Explanation

Private endpoints help secure enterprise AI workloads.

Question 7

What is commonly required when connecting to a deployed model?

A. Deployment name
B. Firewall removal
C. Disabling authentication
D. Public anonymous access

Answer

A. Deployment name

Explanation

Applications typically reference deployment identifiers.

Question 8

Why should applications implement retry policies?

A. To increase hallucinations
B. To recover from transient failures and improve reliability
C. To disable APIs
D. To remove authentication

Answer

B. To recover from transient failures and improve reliability

Explanation

Retry logic improves resiliency.

Question 9

Which service is commonly integrated for RAG search functionality?

A. Azure AI Search
B. Azure DNS
C. Azure Backup
D. Azure Batch

Answer

A. Azure AI Search

Explanation

Azure AI Search supports vector and semantic retrieval.

Question 10

What is the least privilege principle?

A. Give all users full access
B. Grant only the permissions necessary to perform required tasks
C. Disable RBAC
D. Allow anonymous authentication

Answer

B. Grant only the permissions necessary to perform required tasks

Explanation

Least privilege reduces security risk by minimizing unnecessary permissions.

Go to the AI-103 Exam Prep Hub main page

Agentic AI, AI, AI-103, Artificial Intelligence (AI), Azure AI, Microsoft Certification

Evaluate models and apps, including detecting fabrications, relevance, quality, and safety (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Implement generative AI and agentic solutions (30–35%)
   --> Build generative applications by using Foundry
      --> Evaluate models and apps, including detecting fabrications, relevance, quality, and safety

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

Building generative AI applications is only part of the development process.

Organizations must also evaluate whether AI systems are:

  • Accurate
  • Reliable
  • Relevant
  • Safe
  • Grounded
  • Trustworthy

AI systems can generate:

  • Hallucinations
  • Unsafe content
  • Biased responses
  • Irrelevant answers
  • Inconsistent outputs

The AI-103: Develop AI Apps and Agents on Azure certification exam tests your understanding of evaluating models and applications.

For the AI-103 exam, you should understand:

  • Model evaluation
  • Application evaluation
  • Fabrication detection
  • Groundedness
  • Relevance evaluation
  • Quality evaluation
  • Safety evaluation
  • Responsible AI testing
  • Automated evaluators
  • Human evaluation
  • Benchmarking
  • Monitoring and continuous evaluation

Why AI Evaluation Matters

Evaluation is essential because generative AI systems are probabilistic.

This means:

  • Responses may vary
  • Outputs may be incorrect
  • Safety risks may occur
  • Hallucinations may appear

Without evaluation, organizations cannot reliably trust AI systems.

What Is AI Evaluation?

AI evaluation is the process of measuring:

  • Accuracy
  • Safety
  • Reliability
  • Relevance
  • Groundedness
  • User satisfaction

Types of AI Evaluation

Common evaluation categories include:

  • Model evaluation
  • Prompt evaluation
  • Retrieval evaluation
  • Application evaluation
  • Safety evaluation
  • Human evaluation

Model Evaluation

Model evaluation focuses on:

  • Model quality
  • Accuracy
  • Performance
  • Reasoning ability

Application Evaluation

Application evaluation measures:

  • End-to-end user experience
  • Workflow success
  • Tool orchestration quality
  • Groundedness

What Are Fabrications?

Fabrications are generated outputs that:

  • Are incorrect
  • Are unsupported
  • Contain invented facts
  • Misrepresent information

Fabrications are commonly called hallucinations.

Causes of Fabrications

Fabrications may occur because:

  • The model lacks relevant knowledge
  • Prompts are ambiguous
  • Retrieval quality is poor
  • Context is insufficient
  • Safety constraints are weak

Fabrication Detection

Organizations should evaluate whether outputs:

  • Match trusted sources
  • Remain grounded
  • Avoid unsupported claims

Groundedness Evaluation

Groundedness measures whether responses are supported by:

  • Retrieved documents
  • Enterprise data
  • Trusted sources

Importance of Groundedness

Grounded responses:

  • Improve trust
  • Reduce hallucinations
  • Increase explainability

Retrieval Quality Evaluation

RAG systems should evaluate:

  • Search relevance
  • Retrieved chunk quality
  • Citation accuracy
  • Context completeness

Relevance Evaluation

Relevance measures whether responses:

  • Answer the user’s question
  • Stay on-topic
  • Match user intent

Quality Evaluation

Quality evaluations may assess:

  • Clarity
  • Completeness
  • Coherence
  • Fluency
  • Professionalism

Consistency Evaluation

Consistency measures whether models:

  • Produce stable responses
  • Avoid contradictory outputs
  • Maintain predictable behavior

Safety Evaluation

Safety evaluations identify:

  • Harmful outputs
  • Toxic content
  • Unsafe instructions
  • Policy violations

Responsible AI Evaluation

Responsible AI testing focuses on:

  • Fairness
  • Safety
  • Transparency
  • Accountability
  • Privacy

Bias Evaluation

Organizations should evaluate whether models:

  • Produce biased outputs
  • Treat groups unfairly
  • Reinforce stereotypes

Toxicity Detection

Toxicity evaluations identify:

  • Offensive language
  • Hate speech
  • Harassment
  • Abusive content

Jailbreak Testing

Jailbreak testing evaluates whether users can bypass:

  • Safety controls
  • Content filters
  • Guardrails

Adversarial Testing

Adversarial testing intentionally challenges models using:

  • Malicious prompts
  • Edge cases
  • Prompt injection attacks

Prompt Injection Testing

Prompt injection testing evaluates whether:

  • External content manipulates model behavior
  • Instructions override safety policies

Automated Evaluators

Automated evaluators use:

  • Rules
  • Scoring systems
  • AI-based evaluators

To assess model outputs.

AI-Assisted Evaluation

Some systems use LLMs to evaluate:

  • Relevance
  • Groundedness
  • Quality
  • Safety

Human Evaluation

Human reviewers may evaluate:

  • Accuracy
  • Tone
  • Helpfulness
  • Safety
  • Business alignment

Human-in-the-Loop Evaluation

Human-in-the-loop evaluation combines:

  • Automated evaluation
  • Human oversight
  • Expert validation

Benchmarking Models

Benchmarking compares models using:

  • Standard datasets
  • Consistent prompts
  • Defined metrics

A/B Testing

A/B testing compares:

  • Different prompts
  • Different models
  • Different workflows

Evaluation Metrics

Common metrics include:

  • Precision
  • Recall
  • Accuracy
  • Relevance
  • Groundedness
  • Toxicity scores
  • Latency
  • User satisfaction

Precision and Recall

Precision

Measures how many retrieved results are relevant.

Recall

Measures how many relevant results were successfully retrieved.

Latency Evaluation

Organizations should measure:

  • Response times
  • Retrieval delays
  • Tool execution times

Cost Evaluation

Cost evaluation considers:

  • Token usage
  • API calls
  • Infrastructure consumption

User Satisfaction Evaluation

Organizations may measure:

  • User feedback
  • Completion success
  • Satisfaction ratings

Continuous Evaluation

AI systems should be evaluated continuously because:

  • User behavior changes
  • Data evolves
  • Model drift may occur

Model Drift

Model drift occurs when:

  • Performance changes over time
  • Inputs evolve
  • User expectations shift

Monitoring Production Systems

Organizations should monitor:

  • Safety violations
  • Hallucination rates
  • Retrieval failures
  • Latency spikes
  • Cost increases

Evaluation Pipelines

Evaluation pipelines automate:

  • Testing
  • Scoring
  • Reporting
  • Regression analysis

Regression Testing

Regression testing ensures updates do not:

  • Reduce quality
  • Break workflows
  • Increase hallucinations

Azure AI Foundry Evaluation Capabilities

Azure AI Foundry supports:

  • Evaluation workflows
  • Automated evaluators
  • Safety monitoring
  • Groundedness evaluation
  • Prompt testing
  • Trace analysis

Trace Analysis

Trace analysis helps inspect:

  • Tool calls
  • Retrieval steps
  • Agent decisions
  • Workflow execution

Evaluation Datasets

Organizations should create datasets containing:

  • Expected outputs
  • Edge cases
  • Adversarial prompts
  • Real-world scenarios

Synthetic Test Data

Synthetic data may help test:

  • Rare scenarios
  • Adversarial prompts
  • Safety boundaries

Real-World Evaluation Scenarios

Scenario 1: Enterprise Chatbot

Requirements:

  • Accurate responses
  • Citation support
  • Low hallucination rate

Recommended Evaluation:

  • Groundedness testing
  • Retrieval quality evaluation

Scenario 2: Financial Assistant

Requirements:

  • High accuracy
  • Safety compliance
  • Low fabrication risk

Recommended Evaluation:

  • Human review
  • Adversarial testing
  • Approval workflows

Scenario 3: Customer Support Copilot

Requirements:

  • Relevant responses
  • Fast response times
  • Consistent tone

Recommended Evaluation:

  • Latency evaluation
  • Quality scoring
  • A/B testing

Scenario 4: Agentic Workflow System

Requirements:

  • Tool accuracy
  • Safe tool execution
  • Workflow traceability

Recommended Evaluation:

  • Trace analysis
  • Tool execution monitoring
  • HITL evaluation

Common AI-103 Exam Tips

Understand Evaluation Categories

Know the differences between:

  • Relevance
  • Quality
  • Groundedness
  • Safety
  • Consistency

Learn Fabrication Detection Concepts

Understand:

  • Hallucinations
  • Unsupported claims
  • Grounding validation

Understand Safety Testing

Know:

  • Toxicity testing
  • Jailbreak testing
  • Prompt injection evaluation
  • Adversarial testing

Learn Monitoring Concepts

Understand:

  • Continuous evaluation
  • Drift detection
  • Trace analysis
  • Regression testing

Summary

Evaluating generative AI systems is critical for building:

  • Reliable
  • Safe
  • Grounded
  • Trustworthy applications

For the AI-103 exam, you should understand:

  • Fabrication detection
  • Groundedness evaluation
  • Retrieval quality
  • Relevance testing
  • Quality evaluation
  • Safety evaluation
  • Toxicity detection
  • Adversarial testing
  • Human evaluation
  • Automated evaluators
  • Monitoring and drift detection
  • Evaluation pipelines

These concepts are foundational for developing enterprise-grade AI applications and agentic systems on Azure.

Practice Exam Questions

Question 1

What is a fabrication in generative AI?

A. A storage replication process
B. An unsupported or invented response
C. A vector indexing method
D. A deployment strategy

Answer

B. An unsupported or invented response

Explanation

Fabrications, also called hallucinations, are incorrect or invented outputs.

Question 2

What does groundedness measure?

A. GPU performance
B. Whether outputs are supported by trusted sources
C. Network bandwidth
D. Token compression efficiency

Answer

B. Whether outputs are supported by trusted sources

Explanation

Groundedness evaluates factual support from retrieved or trusted data.

Question 3

Which evaluation type focuses on harmful or unsafe outputs?

A. Latency evaluation
B. Safety evaluation
C. Compression evaluation
D. Replication evaluation

Answer

B. Safety evaluation

Explanation

Safety evaluations detect harmful, toxic, or policy-violating outputs.

Question 4

What is the purpose of retrieval quality evaluation in RAG systems?

A. Measure GPU speed
B. Assess search relevance and retrieved context quality
C. Reduce storage redundancy
D. Disable embeddings

Answer

B. Assess search relevance and retrieved context quality

Explanation

Retrieval quality measures how useful and relevant retrieved information is.

Question 5

What is jailbreak testing?

A. Testing storage failures
B. Evaluating attempts to bypass safety controls
C. Measuring retrieval latency
D. Compressing prompts

Answer

B. Evaluating attempts to bypass safety controls

Explanation

Jailbreak testing checks whether users can circumvent AI safety mechanisms.

Question 6

Which metric measures whether responses answer the user’s question appropriately?

A. Relevance
B. Replication
C. Throughput
D. Compression

Answer

A. Relevance

Explanation

Relevance evaluates how well outputs match user intent.

Question 7

Why is continuous evaluation important?

A. To eliminate all infrastructure costs
B. Because models and data can change over time
C. To remove all safety policies
D. To disable monitoring

Answer

B. Because models and data can change over time

Explanation

Continuous evaluation helps detect drift and performance degradation.

Question 8

What is adversarial testing?

A. Testing network redundancy
B. Challenging AI systems with malicious or difficult prompts
C. Increasing vector dimensions
D. Optimizing GPU allocation

Answer

B. Challenging AI systems with malicious or difficult prompts

Explanation

Adversarial testing identifies vulnerabilities and unsafe behaviors.

Question 9

What is a benefit of A/B testing in AI systems?

A. Eliminates monitoring requirements
B. Compares prompts or models to identify better performance
C. Removes the need for evaluation datasets
D. Disables retrieval pipelines

Answer

B. Compares prompts or models to identify better performance

Explanation

A/B testing helps optimize prompts, workflows, and models.

Question 10

Which Azure capability helps inspect workflow execution and tool calls?

A. Trace analysis
B. DNS failover
C. Storage mirroring
D. GPU partitioning

Answer

A. Trace analysis

Explanation

Trace analysis provides visibility into workflow execution and reasoning steps.

Go to the AI-103 Exam Prep Hub main page

Agentic AI, AI, AI-103, Artificial Intelligence (AI), Azure AI, Generative AI, Microsoft Certification

Design workflows, tool-augmented flows, and multistep reasoning pipelines (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Implement generative AI and agentic solutions (30–35%)
   --> Build generative applications by using Foundry
      --> Design workflows, tool-augmented flows, and multistep reasoning pipelines

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

Modern AI systems are evolving beyond simple prompt-response interactions.

Today’s generative AI applications often:

  • Use external tools
  • Perform multistep reasoning
  • Orchestrate workflows
  • Retrieve enterprise data
  • Execute actions autonomously
  • Coordinate across services

These systems are commonly called:

  • Agentic systems
  • Tool-augmented AI systems
  • AI workflow pipelines

The AI-103: Develop AI Apps and Agents on Azure certification exam tests your understanding of designing intelligent workflows and reasoning pipelines.

For the AI-103 exam, you should understand:

  • AI workflows
  • Agent orchestration
  • Tool augmentation
  • Function calling
  • Multistep reasoning
  • Workflow pipelines
  • Retrieval integration
  • Memory integration
  • Planning and execution
  • Human-in-the-loop workflows
  • Monitoring and governance

What Are AI Workflows?

AI workflows are structured sequences of operations that combine:

  • AI reasoning
  • Data retrieval
  • Tool execution
  • Decision-making
  • Automation

Workflows coordinate multiple steps to complete complex tasks.

Why AI Workflows Matter

Simple prompts are often insufficient for:

  • Enterprise automation
  • Complex reasoning
  • Dynamic decision-making
  • Multi-system integration

Workflows allow AI systems to:

  • Break problems into steps
  • Use external tools
  • Validate outputs
  • Iterate toward solutions

What Is Tool Augmentation?

Tool augmentation allows AI systems to use external capabilities.

Examples include:

  • APIs
  • Databases
  • Search engines
  • Calculators
  • Business systems
  • Code interpreters

Why Tool Augmentation Is Important

Language models alone:

  • Cannot access real-time data
  • Cannot execute business actions directly
  • Cannot reliably perform all calculations

Tools extend AI capabilities.

Common Tool-Augmented Scenarios

Examples include:

  • Checking inventory
  • Booking appointments
  • Querying databases
  • Sending emails
  • Executing workflows
  • Calling REST APIs

What Is Function Calling?

Function calling enables models to:

  • Detect when a tool is needed
  • Generate structured tool requests
  • Invoke external services
  • Process returned results

Function Calling Workflow

Typical flow:

  1. User submits request
  2. Model determines tool requirement
  3. Model generates function call
  4. External tool executes
  5. Results return to model
  6. Model generates final response

Structured Tool Inputs

Function calling typically uses:

  • JSON schemas
  • Structured parameters
  • Validated inputs

This improves reliability.

