Module II — Provisioning Azure Resources

Note

Duration: ~15 minutes In this module, you will set up all the Azure resources needed to power the Compliance Compass agent