Tool Selection

Agentic systems may dynamically choose:

  • Which tools to use
  • Which workflows to invoke
  • Which retrieval strategies to apply

Tool Orchestration

Tool orchestration coordinates multiple tools within a workflow.

Examples include:

  • Retrieval + summarization
  • Search + booking systems
  • Database queries + reporting

Sequential Workflows

Sequential workflows execute steps in order.

Example:

  1. Retrieve customer data
  2. Analyze account status
  3. Generate recommendations
  4. Send response

Parallel Workflows

Parallel workflows execute multiple tasks simultaneously.

Benefits include:

  • Faster execution
  • Better scalability
  • Reduced latency

Conditional Workflows

Conditional workflows branch based on:

  • User intent
  • Retrieved data
  • Safety evaluations
  • Confidence scores

What Is Multistep Reasoning?

Multistep reasoning breaks complex problems into smaller steps.

This improves:

  • Accuracy
  • Planning
  • Decision quality

Examples of Multistep Reasoning

Examples include:

  • Research workflows
  • Financial analysis
  • Travel planning
  • Technical troubleshooting

Chain-of-Thought Reasoning

Chain-of-thought reasoning encourages models to:

  • Reason step-by-step
  • Decompose problems
  • Validate intermediate steps

Planning and Execution Models

Agentic systems often separate:

  • Planning
  • Execution

The planner decides:

  • What steps are needed
  • Which tools to use

The executor performs actions.

Planner-Executor Architectures

Planner-executor architectures support:

  • Dynamic workflows
  • Adaptive reasoning
  • Task decomposition

ReAct Pattern

The ReAct (Reason + Act) pattern combines:

  • Reasoning
  • Tool usage
  • Observation
  • Iterative decision-making

Reflection and Self-Correction

Some systems support:

  • Self-evaluation
  • Output refinement
  • Error correction

Retrieval-Augmented Workflows

Workflows often integrate:

  • Vector search
  • RAG pipelines
  • Enterprise grounding

Memory in Agentic Systems

AI systems may use memory for:

  • Conversation history
  • User preferences
  • Workflow state
  • Long-running tasks

Short-Term Memory

Short-term memory stores:

  • Current conversation context
  • Immediate workflow information

Long-Term Memory

Long-term memory stores:

  • Persistent preferences
  • Historical interactions
  • Learned context

Workflow State Management

State management tracks:

  • Current task progress
  • Intermediate outputs
  • Pending actions

Human-in-the-Loop (HITL) Workflows

High-risk workflows may require:

  • Human approvals
  • Validation checkpoints
  • Escalation paths

Approval Gates

Approval gates can prevent:

  • Unsafe actions
  • Unauthorized tool usage
  • Harmful outputs

Safety and Governance

Organizations should enforce:

  • Tool restrictions
  • Permission boundaries
  • Safety filters
  • Approval workflows

Autonomous vs Semi-Autonomous Agents

Autonomous Agents

Can:

  • Make decisions independently
  • Execute workflows automatically

Semi-Autonomous Agents

Require:

  • Human review
  • Approval checkpoints

Workflow Monitoring

Organizations should monitor:

  • Tool usage
  • Failures
  • Safety violations
  • Latency
  • Costs

Trace Logging

Trace logging helps track:

  • Workflow execution
  • Tool calls
  • Reasoning steps
  • Agent decisions

Error Handling in Workflows

Workflow pipelines should handle:

  • API failures
  • Missing data
  • Timeout errors
  • Invalid outputs

Retry Strategies

Common retry strategies include:

  • Automatic retries
  • Fallback workflows
  • Alternative tool selection

Fallback Models

Applications may use fallback models when:

  • Primary models fail
  • Costs exceed thresholds
  • Latency becomes excessive

Workflow Optimization

Optimization strategies include:

  • Parallel processing
  • Caching
  • Smaller models
  • Efficient retrieval

Latency Considerations

Complex workflows may increase latency due to:

  • Multiple model calls
  • Tool invocations
  • Retrieval operations

Cost Considerations

Tool-augmented systems may increase:

  • Token usage
  • API calls
  • Infrastructure costs

Azure AI Foundry Workflow Capabilities

Azure AI Foundry supports:

  • Model orchestration
  • Tool integration
  • Agent workflows
  • Evaluation pipelines
  • Monitoring

Common AI-103 Workflow Scenarios

Scenario 1: Enterprise Research Assistant

Requirements:

  • Multi-document retrieval
  • Summarization
  • Citation generation

Recommended Workflow:

  • RAG + multistep reasoning

Scenario 2: Customer Service Agent

Requirements:

  • CRM access
  • Ticket management
  • Escalation workflows

Recommended Workflow:

  • Tool-augmented agent

Scenario 3: Financial Approval System

Requirements:

  • Risk evaluation
  • Human approvals
  • Audit logging

Recommended Workflow:

  • HITL approval pipeline

Scenario 4: AI Coding Assistant

Requirements:

  • Code generation
  • Code execution
  • Documentation retrieval

Recommended Workflow:

  • Code model + tool orchestration

Common AI-103 Exam Tips

Understand Workflow Patterns

Know:

  • Sequential workflows
  • Parallel workflows
  • Conditional workflows

Learn Tool-Augmented AI Concepts

Understand:

  • Function calling
  • Tool orchestration
  • Dynamic tool selection

Understand Multistep Reasoning

Know:

  • Chain-of-thought reasoning
  • Planner-executor patterns
  • ReAct workflows

Learn Governance Concepts

Understand:

  • HITL workflows
  • Approval gates
  • Monitoring
  • Trace logging

Summary

Modern AI applications increasingly rely on:

  • Workflow orchestration
  • Tool augmentation
  • Multistep reasoning
  • Agentic architectures

For the AI-103 exam, you should understand:

  • AI workflow design
  • Function calling
  • Tool orchestration
  • Sequential and parallel workflows
  • Multistep reasoning
  • Planner-executor architectures
  • ReAct patterns
  • Memory integration
  • HITL workflows
  • Monitoring and governance

These concepts enable organizations to build:

  • Intelligent
  • Autonomous
  • Scalable
  • Governed AI systems

They are foundational for modern generative AI and agentic solutions on Azure.

Practice Exam Questions

Question 1

What is the primary purpose of tool augmentation in AI systems?

A. Reduce storage costs
B. Extend model capabilities using external tools
C. Eliminate prompts
D. Replace vector search

Answer

B. Extend model capabilities using external tools

Explanation

Tool augmentation enables AI systems to interact with APIs, databases, and other services.

Question 2

What does function calling enable a model to do?

A. Generate only static responses
B. Invoke external tools using structured inputs
C. Eliminate workflows
D. Replace embeddings

Answer

B. Invoke external tools using structured inputs

Explanation

Function calling allows models to interact with external services.

Question 3

Which workflow type executes tasks simultaneously?

A. Sequential workflow
B. Parallel workflow
C. Manual workflow
D. Static workflow

Answer

B. Parallel workflow

Explanation

Parallel workflows improve speed by running tasks concurrently.

Question 4

What is multistep reasoning?

A. Compressing vector indexes
B. Breaking complex tasks into smaller reasoning steps
C. Increasing GPU memory
D. Reducing prompt size only

Answer

B. Breaking complex tasks into smaller reasoning steps

Explanation

Multistep reasoning improves problem-solving accuracy.

Question 5

What does the ReAct pattern combine?

A. Compression and storage
B. Reasoning and acting
C. Replication and scaling
D. Encryption and backup

Answer

B. Reasoning and acting

Explanation

ReAct combines reasoning steps with tool usage.

Question 6

What is the purpose of workflow state management?

A. Monitor GPU temperature
B. Track task progress and intermediate outputs
C. Disable logging
D. Replace semantic search

Answer

B. Track task progress and intermediate outputs

Explanation

State management helps maintain workflow continuity.

Question 7

Which architecture separates planning from execution?

A. Static inference architecture
B. Planner-executor architecture
C. Batch storage architecture
D. Compression architecture

Answer

B. Planner-executor architecture

Explanation

Planner-executor systems divide reasoning and execution responsibilities.

Question 8

Why are approval gates important in AI workflows?

A. They increase vector dimensions
B. They prevent unsafe or unauthorized actions
C. They reduce indexing speed
D. They eliminate monitoring requirements

Answer

B. They prevent unsafe or unauthorized actions

Explanation

Approval gates enforce governance and human oversight.

Question 9

Which concept allows AI systems to remember previous interactions?

A. Semantic ranking
B. Memory integration
C. Static chunking
D. GPU partitioning

Answer

B. Memory integration

Explanation

Memory enables contextual continuity and long-running workflows.

Question 10

What is a major challenge of complex AI workflows?

A. Eliminating all costs
B. Increased latency from multiple operations
C. Removing all need for monitoring
D. Preventing all hallucinations automatically

Answer

B. Increased latency from multiple operations

Explanation

Complex workflows may require multiple model calls and tool executions.

Go to the AI-103 Exam Prep Hub main page

AI, AI-103, Artificial Intelligence (AI), Generative AI, Microsoft Certification

Deploy and consume LLMs, small models, code models, and multimodal models (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Implement generative AI and agentic solutions (30–35%)
   --> Build generative applications by using Foundry
      --> Deploy and consume LLMs, small models, code models, and multimodal models

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

Modern AI applications rely on a wide variety of AI models.

Different models are optimized for different workloads, including:

  • Conversational AI
  • Code generation
  • Text summarization
  • Image understanding
  • Audio processing
  • Reasoning tasks
  • Agentic workflows

The AI-103: Develop AI Apps and Agents on Azure certification exam tests your understanding of how to deploy and consume AI models in Azure AI Foundry.

For the AI-103 exam, you should understand:

  • Large language models (LLMs)
  • Small language models (SLMs)
  • Code models
  • Multimodal models
  • Model deployment concepts
  • Model consumption patterns
  • API-based model access
  • Endpoint configuration
  • Performance and cost tradeoffs
  • Model selection strategies
  • Responsible AI considerations

What Are Large Language Models (LLMs)?

Large language models are advanced AI systems trained on massive datasets.

LLMs can:

  • Generate text
  • Summarize documents
  • Answer questions
  • Translate languages
  • Reason across prompts
  • Support conversational AI

Common LLM Use Cases

Typical use cases include:

  • AI assistants
  • Enterprise chatbots
  • Content generation
  • Knowledge retrieval
  • Agent orchestration
  • Workflow automation

Characteristics of LLMs

LLMs typically provide:

  • Strong reasoning
  • Broad general knowledge
  • Advanced conversational abilities
  • Complex instruction following

However, they also:

  • Require more compute
  • Cost more to run
  • May introduce higher latency

What Are Small Language Models (SLMs)?

Small language models are lightweight models optimized for:

  • Faster inference
  • Lower cost
  • Lower latency
  • Edge deployment
  • Specialized tasks

Common SLM Use Cases

SLMs are often used for:

  • Classification
  • Simple chatbots
  • Mobile applications
  • Embedded AI
  • Lightweight assistants

Benefits of Small Models

Advantages include:

  • Reduced infrastructure cost
  • Faster response times
  • Lower resource requirements
  • Easier deployment at scale

LLM vs SLM Tradeoffs

LLMs

Best for:

  • Complex reasoning
  • Broad knowledge
  • Multi-step tasks

Tradeoffs:

  • Higher cost
  • Higher latency
  • Larger infrastructure requirements

SLMs

Best for:

  • Lightweight inference
  • Narrow tasks
  • Cost-sensitive workloads

Tradeoffs:

  • Reduced reasoning capability
  • Smaller context windows
  • Less flexibility

What Are Code Models?

Code models are specialized AI models trained for software development tasks.

These models can:

  • Generate code
  • Explain code
  • Complete functions
  • Debug issues
  • Convert between languages

Common Code Model Use Cases

Typical scenarios include:

  • Developer copilots
  • Code generation
  • Documentation generation
  • Test generation
  • Refactoring assistance

Code Model Capabilities

Code models often support:

  • Multiple programming languages
  • Natural language prompts
  • Code reasoning
  • Syntax understanding

What Are Multimodal Models?

Multimodal models process multiple types of input.

Examples include:

  • Text and images
  • Text and audio
  • Video and text

Multimodal AI Capabilities

Multimodal models may support:

  • Image understanding
  • OCR
  • Visual question answering
  • Audio transcription
  • Speech interaction
  • Video analysis

Common Multimodal Use Cases

Examples include:

  • AI vision assistants
  • Document understanding
  • Medical imaging analysis
  • Voice assistants
  • Image captioning

Model Deployment in Azure AI Foundry

Azure AI Foundry enables developers to:

  • Discover models
  • Deploy models
  • Test models
  • Monitor deployments
  • Consume models through APIs

Model Catalogs

Azure AI Foundry provides access to:

  • Foundation models
  • Open-source models
  • Specialized models
  • Multimodal models

Deployment Concepts

A deployment makes a model available through:

  • APIs
  • Endpoints
  • Applications
  • Agent workflows

Deployment Types

Common deployment options include:

  • Managed online deployments
  • Serverless deployments
  • Real-time inference endpoints
  • Batch inference deployments

Real-Time Inference

Real-time inference is used for:

  • Interactive chat
  • AI assistants
  • Live applications
  • Agent workflows

Batch Inference

Batch inference is used for:

  • Large-scale document processing
  • Offline analysis
  • Scheduled workloads
  • Bulk content generation

Endpoint Configuration

Deployments expose endpoints for application access.

Endpoints may include:

  • Authentication
  • Rate limits
  • Scaling policies
  • Monitoring settings

Authentication and Authorization

Applications may access models using:

  • API keys
  • Managed identities
  • Microsoft Entra ID
  • Role-based access control (RBAC)

Consuming Models Through APIs

Applications consume deployed models using:

  • REST APIs
  • SDKs
  • Client libraries

Prompt-Based Interactions

Generative AI applications commonly interact with models through prompts.

Prompts may include:

  • Instructions
  • Context
  • Examples
  • Retrieved documents

System Prompts

System prompts define:

  • AI behavior
  • Tone
  • Constraints
  • Safety policies

Model Parameters

Common inference parameters include:

  • Temperature
  • Top-p
  • Max tokens
  • Frequency penalty
  • Presence penalty

Temperature

Temperature controls output randomness.

Lower temperature:

  • More deterministic
  • More predictable

Higher temperature:

  • More creative
  • More variable

Context Windows

Context windows determine how much information a model can process in a request.

Larger context windows support:

  • Long conversations
  • Large documents
  • Multi-document grounding

Streaming Responses

Streaming enables applications to receive responses incrementally.

Benefits include:

  • Improved user experience
  • Faster perceived response times

Grounding Models

Grounding improves factual accuracy by providing trusted data.

Grounded applications commonly use:

  • Vector search
  • Retrieval-Augmented Generation (RAG)
  • Enterprise knowledge sources

Model Selection Considerations

Developers should evaluate:

  • Accuracy
  • Cost
  • Latency
  • Context size
  • Reasoning ability
  • Multimodal support
  • Scalability

Choosing Between Models

Use LLMs When:

  • Complex reasoning is required
  • Broad knowledge is needed
  • Multi-step workflows are involved

Use SLMs When:

  • Low latency matters
  • Cost optimization is critical
  • Tasks are narrow or repetitive

Use Code Models When:

  • Building developer tools
  • Generating code
  • Supporting programming workflows

Use Multimodal Models When:

  • Images or audio are required
  • Visual understanding is needed
  • Mixed media inputs are processed

Scaling Model Deployments

Scaling strategies may include:

  • Autoscaling
  • Regional deployments
  • Load balancing
  • Rate limiting

Monitoring Deployments

Organizations should monitor:

  • Latency
  • Throughput
  • Token usage
  • Errors
  • Safety events
  • Cost

Cost Optimization

Cost optimization strategies include:

  • Choosing smaller models
  • Limiting token usage
  • Caching responses
  • Using batch processing

Responsible AI Considerations

Developers should implement:

  • Safety filters
  • Guardrails
  • Content moderation
  • Monitoring
  • Human oversight

Multimodal Safety Concerns

Multimodal systems may require:

  • Image moderation
  • OCR filtering
  • Audio moderation
  • Content safety evaluation

Agentic AI and Model Consumption

AI agents may use:

  • LLMs for reasoning
  • SLMs for lightweight tasks
  • Code models for automation
  • Multimodal models for perception

Common AI-103 Deployment Scenarios

Scenario 1: Enterprise Chatbot

Requirements:

  • Strong reasoning
  • Long conversations
  • Grounded responses

Recommended Model:

  • LLM with RAG

Scenario 2: Mobile AI Assistant

Requirements:

  • Fast responses
  • Low cost
  • Lightweight inference

Recommended Model:

  • Small language model

Scenario 3: Developer Copilot

Requirements:

  • Code generation
  • Programming assistance
  • Syntax awareness

Recommended Model:

  • Code model

Scenario 4: Image-Aware AI Assistant

Requirements:

  • Image analysis
  • OCR
  • Text generation

Recommended Model:

  • Multimodal model

Common AI-103 Exam Tips

Understand Model Categories

Know the differences between:

  • LLMs
  • SLMs
  • Code models
  • Multimodal models

Learn Deployment Concepts

Understand:

  • Endpoints
  • Real-time inference
  • Batch inference
  • Scaling

Learn Consumption Patterns

Know:

  • REST APIs
  • SDKs
  • Prompt engineering
  • System prompts

Understand Cost and Performance Tradeoffs

Know how:

  • Model size affects cost
  • Context size affects latency
  • Scaling impacts performance

Summary

Azure AI Foundry enables developers to deploy and consume a wide range of AI models.

For the AI-103 exam, you should understand:

  • LLMs
  • Small language models
  • Code models
  • Multimodal models
  • Deployment options
  • Model consumption patterns
  • Prompt engineering
  • Scaling strategies
  • Cost optimization
  • Responsible AI controls

Choosing the right model and deployment strategy is essential for building:

  • Scalable
  • Reliable
  • Efficient
  • Responsible AI solutions

These concepts are foundational for generative AI and agentic systems on Azure.

Practice Exam Questions

Question 1

What is a primary strength of large language models (LLMs)?

A. Minimal compute usage
B. Complex reasoning and broad knowledge
C. Guaranteed factual accuracy
D. Extremely low latency

Answer

B. Complex reasoning and broad knowledge

Explanation

LLMs excel at reasoning, conversation, and broad knowledge tasks.

Question 2

Which model type is best suited for lightweight, low-cost inference?

A. Large language model
B. Small language model
C. Multimodal model
D. Vision transformer only

Answer

B. Small language model

Explanation

SLMs are optimized for lower latency and reduced cost.

Question 3

Which model type is specifically optimized for programming tasks?

A. Vision model
B. Code model
C. Embedding model
D. Speech model

Answer

B. Code model

Explanation

Code models are trained for software development workflows.

Question 4

What is a defining feature of multimodal models?

A. They only process text
B. They process multiple input types
C. They eliminate inference costs
D. They require no prompting

Answer

B. They process multiple input types

Explanation

Multimodal models handle text, images, audio, and other media.

Question 5

Which deployment type is best for interactive AI chat applications?

A. Batch inference
B. Real-time inference
C. Archive deployment
D. Offline storage deployment

Answer

B. Real-time inference

Explanation

Interactive applications require low-latency real-time inference.

Question 6

What does the temperature parameter control?

A. Network throughput
B. Output randomness and creativity
C. Storage replication
D. GPU memory allocation

Answer

B. Output randomness and creativity

Explanation

Temperature affects how deterministic or creative outputs become.

Question 7

Which technique improves factual accuracy by using trusted data sources?

A. GPU scaling
B. Retrieval-Augmented Generation (RAG)
C. Semantic caching
D. Compression indexing

Answer

B. Retrieval-Augmented Generation (RAG)

Explanation

RAG grounds model outputs using retrieved enterprise data.

Question 8

What is a major benefit of streaming responses?

A. Reduced storage costs
B. Faster perceived response times
C. Elimination of monitoring
D. Improved vector indexing

Answer

B. Faster perceived response times

Explanation

Streaming improves user experience during response generation.

Question 9

Which authentication method supports passwordless access to Azure AI services?

A. Static credentials only
B. Managed identities
C. Anonymous access
D. Embedded API secrets in code

Answer

B. Managed identities

Explanation

Managed identities support secure, keyless authentication.

Question 10

Which model type is most appropriate for image understanding and OCR tasks?

A. Small language model
B. Multimodal model
C. Traditional relational database
D. Static rules engine

Answer

B. Multimodal model

Explanation

Multimodal models process images and text together.

Go to the AI-103 Exam Prep Hub main page

AI, AI-103, Artificial Intelligence (AI), Microsoft Certification

Configure model and agent deployments (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Plan and manage an Azure AI solution (25–30%)
   --> Set up AI solutions in Foundry
      --> Configure model and agent deployments

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

One of the most important responsibilities for Azure AI developers is configuring and managing model and agent deployments.

Modern AI applications depend on properly configured:

  • Large Language Models (LLMs)
  • Embedding models
  • Multimodal models
  • AI agents
  • Retrieval systems
  • Tool integrations
  • Orchestration workflows

The AI-103: Develop AI Apps and Agents on Azure certification exam tests your ability to configure AI solutions in Azure AI Foundry and related Azure services.

For the AI-103 exam, you should understand:

  • Azure OpenAI model deployments
  • Deployment types
  • Provisioned throughput
  • Model versioning
  • Deployment scaling
  • Agent configuration
  • Tool and function integration
  • Retrieval integration
  • Security configuration
  • Monitoring and evaluation
  • Deployment lifecycle management

What Is a Model Deployment?

A model deployment is a configured instance of an AI model that applications can access through APIs.

Deployments allow developers to:

  • Choose models
  • Configure capacity
  • Control scaling
  • Manage versions
  • Apply security controls
  • Monitor usage

A deployment acts as the operational endpoint for AI inference.

Azure AI Foundry

Azure AI Foundry provides tools and services for:

  • Deploying AI models
  • Configuring AI agents
  • Managing workflows
  • Evaluating AI systems
  • Monitoring AI applications

It integrates with:

  • Azure OpenAI
  • Azure AI Search
  • Prompt Flow
  • Azure AI Content Safety
  • Azure Functions

Types of Models in Azure AI

Common model types include:

  • Large Language Models (LLMs)
  • Small Language Models (SLMs)
  • Embedding models
  • Multimodal models
  • Vision models
  • Speech models

Large Language Models (LLMs)

LLMs are used for:

  • Chatbots
  • AI copilots
  • Summarization
  • Reasoning
  • Tool calling
  • Content generation

Examples include GPT-based models.

Embedding Models

Embedding models convert content into vector representations.

Used for:

  • Vector search
  • Semantic retrieval
  • Similarity matching
  • RAG systems

Multimodal Models

Multimodal models process multiple input types such as:

  • Text
  • Images
  • Audio
  • Documents

Used for:

  • Image analysis
  • Visual reasoning
  • OCR workflows
  • Multimodal agents

Azure OpenAI Deployments

Azure OpenAI deployments expose models through API endpoints.

Deployment configuration includes:

  • Model selection
  • Deployment name
  • Capacity allocation
  • Version selection
  • Region selection
  • Content filtering settings

Deployment Names

Each deployment has a unique deployment name.

Applications use the deployment name when making API requests.

Example:

  • gpt4-copilot-prod
  • embeddings-search-dev

Model Versioning

Models evolve over time.

Versioning helps:

  • Maintain stability
  • Test upgrades
  • Support rollback strategies
  • Compare model behavior

Why Model Versioning Matters

Different versions may:

  • Behave differently
  • Produce different outputs
  • Affect latency
  • Affect costs
  • Impact prompt performance

Deployment Types

Azure AI commonly supports:

  • Standard deployments
  • Provisioned throughput deployments

Standard Deployments

Standard deployments use shared infrastructure.

Advantages:

  • Simpler setup
  • Lower upfront costs
  • Flexible usage

Limitations:

  • Shared capacity
  • Variable latency under heavy load

Provisioned Throughput Deployments

Provisioned throughput reserves dedicated model capacity.

Advantages:

  • Predictable performance
  • Consistent latency
  • Enterprise-grade scaling

Limitations:

  • Higher cost
  • Capacity planning required

When to Use Standard Deployments

Use standard deployments when:

  • Workloads are moderate
  • Usage is variable
  • Cost optimization matters
  • Development/testing environments are used

When to Use Provisioned Throughput

Use provisioned throughput when:

  • High traffic is expected
  • Predictable latency is required
  • Enterprise SLAs exist
  • Production copilots are deployed

Scaling Model Deployments

AI deployments must support varying workloads.

Autoscaling

Autoscaling adjusts resources dynamically based on demand.

Benefits:

  • Improved performance
  • Better cost efficiency
  • Reduced manual intervention

Horizontal Scaling

Horizontal scaling adds additional instances or capacity.

Useful for:

  • High concurrency
  • Enterprise AI systems
  • Large-scale chatbots

Latency Considerations

Latency refers to response time.

Factors affecting latency:

  • Model size
  • Throughput load
  • Geographic distance
  • Retrieval pipelines
  • Tool execution

Choosing the Correct Model

Choosing the correct model is critical.

Use Larger Models When:

  • Advanced reasoning is required
  • Complex workflows exist
  • High-quality generation matters

Use Smaller Models When:

  • Cost efficiency matters
  • Low latency is important
  • Simpler tasks are performed

Agent Deployments

AI agents combine:

  • Models
  • Memory
  • Retrieval
  • Tool calling
  • Workflow orchestration

Agent deployment involves configuring all these components together.

Agent Configuration Components

Common agent configuration elements include:

  • System prompts
  • Tool definitions
  • Function calling
  • Knowledge sources
  • Retrieval settings
  • Memory configuration
  • Safety settings

System Prompts

System prompts define:

  • Agent behavior
  • Role instructions
  • Response style
  • Operational constraints

Well-designed system prompts improve:

  • Reliability
  • Consistency
  • Safety

Tool and Function Integration

Agents may use tools such as:

  • APIs
  • Databases
  • Search services
  • External systems

Function calling enables agents to invoke these tools dynamically.

Retrieval Integration

Many AI agents use Retrieval-Augmented Generation (RAG).

RAG systems commonly integrate:

  • Azure AI Search
  • Embedding models
  • Vector search
  • Knowledge indexes

Knowledge Sources

Agents may connect to:

  • Enterprise documents
  • Databases
  • APIs
  • SharePoint
  • Blob Storage
  • Internal knowledge bases

Memory Configuration

Agents may use:

  • Short-term memory
  • Long-term memory
  • Semantic memory

Common storage systems include:

  • Azure Cosmos DB
  • Azure SQL Database
  • Azure AI Search

Security Configuration

Security is a major AI-103 exam topic.

Microsoft Entra ID

Microsoft Entra ID supports:

  • Authentication
  • Authorization
  • RBAC
  • Identity management

Azure Key Vault

Azure Key Vault securely stores:

  • API keys
  • Secrets
  • Certificates
  • Connection strings

Content Safety Configuration

Azure AI Content Safety helps:

  • Detect harmful content
  • Filter unsafe outputs
  • Apply safety policies

Network Security

Enterprise AI deployments may use:

  • VNets
  • Private Endpoints
  • Firewalls
  • API gateways

Monitoring Deployments

AI deployments require operational monitoring.

Azure Monitor

Azure Monitor provides:

  • Metrics
  • Logging
  • Alerts
  • Diagnostics

Application Insights

Application Insights supports:

  • Telemetry
  • Request tracing
  • Error diagnostics
  • Performance monitoring

Metrics to Monitor

Common metrics include:

  • Latency
  • Token usage
  • Error rates
  • Throughput
  • Tool call failures
  • Retrieval quality

Evaluating AI Deployments

AI systems should be evaluated for:

  • Accuracy
  • Groundedness
  • Safety
  • Relevance
  • Reliability

Prompt Flow

Prompt Flow supports:

  • Workflow orchestration
  • Prompt chaining
  • Tool integration
  • Evaluation pipelines

Prompt Flow is an important AI-103 topic.

CI/CD for AI Deployments

AI deployment pipelines should support:

  • Automated testing
  • Version control
  • Safe releases
  • Rollbacks

Blue-Green Deployments

Blue-green deployments:

  • Reduce downtime
  • Support safer releases
  • Simplify rollback

Canary Deployments

Canary deployments:

  • Roll out changes gradually
  • Reduce deployment risk
  • Support controlled testing

Common AI-103 Deployment Scenarios

Scenario 1: Enterprise AI Copilot

Requirements:

  • High concurrency
  • Secure retrieval
  • Enterprise search
  • Low latency

Recommended Configuration:

  • Provisioned throughput
  • Azure AI Search
  • Entra ID
  • Autoscaling

Scenario 2: Development Chatbot

Requirements:

  • Low cost
  • Rapid experimentation
  • Flexible scaling

Recommended Configuration:

  • Standard deployment
  • App Service
  • Basic monitoring

Scenario 3: AI Agent with Tool Calling

Requirements:

  • API integrations
  • Workflow execution
  • Multi-step reasoning

Recommended Configuration:

  • Azure OpenAI
  • Azure Functions
  • Prompt Flow
  • Tool definitions

Scenario 4: Enterprise Knowledge Assistant

Requirements:

  • Grounded responses
  • Semantic retrieval
  • Document search

Recommended Configuration:

  • Embedding models
  • Azure AI Search
  • Hybrid search
  • RAG pipelines

Cost Optimization Considerations

AI deployments can become expensive.

Common Cost Drivers

  • Token usage
  • Provisioned throughput
  • Search indexing
  • Embedding generation
  • Large models
  • High concurrency

Cost Optimization Strategies

Use Smaller Models When Possible

Smaller models reduce:

  • Latency
  • Compute costs
  • Token usage

Optimize Retrieval

Efficient retrieval reduces:

  • Prompt size
  • Token costs
  • Latency

Use Autoscaling

Autoscaling prevents overprovisioning.

Common AI-103 Exam Tips

Understand Deployment Types

Know the differences between:

  • Standard deployments
  • Provisioned throughput deployments

Learn Agent Configuration Components

Understand:

  • System prompts
  • Tool integration
  • Retrieval settings
  • Memory configuration

Know Security Best Practices

Use:

  • Entra ID
  • RBAC
  • Key Vault
  • Private networking

Understand Monitoring Concepts

Know how to monitor:

  • Latency
  • Token usage
  • Throughput
  • Errors
  • AI quality

Summary

Configuring model and agent deployments is a critical skill for Azure AI developers.

For the AI-103 exam, you should understand:

  • Azure OpenAI deployment configuration
  • Model versioning
  • Deployment scaling
  • Agent architecture
  • Tool integration
  • Retrieval integration
  • Memory configuration
  • Security controls
  • Monitoring and evaluation
  • Deployment lifecycle management

Well-configured deployments improve:

  • Reliability
  • Performance
  • Scalability
  • Security
  • Cost efficiency
  • User experience

These concepts are foundational for building enterprise-grade AI applications and agent-based systems on Azure.

Practice Exam Questions

Question 1

Which deployment type provides dedicated capacity for Azure OpenAI workloads?

A. Shared deployment
B. Provisioned throughput deployment
C. Batch deployment
D. Basic deployment

Answer

B. Provisioned throughput deployment

Explanation

Provisioned throughput reserves dedicated processing capacity.

Question 2

What is the primary purpose of model versioning?

A. Increase storage size
B. Manage model updates and rollback strategies
C. Reduce API authentication
D. Eliminate monitoring

Answer

B. Manage model updates and rollback strategies

Explanation

Versioning helps maintain stability and supports rollback.

Question 3

Which Azure service is MOST commonly used for semantic retrieval in RAG systems?

A. Azure AI Search
B. Azure Backup
C. Azure CDN
D. Azure DNS

Answer

A. Azure AI Search

Explanation

Azure AI Search supports vector and semantic retrieval.

Question 4

What is the purpose of a system prompt in an AI agent?

A. Encrypt embeddings
B. Define agent behavior and instructions
C. Replace APIs
D. Configure storage replication

Answer

B. Define agent behavior and instructions

Explanation

System prompts guide the agent’s role, constraints, and response style.

Question 5

Which Azure service securely stores API keys and secrets?

A. Azure Key Vault
B. Azure Monitor
C. Azure Backup
D. Azure CDN

Answer

A. Azure Key Vault

Explanation

Azure Key Vault securely stores sensitive credentials.

Question 6

Which deployment strategy gradually rolls out updates to a small percentage of users first?

A. Full deployment
B. Canary deployment
C. Offline deployment
D. Batch deployment

Answer

B. Canary deployment

Explanation

Canary deployments reduce deployment risk through gradual rollout.

Question 7

Which type of model is specifically designed for vector generation and semantic similarity?

A. Vision model
B. Embedding model
C. Speech model
D. OCR model

Answer

B. Embedding model

Explanation

Embedding models generate vector representations for semantic retrieval.

Question 8

Which Azure service provides telemetry and request tracing for AI applications?

A. Application Insights
B. Azure DNS
C. Azure Files
D. Azure Firewall

Answer

A. Application Insights

Explanation

Application Insights provides application telemetry and diagnostics.

Question 9

Which feature dynamically adjusts resources based on workload demand?

A. Static allocation
B. Autoscaling
C. Encryption scaling
D. Semantic routing

Answer

B. Autoscaling

Explanation

Autoscaling automatically adjusts capacity based on traffic.

Question 10

Which Azure service is commonly used for workflow orchestration and prompt chaining in AI solutions?

A. Prompt Flow
B. Azure CDN
C. Azure Backup
D. Azure Front Door

Answer

A. Prompt Flow

Explanation

Prompt Flow orchestrates prompts, tools, and AI workflows.

Go to the AI-103 Exam Prep Hub main page

AI, AI-103, Artificial Intelligence (AI), Microsoft Certification

Choose appropriate deployment options (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Plan and manage an Azure AI solution (25–30%)
   --> Set up AI solutions in Foundry
      --> Choose appropriate deployment options

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

One of the most important responsibilities for Azure AI developers is selecting the correct deployment option for AI applications and agent-based solutions.

Modern AI systems can be deployed in many different ways depending on:

  • Scalability requirements
  • Cost constraints
  • Security requirements
  • Latency expectations
  • Geographic distribution
  • Operational complexity
  • AI workload patterns
  • Enterprise governance needs

The AI-103: Develop AI Apps and Agents on Azure certification exam tests your understanding of how to choose appropriate deployment options for:

  • Generative AI applications
  • AI agents
  • APIs
  • RAG systems
  • Vector search solutions
  • Multimodal applications
  • Enterprise AI systems

For the AI-103 exam, you should understand:

  • Azure deployment models
  • Hosting options
  • Serverless deployments
  • Containerized deployments
  • Kubernetes deployments
  • Regional deployments
  • High availability strategies
  • Scaling approaches
  • CI/CD deployment pipelines
  • Model deployment considerations
  • Infrastructure tradeoffs

What Is a Deployment Option?

A deployment option refers to the method used to host and run an AI application or service.

Deployment choices affect:

  • Performance
  • Reliability
  • Cost
  • Security
  • Scalability
  • Maintainability

Choosing the wrong deployment strategy can:

  • Increase costs
  • Reduce performance
  • Complicate maintenance
  • Create scaling problems

Common Azure AI Deployment Components

AI solutions commonly include:

  • AI models
  • APIs
  • Search systems
  • Databases
  • Agent orchestration
  • Storage systems
  • Monitoring tools
  • Security services

Each component may use different deployment approaches.

Azure OpenAI Deployment Options

Azure OpenAI allows developers to deploy:

  • GPT models
  • Embedding models
  • Multimodal models
  • Fine-tuned models

Deployment considerations include:

  • Region availability
  • Throughput requirements
  • Latency requirements
  • Cost optimization
  • Capacity planning

Standard Deployments

What Are Standard Deployments?

Standard deployments provide shared model hosting infrastructure.

Advantages:

  • Lower operational complexity
  • Managed infrastructure
  • Easier setup

Disadvantages:

  • Shared capacity
  • Potential throughput limitations

Provisioned Throughput Deployments

What Is Provisioned Throughput?

Provisioned throughput reserves dedicated processing capacity.

Advantages:

  • Predictable performance
  • Dedicated throughput
  • Lower latency consistency

Disadvantages:

  • Higher cost
  • Capacity planning required

When to Use Provisioned Throughput

Use provisioned throughput when:

  • Workloads are high volume
  • Predictable latency is critical
  • Enterprise SLAs are required
  • Large-scale copilots are deployed

Serverless Deployment Options

What Is Serverless?

Serverless computing automatically manages infrastructure.

Developers focus on code instead of servers.

Azure Functions

Azure Functions provides event-driven serverless compute.

Common AI use cases:

  • Tool calling
  • Workflow execution
  • API processing
  • Lightweight orchestration
  • Event-triggered AI actions

Advantages of Azure Functions

  • Automatic scaling
  • Pay-per-use pricing
  • Rapid deployment
  • Minimal infrastructure management

Limitations of Azure Functions

  • Execution duration limits
  • Cold starts
  • Less suitable for large persistent workloads

When to Use Azure Functions

Use Azure Functions when:

  • Workloads are event-driven
  • Execution is lightweight
  • Cost optimization is important
  • Rapid scaling is required

Azure Container Apps

Azure Container Apps provides serverless container hosting.

Useful for:

  • AI middleware
  • APIs
  • Agent orchestration
  • Background workers
  • Lightweight microservices

Advantages of Container Apps

  • Simplified container deployment
  • Autoscaling support
  • Event-driven scaling
  • Lower operational overhead than Kubernetes

Kubernetes Deployments

Azure Kubernetes Service (AKS)

AKS provides enterprise-grade container orchestration.

Common AI uses:

  • Multi-agent systems
  • Large-scale AI platforms
  • Distributed AI services
  • Complex orchestration
  • High-volume APIs

Advantages of AKS

  • High scalability
  • Advanced orchestration
  • Fine-grained control
  • Container portability
  • Enterprise-grade deployments

Limitations of AKS

  • Higher operational complexity
  • More infrastructure management
  • Requires Kubernetes expertise

When to Use AKS

Use AKS when:

  • Large-scale deployments exist
  • Multiple microservices interact
  • High traffic is expected
  • Advanced orchestration is needed

Platform-as-a-Service (PaaS) Deployments

Azure App Service

Azure App Service hosts:

  • Web apps
  • APIs
  • AI front ends
  • Lightweight enterprise applications

Advantages of App Service

  • Managed platform
  • Easy deployment
  • Autoscaling support
  • Simplified maintenance

When to Use App Service

Use App Service when:

  • Hosting AI web applications
  • Managing APIs
  • Rapid development is needed
  • Full Kubernetes orchestration is unnecessary

Edge and Hybrid Deployments

Some AI workloads require local or hybrid deployments.

Reasons include:

  • Low latency
  • Regulatory requirements
  • Limited connectivity
  • On-premises data processing

Azure Arc

Azure Arc extends Azure management to:

  • On-premises systems
  • Multi-cloud environments
  • Edge deployments

Useful for hybrid AI environments.

Deployment Considerations for AI Agents

AI agents often require multiple deployment layers.

Examples include:

  • LLM hosting
  • Retrieval systems
  • Tool execution services
  • Workflow orchestration
  • Persistent memory systems

Multi-Service Architectures

AI agents commonly use:

  • Azure OpenAI
  • Azure AI Search
  • Azure Functions
  • Cosmos DB
  • APIs
  • Orchestration workflows

Different components may use different deployment options.

Geographic Deployment Considerations

AI systems may require global deployment strategies.

Regional Deployments

Deploying resources in a specific region helps:

  • Reduce latency
  • Meet compliance requirements
  • Improve user experience

Multi-Region Deployments

Multi-region deployments improve:

  • Availability
  • Disaster recovery
  • Global performance

Availability Zones

Availability Zones provide redundancy across isolated datacenters.

Benefits include:

  • Higher uptime
  • Fault tolerance
  • Improved resilience

High Availability Design

Enterprise AI applications often require:

  • Redundant infrastructure
  • Automatic failover
  • Load balancing
  • Disaster recovery

Load Balancing

Azure Load Balancer and Azure Application Gateway distribute traffic across services.

Benefits:

  • Scalability
  • High availability
  • Traffic optimization

Autoscaling

Autoscaling dynamically adjusts infrastructure based on demand.

Supported by:

  • AKS
  • Azure Functions
  • App Service
  • Container Apps

Deployment Security Considerations

Security is a major AI-103 exam topic.

Microsoft Entra ID

Microsoft Entra ID supports:

  • Authentication
  • Authorization
  • Identity management
  • RBAC

Azure Key Vault

Azure Key Vault securely stores:

  • Secrets
  • API keys
  • Certificates
  • Connection strings

Private Endpoints

Private Endpoints provide secure private connectivity between Azure services.

Useful for:

  • Enterprise AI systems
  • Sensitive data workloads
  • Compliance-driven deployments

CI/CD for AI Deployments

What Is CI/CD?

CI/CD stands for:

  • Continuous Integration
  • Continuous Deployment

CI/CD automates:

  • Testing
  • Deployment
  • Validation
  • Release management

Azure DevOps

Azure DevOps supports:

  • Build pipelines
  • Release pipelines
  • Source control
  • Automated deployments

GitHub Actions

GitHub Actions supports:

  • Workflow automation
  • CI/CD pipelines
  • Deployment automation

Commonly used for AI application deployments.

Blue-Green Deployments

Blue-green deployments reduce downtime during releases.

How it works:

  • One environment remains active
  • A second environment receives updates
  • Traffic shifts after validation

Benefits:

  • Safer releases
  • Reduced downtime
  • Easier rollback

Canary Deployments

Canary deployments release updates gradually to a small percentage of users.

Benefits:

  • Reduced deployment risk
  • Easier issue detection
  • Safer experimentation

Monitoring Deployment Health

AI deployments should monitor:

  • Latency
  • Throughput
  • Token usage
  • Errors
  • Model failures
  • Tool call failures
  • Retrieval quality

Azure Monitor

Azure Monitor provides:

  • Metrics
  • Logging
  • Alerts
  • Diagnostics

Application Insights

Application Insights supports:

  • Telemetry
  • Request tracing
  • Dependency tracking
  • Error diagnostics

Cost Optimization Considerations

AI deployments can become expensive.

Common Cost Drivers

  • Token consumption
  • GPU usage
  • High-scale orchestration
  • Search indexing
  • Storage
  • Data transfer

Cost Optimization Strategies

Use Smaller Models When Appropriate

Smaller models reduce:

  • Compute costs
  • Token usage
  • Latency

Use Serverless When Appropriate

Serverless deployments reduce idle infrastructure costs.

Use Autoscaling

Autoscaling prevents overprovisioning.

Common AI-103 Deployment Scenarios

Scenario 1: Enterprise AI Chatbot

Requirements:

  • High availability
  • Secure authentication
  • Enterprise search

Recommended Deployment:

  • Azure OpenAI
  • App Service
  • Azure AI Search
  • Entra ID

Scenario 2: Large-Scale AI Agent Platform

Requirements:

  • Multiple AI agents
  • Heavy orchestration
  • High concurrency

Recommended Deployment:

  • AKS
  • Azure Functions
  • Cosmos DB
  • Prompt Flow

Scenario 3: Lightweight AI API

Requirements:

  • Rapid deployment
  • Cost optimization
  • Moderate scale

Recommended Deployment:

  • Azure Functions
  • Container Apps

Scenario 4: Global AI Application

Requirements:

  • Global users
  • Low latency
  • Disaster recovery

Recommended Deployment:

  • Multi-region deployment
  • Availability Zones
  • Load balancing

Common AI-103 Exam Tips

Understand Deployment Tradeoffs

Know when to use:

  • App Service vs AKS
  • Functions vs Containers
  • Standard vs Provisioned Throughput

Know High Availability Concepts

Understand:

  • Availability Zones
  • Multi-region deployments
  • Load balancing
  • Failover strategies

Learn Security Best Practices

Know how to use:

  • Entra ID
  • RBAC
  • Key Vault
  • Private Endpoints

Understand Agent Deployment Needs

AI agents commonly require:

  • Tool orchestration
  • Retrieval systems
  • Persistent memory
  • API integrations

Summary

Choosing the correct deployment option is critical for successful AI applications and agent-based systems.

For the AI-103 exam, you should understand:

  • Azure deployment models
  • Serverless deployment options
  • Kubernetes deployments
  • PaaS hosting options
  • Multi-region architectures
  • High availability design
  • Security considerations
  • CI/CD pipelines
  • Scaling strategies
  • AI deployment tradeoffs

Strong deployment architecture skills help ensure AI systems are:

  • Reliable
  • Scalable
  • Secure
  • Cost-effective
  • Maintainable

Practice Exam Questions

Question 1

Which Azure service is BEST suited for enterprise-scale container orchestration for AI applications?

A. Azure App Service
B. Azure Kubernetes Service (AKS)
C. Azure DNS
D. Azure Backup

Answer

B. Azure Kubernetes Service (AKS)

Explanation

AKS provides enterprise-grade container orchestration and scalability.

Question 2

Which deployment option provides dedicated throughput capacity for Azure OpenAI models?

A. Shared deployment
B. Provisioned throughput deployment
C. Consumption deployment
D. Basic deployment

Answer

B. Provisioned throughput deployment

Explanation

Provisioned throughput reserves dedicated model processing capacity.

Question 3

Which Azure service is MOST appropriate for lightweight event-driven AI workflows?

A. Azure Functions
B. Azure Firewall
C. Azure Backup
D. Azure CDN

Answer

A. Azure Functions

Explanation

Azure Functions supports serverless event-driven execution.

Question 4

What is the primary benefit of Availability Zones?

A. Lower token usage
B. Increased embedding size
C. Improved fault tolerance
D. Reduced API authentication

Answer

C. Improved fault tolerance

Explanation

Availability Zones provide redundancy across isolated datacenters.

Question 5

Which Azure service is commonly used to host AI web applications and APIs with minimal infrastructure management?

A. Azure App Service
B. Azure Load Balancer
C. Azure DNS
D. Azure Monitor

Answer

A. Azure App Service

Explanation

Azure App Service is a managed PaaS platform for hosting web applications and APIs.

Question 6

Which deployment strategy gradually releases updates to a subset of users first?

A. Blue-green deployment
B. Canary deployment
C. Full rollback deployment
D. Batch deployment

Answer

B. Canary deployment

Explanation

Canary deployments release updates incrementally to reduce risk.

Question 7

Which Azure service securely stores API keys and secrets for AI applications?

A. Azure Key Vault
B. Azure CDN
C. Azure Firewall
D. Azure Backup

Answer

A. Azure Key Vault

Explanation

Azure Key Vault securely manages secrets and credentials.

Question 8

Which Azure deployment option is MOST appropriate for serverless container hosting?

A. Azure Container Apps
B. Azure Backup
C. Azure DNS
D. Azure Files

Answer

A. Azure Container Apps

Explanation

Azure Container Apps provides simplified serverless container deployment.

Question 9

Which deployment architecture improves global application availability and disaster recovery?

A. Single-region deployment
B. Multi-region deployment
C. Local-only deployment
D. Single-container deployment

Answer

B. Multi-region deployment

Explanation

Multi-region deployments improve resilience and geographic performance.

Question 10

Which Azure monitoring service provides application telemetry and request diagnostics?

A. Application Insights
B. Azure CDN
C. Azure DNS
D. Azure Policy

Answer

A. Application Insights

Explanation

Application Insights provides monitoring and telemetry for applications.

Go to the AI-103 Exam Prep Hub main page

AI, AI-103, Artificial Intelligence (AI), Microsoft Certification

Design Azure infrastructure for AI Apps and agent-based solutions (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Plan and manage an Azure AI solution (25–30%)
   --> Set up AI solutions in Foundry
      --> Design Azure infrastructure for AI Apps and agent-based solutions

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

Designing infrastructure for AI applications and agent-based systems is one of the most important responsibilities for Azure AI developers.

Modern AI solutions are not simply standalone models. They are distributed cloud systems that combine:

  • AI services
  • APIs
  • Databases
  • Search systems
  • Storage
  • Networking
  • Security controls
  • Monitoring systems
  • Agent orchestration components

The AI-103: Develop AI Apps and Agents on Azure certification exam tests your ability to design Azure infrastructure that supports:

  • Generative AI applications
  • AI agents
  • Retrieval-Augmented Generation (RAG)
  • Vector search
  • Multimodal AI systems
  • Scalable AI architectures
  • Secure enterprise AI deployments

For the AI-103 exam, you should understand:

  • Core Azure infrastructure services
  • AI architecture patterns
  • Scalability and performance design
  • Networking and security
  • Identity and access management
  • Storage and databases
  • Monitoring and observability
  • Cost optimization
  • High availability and disaster recovery
  • Infrastructure choices for AI agents

Core Components of AI Infrastructure

AI applications commonly require multiple infrastructure layers.

Typical components include:

  1. AI model services
  2. Compute resources
  3. Storage systems
  4. Search and retrieval systems
  5. Networking components
  6. Security services
  7. Monitoring systems
  8. Workflow orchestration
  9. API management
  10. Identity management

Azure AI Services Layer

Azure OpenAI

Azure OpenAI provides:

  • Large Language Models (LLMs)
  • Embedding models
  • Multimodal models
  • Conversational AI capabilities

Azure OpenAI is commonly used for:

  • AI copilots
  • Chatbots
  • AI agents
  • Summarization
  • Content generation
  • Tool calling

Azure AI Search

Azure AI Search supports:

  • Vector search
  • Semantic search
  • Hybrid search
  • Enterprise retrieval
  • RAG architectures

It is commonly used for:

  • Knowledge grounding
  • Enterprise search
  • AI assistant retrieval

Azure AI Vision

Azure AI Vision provides:

  • OCR
  • Image analysis
  • Object detection
  • Caption generation
  • Visual understanding

Azure AI Document Intelligence

Azure AI Document Intelligence supports:

  • Invoice extraction
  • Form processing
  • Layout analysis
  • OCR workflows
  • Structured document extraction

Compute Infrastructure for AI Applications

Azure App Service

Azure App Service is commonly used to host:

  • Web applications
  • AI front ends
  • APIs
  • Lightweight AI services

Advantages:

  • Managed platform
  • Easy scaling
  • Simplified deployment

Azure Kubernetes Service (AKS)

AKS provides container orchestration for:

  • Large-scale AI applications
  • Microservices
  • Agent orchestration systems
  • Distributed AI workloads

Advantages:

  • High scalability
  • Container management
  • Advanced orchestration
  • Enterprise-grade deployments

When to Use AKS

Use AKS when:

  • Complex orchestration is required
  • Multiple services interact
  • High scalability is needed
  • Microservice architectures are used

Azure Functions

Azure Functions provides serverless compute.

Common AI use cases:

  • Tool execution
  • Event-driven workflows
  • API integrations
  • Lightweight processing
  • Agent tool calling

Advantages:

  • Pay-per-use pricing
  • Automatic scaling
  • Fast development

Azure Container Apps

Azure Container Apps provides simplified container hosting.

Useful for:

  • API services
  • AI middleware
  • Lightweight agent services
  • Event-driven AI components

Choosing the Correct Compute Service

Use Azure App Service When:

  • Hosting simple AI web apps
  • Managing APIs
  • Rapid deployment is needed

Use AKS When:

  • Large-scale orchestration is required
  • Complex microservices exist
  • Advanced scalability is necessary

Use Azure Functions When:

  • Event-driven execution is needed
  • Tool calling is required
  • Lightweight compute is sufficient

Use Azure Container Apps When:

  • Container simplicity is preferred
  • Serverless containers are desired

Storage Infrastructure

AI systems often require multiple storage solutions.

Azure Blob Storage

Azure Blob Storage supports:

  • Document storage
  • Training data
  • Images
  • Videos
  • Logs
  • AI datasets

Common AI uses:

  • RAG document storage
  • Knowledge repositories
  • Media storage

Azure Cosmos DB

Azure Cosmos DB provides:

  • Globally distributed NoSQL storage
  • Low-latency access
  • High scalability

Common AI uses:

  • Agent memory
  • Session storage
  • User profiles
  • Conversation history

Azure SQL Database

Azure SQL Database supports:

  • Structured enterprise data
  • Relational workloads
  • Transactional systems

Common AI uses:

  • Enterprise integration
  • Business systems
  • Structured metadata

Vector Storage

Vector-enabled storage supports:

  • Embedding storage
  • Similarity search
  • Semantic retrieval

Common services include:

  • Azure AI Search
  • Azure Cosmos DB
  • Azure SQL Database

Networking Infrastructure

AI solutions require secure and scalable networking.

Virtual Networks (VNets)

VNets provide:

  • Network isolation
  • Secure communication
  • Private connectivity

Use VNets when:

  • Enterprise security is required
  • Private networking is necessary
  • Sensitive data is involved

Private Endpoints

Private Endpoints allow Azure services to be accessed privately through VNets.

Benefits:

  • Improved security
  • Reduced public exposure
  • Enterprise compliance support

API Management

Azure API Management helps:

  • Secure APIs
  • Throttle requests
  • Monitor API usage
  • Apply policies
  • Manage agent APIs

This is important for:

  • AI agents
  • Tool integrations
  • Enterprise API governance

Load Balancing

Azure Load Balancer and Application Gateway help:

  • Distribute traffic
  • Improve availability
  • Scale AI applications

Identity and Security

Security is a major AI-103 exam topic.

Microsoft Entra ID

Microsoft Entra ID provides:

  • Authentication
  • Authorization
  • Identity management
  • Role-based access control (RBAC)

AI applications use Entra ID for:

  • User authentication
  • API access control
  • Secure enterprise integration

Role-Based Access Control (RBAC)

RBAC ensures users and services only access authorized resources.

Examples:

  • Restricting AI model access
  • Controlling storage access
  • Securing search indexes

Azure Key Vault

Azure Key Vault stores:

  • Secrets
  • API keys
  • Certificates
  • Connection strings

Never hardcode secrets in AI applications.

Azure AI Content Safety

Azure AI Content Safety helps:

  • Detect harmful content
  • Filter unsafe outputs
  • Support responsible AI practices

Monitoring and Observability

AI systems require monitoring for:

  • Reliability
  • Performance
  • Cost
  • Failures
  • Hallucinations
  • API latency

Azure Monitor

Azure Monitor collects:

  • Metrics
  • Logs
  • Alerts
  • Performance data

Application Insights

Application Insights supports:

  • Application telemetry
  • Request tracing
  • Error tracking
  • Dependency monitoring

Useful for:

  • AI apps
  • APIs
  • Agent workflows

Logging AI Systems

AI systems should log:

  • Prompts
  • Responses
  • Errors
  • Tool calls
  • Latency
  • Retrieval quality

Logging helps:

  • Troubleshooting
  • Auditing
  • Evaluation
  • Compliance

Scalability Design

AI applications may experience:

  • High traffic
  • Large token volumes
  • Heavy retrieval workloads
  • Concurrent agent operations

Infrastructure must scale effectively.

Horizontal Scaling

Horizontal scaling adds more instances.

Examples:

  • Additional API servers
  • More containers
  • More worker nodes

Vertical Scaling

Vertical scaling increases resource capacity.

Examples:

  • More CPU
  • More memory
  • Larger VM sizes

Autoscaling

Autoscaling dynamically adjusts resources based on demand.

Common services supporting autoscaling:

  • AKS
  • Azure Functions
  • App Service
  • Container Apps

High Availability and Disaster Recovery

Enterprise AI systems require resilience.

Availability Zones

Availability Zones improve fault tolerance.

Benefits:

  • Redundancy
  • Improved uptime
  • Reduced outage risk

Geo-Redundancy

Geo-redundancy replicates data across regions.

Useful for:

  • Disaster recovery
  • Business continuity
  • Global applications

Backup and Recovery

AI systems should back up:

  • Knowledge indexes
  • Databases
  • Configuration data
  • Logs
  • Agent memory

Infrastructure for AI Agents

AI agents often require additional infrastructure components.

Agent Orchestration

AI agents may require orchestration services such as:

  • Prompt Flow
  • Azure Functions
  • Logic Apps
  • AKS workflows

Retrieval Infrastructure

Agent systems commonly use:

  • Azure AI Search
  • Embeddings
  • Vector indexes
  • RAG pipelines

Persistent Memory Infrastructure

Persistent memory may use:

  • Azure Cosmos DB
  • Azure SQL Database
  • Blob Storage

Tool Integration Infrastructure

Agents often integrate with:

  • REST APIs
  • Databases
  • External SaaS systems
  • Enterprise workflows

Common AI-103 Architecture Scenarios

Scenario 1: Enterprise AI Copilot

Requirements:

  • Conversational AI
  • Enterprise search
  • Secure authentication
  • Document retrieval

Recommended Infrastructure:

  • Azure OpenAI
  • Azure AI Search
  • Entra ID
  • Blob Storage
  • App Service

Scenario 2: Large-Scale Multi-Agent System

Requirements:

  • Multiple AI agents
  • High scalability
  • Distributed orchestration

Recommended Infrastructure:

  • AKS
  • Azure Functions
  • Prompt Flow
  • Cosmos DB

Scenario 3: AI Invoice Processing Solution

Requirements:

  • OCR
  • Document extraction
  • Workflow automation

Recommended Infrastructure:

  • Azure AI Document Intelligence
  • Blob Storage
  • Logic Apps
  • Azure Functions

Scenario 4: Global AI Chat Platform

Requirements:

  • Global availability
  • High concurrency
  • Disaster recovery

Recommended Infrastructure:

  • Geo-redundant storage
  • Availability Zones
  • Load balancing
  • Autoscaling

Cost Optimization Considerations

AI infrastructure can become expensive.

Common Cost Drivers

  • Token usage
  • Vector storage
  • GPU workloads
  • Data transfer
  • Search indexing
  • High-scale orchestration

Cost Optimization Strategies

Use Smaller Models When Appropriate

Smaller models reduce:

  • Compute usage
  • Token costs
  • Latency

Use Autoscaling

Autoscaling reduces idle resource costs.

Optimize Retrieval Pipelines

Efficient chunking and indexing reduce:

  • Search costs
  • Storage requirements
  • Retrieval latency

Common AI-103 Exam Tips

Understand Infrastructure Tradeoffs

Know when to use:

  • AKS vs App Service
  • Functions vs Containers
  • Cosmos DB vs SQL Database

Learn Security Best Practices

Know how to use:

  • Entra ID
  • RBAC
  • Key Vault
  • Private Endpoints

Understand RAG Infrastructure

RAG commonly uses:

  • Azure OpenAI
  • Azure AI Search
  • Embeddings
  • Storage systems

Know Agent Infrastructure Patterns

AI agents commonly require:

  • Workflow orchestration
  • Tool integration
  • Persistent memory
  • Retrieval systems

Summary

Designing Azure infrastructure for AI applications requires balancing:

  • Scalability
  • Security
  • Performance
  • Cost
  • Reliability
  • Maintainability

For the AI-103 exam, you should understand:

  • Azure AI service architecture
  • Compute options
  • Storage design
  • Networking and security
  • Monitoring and observability
  • High availability
  • Agent infrastructure patterns
  • RAG infrastructure
  • Infrastructure scaling strategies

Strong infrastructure design skills are essential for deploying production-grade AI apps and agent-based systems on Azure.

Practice Exam Questions

Question 1

Which Azure service is MOST appropriate for enterprise vector search and RAG retrieval?

A. Azure AI Search
B. Azure Backup
C. Azure CDN
D. Azure DNS

Answer

A. Azure AI Search

Explanation

Azure AI Search supports vector search, semantic search, and retrieval for RAG systems.

Question 2

Which Azure compute service is BEST suited for large-scale containerized AI microservices?

A. Azure App Service
B. Azure Kubernetes Service (AKS)
C. Azure Files
D. Azure CDN

Answer

B. Azure Kubernetes Service (AKS)

Explanation

AKS provides advanced container orchestration and scalability.

Question 3

Which Azure service is MOST appropriate for storing API keys and secrets securely?

A. Azure Key Vault
B. Azure Monitor
C. Azure DNS
D. Azure Load Balancer

Answer

A. Azure Key Vault

Explanation

Azure Key Vault securely stores secrets, certificates, and keys.

Question 4

Which Azure service provides serverless execution for lightweight AI workflows and tool calling?

A. Azure Functions
B. Azure Backup
C. Azure CDN
D. Azure Firewall

Answer

A. Azure Functions

Explanation

Azure Functions supports event-driven serverless compute.

Question 5

What is the primary purpose of Availability Zones?

A. Reduce token usage
B. Improve fault tolerance and uptime
C. Replace backups
D. Encrypt embeddings

Answer

B. Improve fault tolerance and uptime

Explanation

Availability Zones provide redundancy across isolated datacenter locations.

Question 6

Which Azure service is MOST commonly used for globally distributed NoSQL storage in AI applications?

A. Azure Cosmos DB
B. Azure DNS
C. Azure Files
D. Azure CDN

Answer

A. Azure Cosmos DB

Explanation

Azure Cosmos DB provides scalable globally distributed NoSQL storage.

Question 7

Which Azure networking feature enables private access to Azure services from a VNet?

A. Private Endpoint
B. Public IP
C. Load Balancer
D. Traffic Manager

Answer

A. Private Endpoint

Explanation

Private Endpoints provide secure private connectivity.

Question 8

Which Azure monitoring service provides application telemetry and request tracing?

A. Application Insights
B. Azure CDN
C. Azure Policy
D. Azure ExpressRoute

Answer

A. Application Insights

Explanation

Application Insights provides telemetry and diagnostics for applications.

Question 9

Which Azure identity service provides authentication and RBAC support for AI applications?

A. Microsoft Entra ID
B. Azure CDN
C. Azure Firewall
D. Azure Front Door

Answer

A. Microsoft Entra ID

Explanation

Microsoft Entra ID provides identity and access management.

Question 10

Which scaling strategy adds additional instances to support increased AI workload demand?

A. Vertical scaling
B. Horizontal scaling
C. Encryption scaling
D. Semantic scaling

Answer

B. Horizontal scaling

Explanation

Horizontal scaling adds more instances to distribute workloads.

Go to the AI-103 Exam Prep Hub main page

AI, AI-103, Artificial Intelligence (AI), Microsoft Certification

Choose appropriate memory, tool, and knowledge integration services for agent solutions (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Plan and manage an Azure AI solution (25–30%)
   --> Choose the appropriate Foundry services for generative AI and agents
      --> Choose appropriate memory, tool, and knowledge integration services for agent solutions

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

Modern AI agents are far more advanced than traditional chatbots.

AI agents can:

  • Reason through problems
  • Plan tasks
  • Access tools
  • Retrieve knowledge
  • Maintain conversational memory
  • Execute workflows
  • Interact with enterprise systems
  • Coordinate multi-step operations

The AI-103: Develop AI Apps and Agents on Azure certification exam places significant emphasis on understanding how to design and implement these agent capabilities using Azure AI Foundry and related Azure services.

One of the most important skills tested on the exam is the ability to choose appropriate:

  • Memory systems
  • Tool integration services
  • Knowledge integration services
  • Retrieval architectures
  • Agent orchestration tools

For the AI-103 exam, you should understand:

  • Different types of agent memory
  • Tool calling and function calling
  • Retrieval-Augmented Generation (RAG)
  • Knowledge grounding
  • Azure AI Search integration
  • Agent orchestration workflows
  • External API integration
  • Vector search and embeddings
  • Enterprise knowledge integration
  • Security and governance considerations

What Are AI Agents?

AI agents are AI-powered systems capable of:

  • Interpreting goals
  • Planning actions
  • Using tools
  • Retrieving information
  • Maintaining context
  • Completing tasks autonomously or semi-autonomously

Unlike traditional chatbots, AI agents can:

  • Interact with APIs
  • Execute workflows
  • Use memory
  • Retrieve enterprise knowledge
  • Chain actions together
  • Adapt dynamically to user requests

Components of an AI Agent Architecture

Modern AI agent solutions commonly include:

  1. Large Language Models (LLMs)
  2. Memory systems
  3. Retrieval systems
  4. Knowledge integration
  5. Tool and function calling
  6. Workflow orchestration
  7. Security and governance controls

Azure AI Foundry and Agent Solutions

Azure AI Foundry provides services and tools that help developers:

  • Build AI agents
  • Integrate tools
  • Connect enterprise knowledge
  • Implement RAG
  • Orchestrate workflows
  • Evaluate agent behavior
  • Monitor AI systems

Core services often include:

  • Azure OpenAI
  • Azure AI Search
  • Prompt Flow
  • Azure AI Content Safety
  • Azure Functions
  • Azure Logic Apps
  • Azure Cosmos DB
  • Azure SQL Database

Memory in AI Agents

What Is Agent Memory?

Memory enables AI agents to retain and use information over time.

Memory allows agents to:

  • Maintain conversational context
  • Remember user preferences
  • Track workflow state
  • Store historical interactions
  • Support long-running tasks

Without memory, every interaction becomes isolated.

Types of Agent Memory

The AI-103 exam may test multiple memory types.

Short-Term Memory

What Is Short-Term Memory?

Short-term memory stores temporary conversational context.

Examples:

  • Current chat history
  • Active task context
  • Immediate instructions

Characteristics of Short-Term Memory

  • Session-based
  • Temporary
  • Fast access
  • Often stored in prompts or session state

When to Use Short-Term Memory

Use short-term memory for:

  • Conversational continuity
  • Current workflow tracking
  • Multi-turn conversations

Long-Term Memory

What Is Long-Term Memory?

Long-term memory stores persistent information across sessions.

Examples:

  • User preferences
  • Historical interactions
  • Persistent profiles
  • Prior decisions

Characteristics of Long-Term Memory

  • Persistent storage
  • Cross-session continuity
  • Larger storage capacity
  • Supports personalization

Azure Services for Long-Term Memory

Common services include:

  • Azure Cosmos DB
  • Azure SQL Database
  • Azure Storage
  • Vector databases

When to Use Long-Term Memory

Use long-term memory when:

  • Personalization is required
  • User preferences must persist
  • Historical context matters
  • Long-running workflows exist

Semantic Memory

What Is Semantic Memory?

Semantic memory stores knowledge in embeddings or vectorized formats.

This enables:

  • Semantic retrieval
  • Knowledge recall
  • Contextual understanding
  • Similarity matching

Semantic Memory in AI Agents

Semantic memory often uses:

  • Embedding models
  • Vector search
  • Azure AI Search

This allows agents to retrieve relevant information dynamically.

Episodic Memory

What Is Episodic Memory?

Episodic memory stores records of past interactions and events.

Examples:

  • Past conversations
  • Completed workflows
  • User activity history

This helps agents maintain continuity across interactions.

Choosing the Correct Memory Type

Use Short-Term Memory When:

  • Managing active conversations
  • Maintaining immediate context
  • Supporting temporary tasks

Use Long-Term Memory When:

  • Storing persistent user information
  • Personalizing experiences
  • Maintaining history across sessions

Use Semantic Memory When:

  • Retrieving knowledge semantically
  • Supporting RAG
  • Performing contextual retrieval

Use Episodic Memory When:

  • Tracking prior interactions
  • Supporting historical continuity

Knowledge Integration

What Is Knowledge Integration?

Knowledge integration connects AI agents to external information sources.

Examples:

  • Enterprise documents
  • Databases
  • Knowledge bases
  • APIs
  • Websites
  • Internal systems

Knowledge integration helps agents:

  • Provide grounded answers
  • Access current information
  • Reduce hallucinations
  • Support enterprise use cases

Retrieval-Augmented Generation (RAG)

What Is RAG?

RAG combines:

  • Retrieval systems
  • Search indexes
  • Embeddings
  • LLMs

RAG enables agents to retrieve external information before generating responses.

Azure AI Search for Knowledge Integration

Azure AI Search is a core service for:

  • Vector search
  • Semantic search
  • Hybrid search
  • Enterprise retrieval
  • Knowledge grounding

It enables agents to:

  • Search enterprise documents
  • Retrieve semantically relevant content
  • Access indexed knowledge

Hybrid Search

Hybrid search combines:

  • Keyword search
  • Semantic ranking
  • Vector search

Hybrid search is often the preferred approach for enterprise AI agents.

Embeddings and Knowledge Retrieval

Embedding models convert content into vector representations.

Embeddings support:

  • Semantic similarity
  • Vector retrieval
  • Knowledge recall
  • RAG pipelines

Azure OpenAI embedding models are commonly used.

Knowledge Sources for AI Agents

AI agents may integrate with:

  • Azure Blob Storage
  • SharePoint
  • Databases
  • REST APIs
  • Enterprise document repositories
  • CRM systems
  • ERP systems

Tool Integration

What Is Tool Integration?

Tool integration enables AI agents to interact with external systems.

Examples include:

  • APIs
  • Databases
  • Email systems
  • Calendars
  • Search services
  • Workflow systems

Tool integration allows agents to perform actions instead of only generating text.

Tool Calling and Function Calling

LLMs can invoke:

  • Tools
  • Functions
  • APIs

Examples:

  • Retrieve weather data
  • Send emails
  • Query databases
  • Create support tickets
  • Execute workflows

Azure Services for Tool Integration

Common services include:

  • Azure Functions
  • Azure Logic Apps
  • REST APIs
  • Azure API Management

Azure Functions

Azure Functions provides serverless compute for:

  • API integrations
  • Business logic
  • Event-driven workflows
  • Tool execution

AI agents often call Azure Functions to execute tasks.

Azure Logic Apps

Azure Logic Apps supports:

  • Workflow automation
  • Enterprise integrations
  • Connector-based orchestration

Logic Apps are useful when:

  • Multiple systems must interact
  • Low-code orchestration is preferred
  • Enterprise automation is needed

Azure API Management

Azure API Management helps:

  • Secure APIs
  • Manage API access
  • Monitor API usage
  • Apply governance policies

Useful for enterprise AI agent integrations.

Prompt Flow

Prompt Flow is a Foundry tool for:

  • Building AI workflows
  • Orchestrating prompts
  • Chaining tools
  • Managing agent pipelines
  • Evaluating workflows

Prompt Flow is a major AI-103 exam topic.

Multi-Agent Systems

Some AI architectures use multiple specialized agents.

Examples:

  • Research agent
  • Scheduling agent
  • Data retrieval agent
  • Customer service agent

Multi-agent systems may improve:

  • Scalability
  • Specialization
  • Workflow separation

Orchestration Services

Agent orchestration coordinates:

  • Memory
  • Retrieval
  • Tool execution
  • Workflow management

Common orchestration tools include:

  • Prompt Flow
  • Azure Functions
  • Logic Apps
  • Custom orchestration frameworks

Security and Governance

AI agent systems require:

  • Authentication
  • Authorization
  • Data protection
  • Content filtering
  • Responsible AI controls

Azure AI Content Safety

Azure AI Content Safety helps:

  • Detect harmful content
  • Prevent unsafe outputs
  • Support responsible AI deployments

Role-Based Access Control (RBAC)

RBAC ensures agents only access authorized resources.

This is especially important for:

  • Enterprise knowledge systems
  • Confidential data
  • Regulated environments

Monitoring and Observability

AI agent systems should monitor:

  • Tool usage
  • Latency
  • Errors
  • Retrieval quality
  • Hallucinations
  • Token usage

Monitoring improves:

  • Reliability
  • Performance
  • Troubleshooting

Common AI-103 Scenarios

Scenario 1: Enterprise Copilot

Requirements:

  • Access enterprise documents
  • Remember user preferences
  • Retrieve current information
  • Support conversational interactions

Recommended Services:

  • Azure OpenAI
  • Azure AI Search
  • Embedding models
  • Long-term memory storage

Scenario 2: AI Travel Assistant

Requirements:

  • Access calendars
  • Book hotels
  • Query APIs
  • Manage workflows

Recommended Services:

  • Azure OpenAI
  • Tool/function calling
  • Azure Functions
  • Prompt Flow

Scenario 3: Customer Support Agent

Requirements:

  • Retrieve support documents
  • Track prior interactions
  • Escalate tickets

Recommended Services:

  • Azure AI Search
  • Episodic memory
  • Azure Functions
  • CRM integration

Scenario 4: Personalized Learning Assistant

Requirements:

  • Remember learning preferences
  • Track progress
  • Recommend materials

Recommended Services:

  • Long-term memory
  • Semantic retrieval
  • Azure Cosmos DB

Common AI-103 Exam Tips

Understand Memory Types

Know the differences between:

  • Short-term memory
  • Long-term memory
  • Semantic memory
  • Episodic memory

Know When to Use RAG

Use RAG when:

  • External knowledge is required
  • Current data is needed
  • Hallucination reduction matters

Learn Tool Calling Concepts

Agents use:

  • Function calling
  • APIs
  • Workflows
  • Tool orchestration

This is commonly tested.

Understand Azure Service Roles

Azure AI Search

Used for:

  • Retrieval
  • Vector search
  • Grounding

Azure Functions

Used for:

  • Executing logic
  • Tool integration

Prompt Flow

Used for:

  • Workflow orchestration
  • Agent pipelines

Azure Cosmos DB

Used for:

  • Persistent memory
  • Long-term storage

Summary

AI agents require more than just language models.

Successful agent solutions combine:

  • Memory systems
  • Retrieval systems
  • Knowledge grounding
  • Tool integration
  • Workflow orchestration
  • Security controls

For the AI-103 exam, you should understand:

  • Different memory architectures
  • Tool and function calling
  • RAG workflows
  • Azure AI Search integration
  • Knowledge retrieval strategies
  • Prompt Flow orchestration
  • Persistent memory services
  • Enterprise AI integration patterns

Understanding how these services work together is critical for building scalable and intelligent AI agent solutions.

Practice Exam Questions

Question 1

Which type of memory is MOST appropriate for maintaining conversational context during a single chat session?

A. Long-term memory
B. Semantic memory
C. Short-term memory
D. Episodic memory

Answer

C. Short-term memory

Explanation

Short-term memory maintains active conversational context within a session.

Question 2

Which Azure service is MOST commonly used for semantic retrieval and grounding in AI agents?

A. Azure AI Search
B. Azure Backup
C. Azure DNS
D. Azure Firewall

Answer

A. Azure AI Search

Explanation

Azure AI Search provides vector search and semantic retrieval capabilities.

Question 3

What is the primary purpose of Retrieval-Augmented Generation (RAG)?

A. Replace embeddings
B. Reduce retrieval latency only
C. Ground responses using retrieved information
D. Eliminate vector search

Answer

C. Ground responses using retrieved information

Explanation

RAG retrieves external information to improve groundedness and reduce hallucinations.

Question 4

Which Azure service is MOST appropriate for serverless tool execution within AI agents?

A. Azure Functions
B. Azure CDN
C. Azure Backup
D. Azure Policy

Answer

A. Azure Functions

Explanation

Azure Functions supports serverless execution of business logic and APIs.

Question 5

Which memory type stores knowledge using embeddings and vector representations?

A. Short-term memory
B. Semantic memory
C. Transactional memory
D. Procedural memory

Answer

B. Semantic memory

Explanation

Semantic memory stores information in vectorized forms for retrieval.

Question 6

Which Foundry tool is primarily used for orchestrating AI workflows and agent pipelines?

A. Azure Backup
B. Prompt Flow
C. Azure DNS
D. Azure Storage Explorer

Answer

B. Prompt Flow

Explanation

Prompt Flow supports workflow orchestration and prompt chaining.

Question 7

What is the primary advantage of long-term memory in AI agents?

A. Faster GPU performance
B. Persistent cross-session personalization
C. Lower token usage only
D. Reduced API calls

Answer

B. Persistent cross-session personalization

Explanation

Long-term memory enables persistent storage of preferences and history.

Question 8

Which Azure service is MOST appropriate for low-code workflow automation in enterprise agent systems?

A. Azure Logic Apps
B. Azure DNS
C. Azure Monitor
D. Azure DevTest Labs

Answer

A. Azure Logic Apps

Explanation

Azure Logic Apps provides low-code workflow orchestration and integrations.

Question 9

Which capability allows AI agents to invoke APIs and external systems dynamically?

A. OCR
B. Function calling
C. Metadata filtering
D. Image segmentation

Answer

B. Function calling

Explanation

Function calling enables AI models to interact with external tools and services.

Question 10

Which Azure service is MOST appropriate for persistent scalable storage of AI agent memory?

A. Azure Cosmos DB
B. Azure CDN
C. Azure Firewall
D. Azure ExpressRoute

Answer

A. Azure Cosmos DB

Explanation

Azure Cosmos DB is commonly used for scalable persistent memory storage.

Go to the AI-103 Exam Prep Hub main page

AI, AI-103, Artificial Intelligence (AI), Microsoft Certification

Choose an appropriate method for retrieval and indexing (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Plan and manage an Azure AI solution (25–30%)
   --> Choose the appropriate Foundry services for generative AI and agents
      --> Choose an appropriate method for retrieval and indexing

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

One of the most important concepts in modern AI applications is the ability to retrieve the correct information efficiently and accurately.

The AI-103: Develop AI Apps and Agents on Azure certification exam heavily tests knowledge related to:

  • Retrieval methods
  • Indexing strategies
  • Vector search
  • Semantic search
  • Retrieval-Augmented Generation (RAG)
  • Hybrid search
  • Embeddings
  • Knowledge grounding

Modern AI systems are often only as effective as their retrieval systems.

Even highly advanced Large Language Models (LLMs) can:

  • Hallucinate
  • Provide outdated information
  • Miss relevant context

Retrieval and indexing systems solve these problems by providing grounded, relevant, and searchable information to AI applications.

For the AI-103 exam, you should understand:

  • Different retrieval methods
  • Different indexing approaches
  • When to use vector search
  • When keyword search is appropriate
  • When hybrid search is preferred
  • How embeddings support retrieval
  • How Azure AI Search supports enterprise AI systems
  • How RAG architectures work

What Is Retrieval?

Retrieval is the process of locating and returning relevant information from a data source.

Examples include:

  • Searching documents
  • Finding relevant knowledge articles
  • Retrieving product descriptions
  • Returning similar documents
  • Finding semantically related content

Retrieval is essential for:

  • AI copilots
  • Enterprise chatbots
  • Knowledge assistants
  • Search applications
  • Recommendation systems
  • AI agents

What Is Indexing?

Indexing is the process of organizing data to make retrieval efficient.

An index acts like a searchable map of content.

Without indexing:

  • Searches are slower
  • Retrieval is inefficient
  • AI systems scale poorly

Indexes may include:

  • Keywords
  • Metadata
  • Embeddings
  • Semantic relationships
  • Document structure

Why Retrieval and Indexing Matter in AI

Modern generative AI applications often use Retrieval-Augmented Generation (RAG).

RAG combines:

  • Retrieval systems
  • Search indexes
  • Embeddings
  • LLMs

This allows AI systems to:

  • Access current information
  • Use enterprise knowledge
  • Reduce hallucinations
  • Provide grounded answers
  • Improve accuracy

Azure Services for Retrieval and Indexing

The primary Azure service for retrieval and indexing is:

  • Azure AI Search

Additional supporting services include:

  • Azure OpenAI
  • Embedding models
  • Azure Cosmos DB
  • Azure SQL Database
  • Azure Blob Storage

Azure AI Search

Azure AI Search is Microsoft’s enterprise search platform.

It supports:

  • Full-text search
  • Semantic search
  • Vector search
  • Hybrid search
  • AI enrichment
  • Indexing pipelines

Azure AI Search is a core AI-103 exam topic.

Retrieval Methods

There are several major retrieval methods you must understand for AI-103.

Keyword Search

What Is Keyword Search?

Keyword search retrieves documents based on exact word matches.

Example:

Searching for:

“cloud security”

Returns documents containing those exact terms.

Advantages of Keyword Search

  • Fast
  • Simple
  • Efficient for exact matches
  • Mature technology
  • Works well for structured terminology

Limitations of Keyword Search

Keyword search struggles with:

  • Synonyms
  • Contextual meaning
  • Natural language understanding
  • Conceptual similarity

Example:

A search for:

“car”

May not return documents containing:

“vehicle”

When to Use Keyword Search

Use keyword search when:

  • Exact term matching is important
  • Queries are highly structured
  • Performance and simplicity matter
  • Semantic understanding is unnecessary

Semantic Search

What Is Semantic Search?

Semantic search understands meaning and context rather than relying only on exact words.

It uses AI to interpret:

  • Intent
  • Context
  • Relationships between concepts

Example of Semantic Search

A query for:

“How do I secure cloud infrastructure?”

May retrieve documents about:

  • Azure security
  • Network protection
  • Cloud compliance

Even if the exact words differ.

Advantages of Semantic Search

  • Better contextual understanding
  • Improved relevance
  • More natural interactions
  • Better user experience

Limitations of Semantic Search

  • More computationally expensive
  • May increase latency
  • Requires more advanced indexing

When to Use Semantic Search

Use semantic search when:

  • Natural language queries are common
  • Relevance is important
  • Users may not know exact terminology
  • Context matters

Vector Search

What Is Vector Search?

Vector search retrieves information using embeddings.

Embeddings are numerical vector representations of content.

Documents with similar meaning have vectors that are mathematically close.

How Vector Search Works

  1. Documents are converted into embeddings
  2. Embeddings are stored in a vector index
  3. User queries are converted into embeddings
  4. Similarity algorithms identify related vectors
  5. Relevant documents are returned

Advantages of Vector Search

  • Excellent semantic similarity matching
  • Supports RAG architectures
  • Finds conceptually related content
  • Works well with natural language queries

Limitations of Vector Search

  • Higher storage requirements
  • More computational overhead
  • Requires embedding generation
  • More complex implementation

When to Use Vector Search

Use vector search when:

  • Building RAG systems
  • Implementing AI copilots
  • Performing semantic retrieval
  • Supporting conversational AI
  • Searching unstructured content

Hybrid Search

What Is Hybrid Search?

Hybrid search combines:

  • Keyword search
  • Semantic search
  • Vector search

This approach often produces the best retrieval quality.

Why Hybrid Search Matters

Hybrid search combines the strengths of multiple retrieval approaches.

Benefits include:

  • Exact keyword matching
  • Semantic understanding
  • Contextual similarity
  • Improved ranking quality

When to Use Hybrid Search

Use hybrid search when:

  • High retrieval quality is required
  • Enterprise search is needed
  • AI copilots require strong grounding
  • Search relevance is critical

Hybrid search is commonly used in production RAG systems.

Embeddings

What Are Embeddings?

Embeddings are numerical representations of data.

Embedding models transform:

  • Text
  • Images
  • Documents

Into vectors.

Embeddings capture semantic meaning.

Embedding Models

Azure OpenAI provides embedding models used for:

  • Vector search
  • Similarity matching
  • RAG systems
  • Recommendation systems

Chunking Strategies

What Is Chunking?

Chunking is the process of breaking large documents into smaller sections before indexing.

Chunking improves retrieval quality because:

  • Smaller chunks are easier to match
  • Context becomes more precise
  • Retrieval relevance improves

Common Chunking Methods

Fixed-Size Chunking

Documents are split into equal-sized chunks.

Advantages:

  • Simple
  • Easy to implement

Disadvantages:

  • May split important context

Semantic Chunking

Documents are split based on meaning or structure.

Advantages:

  • Better contextual integrity
  • Improved retrieval quality

Disadvantages:

  • More complex

Overlapping Chunks

Adjacent chunks share some content.

Advantages:

  • Preserves context continuity
  • Improves retrieval accuracy

Disadvantages:

  • Increased storage usage

Choosing a Chunking Strategy

Use Fixed-Size Chunking When:

  • Simplicity is important
  • Documents are uniform
  • Rapid implementation is needed

Use Semantic Chunking When:

  • Context preservation matters
  • Documents contain sections/topics
  • Retrieval quality is critical

Use Overlapping Chunks When:

  • Context continuity is important
  • Long-form content is indexed

Metadata Filtering

Indexes may include metadata such as:

  • Author
  • Date
  • Department
  • Category
  • Security level

Metadata filtering improves:

  • Precision
  • Security
  • Retrieval efficiency

Example Metadata Filtering Scenario

An enterprise chatbot retrieves only documents:

  • From HR
  • Created within the last year
  • Approved for employee access

Metadata filters help enforce these constraints.

Retrieval-Augmented Generation (RAG)

What Is RAG?

Retrieval-Augmented Generation combines retrieval systems with LLMs.

The workflow:

  1. User submits a query
  2. Query becomes an embedding
  3. Vector search retrieves relevant documents
  4. Retrieved content is added to the prompt
  5. LLM generates grounded response

Benefits of RAG

RAG helps:

  • Reduce hallucinations
  • Use current enterprise data
  • Avoid retraining models
  • Improve factual accuracy
  • Support enterprise AI assistants

Choosing Retrieval Methods for RAG

Keyword Search

Best for:

  • Exact terminology
  • Compliance searches
  • Structured queries

Vector Search

Best for:

  • Semantic similarity
  • Natural language queries
  • Conversational AI

Hybrid Search

Best for:

  • Enterprise copilots
  • High-quality retrieval
  • Production RAG systems

Indexing Pipelines

What Is an Indexing Pipeline?

An indexing pipeline automates:

  • Data ingestion
  • Document parsing
  • Chunking
  • Embedding generation
  • Metadata extraction
  • Index updates

AI Enrichment

Azure AI Search supports AI enrichment during indexing.

AI enrichment may include:

  • OCR
  • Entity extraction
  • Key phrase extraction
  • Language detection
  • Image analysis

Incremental Indexing

Incremental indexing updates only changed documents.

Benefits:

  • Faster indexing
  • Lower compute costs
  • Better scalability

Full Reindexing

Full reindexing rebuilds the entire index.

Use when:

  • Schema changes occur
  • Embedding models change
  • Large structural updates are required

Choosing an Indexing Strategy

Use Incremental Indexing When:

  • Data changes frequently
  • Efficiency matters
  • Large datasets exist

Use Full Reindexing When:

  • Major schema updates occur
  • Embedding strategy changes
  • Large-scale restructuring is required

Security and Access Control

Retrieval systems often include:

  • Role-based access control
  • Document-level security
  • Metadata-based filtering

This ensures users retrieve only authorized content.

Common AI-103 Scenarios

Scenario 1: Enterprise Knowledge Assistant

Requirements:

  • Conversational search
  • Semantic retrieval
  • Enterprise grounding

Recommended Approach:

  • Azure AI Search
  • Embeddings
  • Hybrid search
  • RAG

Scenario 2: Compliance Document Search

Requirements:

  • Exact terminology
  • Legal references
  • Precision retrieval

Recommended Approach:

  • Keyword search
  • Metadata filtering

Scenario 3: AI Copilot

Requirements:

  • Natural language queries
  • Contextual retrieval
  • Strong relevance

Recommended Approach:

  • Hybrid search
  • Vector search
  • Embeddings

Scenario 4: Product Recommendation System

Requirements:

  • Similarity matching
  • Semantic relationships

Recommended Approach:

  • Embeddings
  • Vector search

Common AI-103 Exam Tips

Understand Retrieval Tradeoffs

Keyword Search

  • Fast
  • Exact matching
  • Weak semantic understanding

Semantic Search

  • Better contextual understanding
  • More advanced relevance

Vector Search

  • Best for semantic similarity
  • Requires embeddings

Hybrid Search

  • Often best overall retrieval quality

Know the Relationship Between Embeddings and Vector Search

Embeddings enable vector search.

Without embeddings, vector search cannot function.

Understand RAG Architectures

RAG combines:

  • Retrieval
  • Indexing
  • Vector search
  • LLMs

This is one of the MOST important AI-103 topics.

Learn Chunking Concepts

Chunking affects:

  • Retrieval quality
  • Context preservation
  • Index efficiency

Chunking questions commonly appear in scenario-based exam questions.

Summary

Retrieval and indexing are foundational components of modern AI systems.

For the AI-103 exam, you should understand:

  • Keyword search
  • Semantic search
  • Vector search
  • Hybrid search
  • Embeddings
  • Chunking strategies
  • Metadata filtering
  • Indexing pipelines
  • Incremental indexing
  • RAG architectures
  • Azure AI Search capabilities

Choosing the correct retrieval and indexing approach directly affects:

  • AI accuracy
  • Groundedness
  • Scalability
  • Cost
  • Performance
  • User experience

Strong retrieval systems are essential for enterprise AI copilots, chatbots, and AI agents.

Practice Exam Questions

Question 1

Which retrieval method relies primarily on exact word matching?

A. Vector search
B. Semantic search
C. Keyword search
D. Hybrid search

Answer

C. Keyword search

Explanation

Keyword search retrieves content using exact lexical matches.

Question 2

Which retrieval method uses embeddings to identify semantically similar content?

A. Keyword search
B. Vector search
C. Lexical search
D. Metadata search

Answer

B. Vector search

Explanation

Vector search uses embeddings to perform similarity matching.

Question 3

What is the primary benefit of Retrieval-Augmented Generation (RAG)?

A. Eliminates embeddings
B. Improves groundedness using retrieved information
C. Removes the need for indexing
D. Replaces semantic search

Answer

B. Improves groundedness using retrieved information

Explanation

RAG improves factual accuracy by grounding responses with retrieved data.

Question 4

Which Azure service is MOST commonly used for enterprise vector search?

A. Azure AI Search
B. Azure DNS
C. Azure Backup
D. Azure Load Balancer

Answer

A. Azure AI Search

Explanation

Azure AI Search provides vector indexing and retrieval capabilities.

Question 5

What is the purpose of chunking during indexing?

A. Encrypt documents
B. Break documents into smaller searchable sections
C. Compress embeddings
D. Eliminate metadata

Answer

B. Break documents into smaller searchable sections

Explanation

Chunking improves retrieval quality and contextual matching.

Question 6

Which search method combines vector search, semantic ranking, and keyword matching?

A. Binary search
B. Metadata search
C. Hybrid search
D. OCR search

Answer

C. Hybrid search

Explanation

Hybrid search combines multiple retrieval methods.

Question 7

What is the primary purpose of embeddings?

A. Encrypt data
B. Create semantic vector representations
C. Compress images
D. Improve OCR quality

Answer

B. Create semantic vector representations

Explanation

Embeddings convert content into vectors representing semantic meaning.

Question 8

Which chunking strategy helps preserve context continuity between adjacent chunks?

A. Fixed chunking
B. Metadata chunking
C. Overlapping chunks
D. Compression chunking

Answer

C. Overlapping chunks

Explanation

Overlapping chunks preserve continuity across document sections.

Question 9

When is incremental indexing MOST appropriate?

A. When rebuilding the entire schema
B. When documents change frequently
C. When changing embedding models
D. When deleting the index

Answer

B. When documents change frequently

Explanation

Incremental indexing updates only modified documents.

Question 10

Which retrieval approach is MOST appropriate for enterprise AI copilots requiring high-quality relevance?

A. Keyword search only
B. Hybrid search
C. Metadata filtering only
D. OCR search

Answer

B. Hybrid search

Explanation

Hybrid search combines multiple retrieval methods for improved relevance.

Go to the AI-103 Exam Prep Hub main page

AI, AI-103, Artificial Intelligence (AI), Microsoft Certification

Choose an appropriate model for each task, including large language models (LLMs), small language models, multimodal models, and Foundry Tools (AI-103 Exam Prep)

 
This post is a part of the AI-103: Develop AI Apps and Agents on Azure Exam Prep Hub. 
This topic falls under these sections:
Plan and manage an Azure AI solution (25–30%)
   --> Choose the appropriate Foundry services for generative AI and agents
      --> Choose an appropriate model for each task, including large language models (LLMs), small language models, multimodal models, and Foundry Tools

Note that there are 10 practice questions (with answers and explanations) at the end of each section to help you solidify your knowledge of the material. Also, there are 2 practice tests with 60 questions each available from the hub's main page below the exam topics section.

Introduction

One of the most important skills for the AI-103: Develop AI Apps and Agents on Azure certification exam is understanding how to choose the correct AI model and supporting Azure AI Foundry tools for a given business or technical scenario.

Modern AI development is no longer about simply selecting “an AI model.” Instead, developers must evaluate:

  • The type of task being performed
  • Cost constraints
  • Latency requirements
  • Accuracy expectations
  • Reasoning complexity
  • Context window needs
  • Multimodal capabilities
  • Deployment environment
  • Security and governance requirements
  • Agent orchestration requirements

Azure AI Foundry provides access to multiple categories of models and tools that help developers build generative AI applications and AI agents efficiently.

For the AI-103 exam, you should understand:

  • When to use Large Language Models (LLMs)
  • When Small Language Models (SLMs) are preferable
  • When multimodal models are required
  • How Azure AI Foundry tools support model selection and orchestration
  • Tradeoffs between performance, cost, speed, and capability
  • Common real-world scenarios for each model category

Azure AI Foundry Overview

Azure AI Foundry is Microsoft’s unified platform for building, evaluating, deploying, and managing AI applications and agents.

Azure AI Foundry provides:

  • Access to foundation models
  • Agent development capabilities
  • Prompt engineering tools
  • Evaluation tools
  • Safety and content filtering
  • Retrieval-augmented generation (RAG) support
  • Fine-tuning capabilities
  • Monitoring and observability
  • Integration with Azure AI services

Azure AI Foundry enables developers to:

  • Compare multiple models
  • Test prompts
  • Evaluate outputs
  • Build AI agents
  • Connect enterprise data
  • Deploy scalable AI applications

For the AI-103 exam, understanding the relationship between model capabilities and Azure AI Foundry tools is extremely important.

Understanding Model Categories

The exam focuses heavily on selecting the correct model type for specific tasks.

The major categories include:

  1. Large Language Models (LLMs)
  2. Small Language Models (SLMs)
  3. Multimodal Models
  4. Embedding Models
  5. Specialized Models

Each category serves different purposes.

Large Language Models (LLMs)

What Are Large Language Models?

Large Language Models are advanced AI models trained on massive datasets containing text, code, and other information.

LLMs are designed for:

  • Natural language understanding
  • Natural language generation
  • Complex reasoning
  • Summarization
  • Coding assistance
  • Question answering
  • Conversational AI
  • Agent workflows
  • Content creation

Examples include:

  • GPT-4 family models
  • GPT-4o models
  • GPT-4 Turbo
  • Phi large models
  • Other frontier foundation models available in Azure AI Foundry

Characteristics of LLMs

Strengths

LLMs are excellent at:

Complex Reasoning

Examples:

  • Multi-step problem solving
  • Data interpretation
  • Logical analysis
  • Decision support

Advanced Content Generation

Examples:

  • Marketing content
  • Technical documentation
  • Email drafting
  • Knowledge-base generation

Conversational Experiences

Examples:

  • AI chatbots
  • AI copilots
  • Virtual assistants
  • Interactive tutoring systems

Agentic Workflows

LLMs are commonly used as the “reasoning engine” behind AI agents.

They can:

  • Plan tasks
  • Determine next actions
  • Call tools
  • Use memory
  • Chain workflows
  • Interact with APIs

Limitations of LLMs

Although powerful, LLMs have tradeoffs.

Higher Cost

LLMs generally:

  • Require more compute
  • Cost more per token
  • Increase infrastructure expenses

Increased Latency

Larger models may:

  • Respond more slowly
  • Increase application response times
  • Affect real-time user experiences

Resource Requirements

LLMs require:

  • More GPU resources
  • More memory
  • Larger deployments

Overkill for Simple Tasks

Using GPT-4-level reasoning for basic classification or short summarization tasks may be unnecessary and expensive.

When to Use LLMs

Choose an LLM when tasks require:

  • Advanced reasoning
  • Long-context understanding
  • High-quality content generation
  • Complex conversational behavior
  • Tool calling and agent orchestration
  • Coding assistance
  • Sophisticated summarization
  • Enterprise copilots

Example LLM Scenarios

Scenario 1: Enterprise AI Copilot

A company wants an AI assistant that:

  • Reads internal documentation
  • Answers employee questions
  • Generates summaries
  • Explains policies
  • Uses tools and APIs

Best choice:

  • Large Language Model with RAG integration

Reason:

  • Requires reasoning and conversational understanding.

Scenario 2: AI Coding Assistant

A development team needs:

  • Code generation
  • Debugging suggestions
  • Refactoring support
  • Documentation generation

Best choice:

  • Advanced LLM

Reason:

  • Coding tasks require complex contextual reasoning.

Small Language Models (SLMs)

What Are Small Language Models?

Small Language Models are more lightweight AI models optimized for:

  • Faster responses
  • Lower costs
  • Lower resource consumption
  • Edge deployments
  • Narrower tasks

Examples include:

  • Smaller Phi models
  • Compact transformer-based models
  • Task-specific lightweight models

Characteristics of SLMs

Strengths

Lower Cost

SLMs:

  • Consume fewer resources
  • Cost less to run
  • Reduce token usage costs

Faster Inference

SLMs typically:

  • Respond more quickly
  • Improve responsiveness
  • Support near real-time interactions

Edge and Mobile Suitability

SLMs may run:

  • On edge devices
  • On mobile hardware
  • In constrained environments

Efficient for Narrow Tasks

SLMs work well for:

  • Classification
  • Basic summarization
  • Intent detection
  • Simple chat interactions
  • Lightweight automation

Limitations of SLMs

Reduced Reasoning Ability

Compared to LLMs, SLMs may struggle with:

  • Complex logic
  • Long context handling
  • Multi-step reasoning
  • Sophisticated conversations

Lower Output Quality

Outputs may:

  • Be less nuanced
  • Contain reduced detail
  • Provide weaker contextual understanding

When to Use SLMs

Choose an SLM when:

  • Speed is critical
  • Cost optimization matters
  • Tasks are relatively simple
  • Edge deployment is needed
  • High throughput is required
  • Lightweight AI experiences are sufficient

Example SLM Scenarios

Scenario 1: Customer Intent Classification

An application classifies support tickets into categories such as:

  • Billing
  • Technical support
  • Returns
  • Sales

Best choice:

  • Small Language Model

Reason:

  • Classification is relatively simple and does not require advanced reasoning.

Scenario 2: Edge Device Assistant

A manufacturing company deploys an AI assistant on factory equipment with limited compute.

Best choice:

  • Small Language Model

Reason:

  • Edge environments benefit from lightweight models.

Multimodal Models

What Are Multimodal Models?

Multimodal models can process multiple data types simultaneously.

Examples include:

  • Text
  • Images
  • Audio
  • Video
  • Documents

These models combine information across modalities to produce richer outputs.

Capabilities of Multimodal Models

Multimodal models can:

  • Analyze images and answer questions about them
  • Generate captions from images
  • Extract information from documents
  • Process speech and text together
  • Understand charts and diagrams
  • Support visual reasoning

Common Multimodal Tasks

Image Understanding

Examples:

  • Object detection
  • Scene analysis
  • Image captioning
  • Visual question answering

Document Intelligence

Examples:

  • Invoice extraction
  • Receipt processing
  • Form analysis
  • OCR workflows

Audio + Text Experiences

Examples:

  • Voice assistants
  • Meeting summarization
  • Speech transcription
  • Audio analysis

When to Use Multimodal Models

Choose multimodal models when applications involve:

  • Images and text together
  • Document processing
  • Speech interactions
  • Visual understanding
  • Cross-modal reasoning

Example Multimodal Scenarios

Scenario 1: Invoice Processing

A company needs to:

  • Read invoices
  • Extract totals
  • Identify vendors
  • Validate line items

Best choice:

  • Multimodal document processing model

Reason:

  • The solution must interpret both layout and text.

Scenario 2: Retail Image Assistant

Users upload photos of products and ask questions about them.

Best choice:

  • Multimodal model

Reason:

  • Requires simultaneous image and text understanding.

Embedding Models

What Are Embedding Models?

Embedding models convert text or other content into vector representations.

These vectors capture semantic meaning.

Embedding models are essential for:

  • Semantic search
  • Retrieval-Augmented Generation (RAG)
  • Similarity matching
  • Recommendation systems
  • Knowledge retrieval

Retrieval-Augmented Generation (RAG)

RAG combines:

  • Embedding models
  • Vector databases
  • LLMs

Workflow:

  1. Convert documents into embeddings
  2. Store embeddings in a vector index
  3. Convert user query into embeddings
  4. Retrieve relevant content
  5. Send retrieved data to the LLM

RAG improves:

  • Accuracy
  • Freshness of information
  • Enterprise grounding
  • Hallucination reduction

Specialized Models

Some tasks are better handled by specialized AI models instead of general-purpose LLMs.

Examples:

  • Translation models
  • Speech models
  • OCR models
  • Vision models
  • Classification models

Why Specialized Models Matter

Specialized models may provide:

  • Better accuracy
  • Lower cost
  • Faster performance
  • Simpler deployment

Example:

Using a dedicated OCR service is often more efficient than asking an LLM to read text from images.

Model Selection Factors

The AI-103 exam heavily tests your ability to select the correct model based on requirements.

Factor 1: Task Complexity

Use LLMs For:

  • Advanced reasoning
  • Multi-step workflows
  • Complex conversations

Use SLMs For:

  • Simple classification
  • Lightweight interactions
  • Fast automation

Factor 2: Cost

LLMs

  • Higher operational cost
  • More expensive inference

SLMs

  • Lower operational cost
  • Better for high-volume workloads

Factor 3: Latency

Low-Latency Requirements

Prefer:

  • SLMs
  • Lightweight models

Complex Processing

Prefer:

  • LLMs

Even if response time increases.

Factor 4: Context Window

Some tasks require processing:

  • Long documents
  • Large conversations
  • Extensive histories

Choose models with larger context windows for:

  • Legal analysis
  • Knowledge assistants
  • Long-form summarization

Factor 5: Multimodal Requirements

If the application involves:

  • Images
  • Audio
  • Video
  • Documents

Choose multimodal-capable models.

Factor 6: Deployment Environment

Cloud-Hosted Applications

May use:

  • Large frontier models
  • GPU-intensive deployments

Edge or Mobile Deployments

Prefer:

  • Small models
  • Quantized models
  • Lightweight inference

Azure AI Foundry Tools

Azure AI Foundry includes numerous tools that support model selection and AI application development.

Model Catalog

The Model Catalog allows developers to:

  • Browse available models
  • Compare capabilities
  • Review benchmarks
  • Deploy models
  • Evaluate pricing

The catalog includes:

  • Microsoft-hosted models
  • Open-source models
  • Partner models
  • Frontier models

Prompt Flow

Prompt Flow helps developers:

  • Build AI workflows
  • Chain prompts together
  • Integrate tools
  • Evaluate prompts
  • Test model behavior

Prompt Flow is useful for:

  • Agent orchestration
  • RAG pipelines
  • Multi-step AI workflows

AI Agent Development Tools

Azure AI Foundry supports AI agents that can:

  • Use tools
  • Access data
  • Maintain memory
  • Perform actions
  • Execute workflows

Agent frameworks may include:

  • Tool calling
  • Function calling
  • Retrieval integration
  • Multi-agent orchestration

Evaluation Tools

Evaluation tools help developers assess:

  • Accuracy
  • Groundedness
  • Safety
  • Relevance
  • Latency
  • Cost

Evaluation is critical because model quality varies by task.

Content Safety Tools

Azure AI Foundry includes safety features such as:

  • Content filtering
  • Harm detection
  • Prompt injection detection
  • Responsible AI controls

These tools help ensure safe AI deployments.

Fine-Tuning Tools

Fine-tuning allows developers to customize models using:

  • Domain-specific data
  • Proprietary terminology
  • Specialized workflows

Fine-tuning may improve:

  • Accuracy
  • Consistency
  • Industry-specific responses

However, fine-tuning also:

  • Increases cost
  • Requires data preparation
  • Adds operational complexity

Choosing Between Prompt Engineering, RAG, and Fine-Tuning

This is a very important AI-103 exam topic.

Prompt Engineering

Use when:

  • You need quick customization
  • Tasks are general-purpose
  • No private data integration is needed

Advantages:

  • Fast
  • Cheap
  • Easy to maintain

RAG

Use when:

  • You need current or proprietary data
  • You want grounding in enterprise content
  • You need dynamic knowledge retrieval

Advantages:

  • Reduces hallucinations
  • Keeps knowledge current
  • Avoids retraining

Fine-Tuning

Use when:

  • Consistent specialized outputs are required
  • Domain language is highly unique
  • Behavioral customization is necessary

Advantages:

  • Tailored responses
  • Better domain alignment

Real-World Model Selection Examples

Example 1: FAQ Chatbot

Requirements:

  • Low cost
  • Fast responses
  • Basic conversational support

Best Choice:

  • Small Language Model + RAG

Example 2: Legal Document Assistant

Requirements:

  • Long-context understanding
  • Detailed summarization
  • Advanced reasoning

Best Choice:

  • Large Language Model with large context window

Example 3: Mobile AI App

Requirements:

  • Offline capability
  • Fast performance
  • Low resource usage

Best Choice:

  • Small Language Model

Example 4: Image-Based Customer Support

Requirements:

  • Analyze uploaded photos
  • Understand text and images
  • Generate responses

Best Choice:

  • Multimodal model

Key AI-103 Exam Tips

Understand Tradeoffs

You should know:

  • Bigger models are not always better
  • Simpler tasks may not require advanced LLMs
  • Cost and latency matter
  • Specialized models may outperform general models

Know Common Pairings

LLM + RAG

Used for:

  • Enterprise chatbots
  • Knowledge assistants
  • AI copilots

Embeddings + Vector Search

Used for:

  • Semantic search
  • Knowledge retrieval
  • Similarity matching

Multimodal Models

Used for:

  • Vision AI
  • Document processing
  • Audio interactions

Learn the Azure AI Foundry Ecosystem

Know the purpose of:

  • Model Catalog
  • Prompt Flow
  • Evaluation tools
  • Agent tools
  • Safety systems
  • Fine-tuning workflows

Summary

Selecting the correct AI model is one of the most important responsibilities for an Azure AI developer.

For the AI-103 exam, you should understand:

  • The differences between LLMs and SLMs
  • When multimodal models are required
  • How embedding models support RAG
  • When specialized models outperform general-purpose models
  • The tradeoffs between cost, speed, and reasoning capability
  • How Azure AI Foundry tools support AI development and orchestration

In real-world AI systems, choosing the correct model can dramatically improve:

  • Performance
  • User experience
  • Scalability
  • Operational cost
  • Reliability
  • Maintainability

A strong understanding of model selection is essential for designing effective Azure AI applications and AI agents.

Practice Exam Questions

Question 1

A company is building an enterprise AI assistant that must answer complex employee questions using internal documentation and perform multi-step reasoning. Which model type is MOST appropriate?

A. Small Language Model (SLM)
B. Embedding model only
C. Large Language Model (LLM)
D. OCR model

Answer

C. Large Language Model (LLM)

Explanation

Complex reasoning and conversational understanding are best handled by LLMs.

Question 2

Which model type is generally BEST for low-cost, low-latency classification tasks?

A. Large multimodal model
B. Small Language Model (SLM)
C. GPT-4-class reasoning model
D. Vision foundation model

Answer

B. Small Language Model (SLM)

Explanation

SLMs are optimized for lightweight and cost-efficient tasks.

Question 3

A solution must process uploaded invoices and extract totals, vendor names, and line items. Which model type is MOST appropriate?

A. Embedding model
B. Small Language Model
C. Multimodal model
D. Translation model

Answer

C. Multimodal model

Explanation

Invoice extraction requires understanding both layout and text.

Question 4

What is the primary purpose of embedding models?

A. Image generation
B. Semantic vector representation
C. Audio transcription
D. Tool orchestration

Answer

B. Semantic vector representation

Explanation

Embedding models convert content into vectors for semantic search and retrieval.

Question 5

Which Azure AI Foundry tool helps developers chain prompts, integrate tools, and build AI workflows?

A. Azure Monitor
B. Prompt Flow
C. Azure Policy
D. Azure Functions

Answer

B. Prompt Flow

Explanation

Prompt Flow is designed for workflow orchestration and prompt pipelines.

Question 6

A mobile AI application must operate with minimal compute resources and very fast response times. Which model type is MOST appropriate?

A. Large Language Model
B. Small Language Model
C. Large multimodal model
D. High-context reasoning model

Answer

B. Small Language Model

Explanation

SLMs are optimized for lightweight and edge deployments.

Question 7

Which approach is BEST when an AI chatbot must use current enterprise data without retraining the model?

A. Fine-tuning only
B. Prompt engineering only
C. Retrieval-Augmented Generation (RAG)
D. Quantization

Answer

C. Retrieval-Augmented Generation (RAG)

Explanation

RAG retrieves current information dynamically without retraining.

Question 8

Which factor MOST strongly indicates that a multimodal model is required?

A. Need for vector embeddings
B. Need for faster response times
C. Need to process images and text together
D. Need for lower cost

Answer

C. Need to process images and text together

Explanation

Multimodal models handle multiple input modalities simultaneously.

Question 9

What is a major tradeoff of using larger language models?

A. Reduced reasoning capability
B. Lower context windows
C. Increased operational cost
D. Inability to support agents

Answer

C. Increased operational cost

Explanation

Larger models typically require more compute resources and cost more.

Question 10

Which Azure AI Foundry capability helps evaluate model quality, safety, and groundedness?

A. Azure Load Balancer
B. Evaluation tools
C. Azure Backup
D. Traffic Manager

Answer

B. Evaluation tools

Explanation

Evaluation tools assess output quality, safety, and performance metrics.

Go to the AI-103 Exam Prep Hub main page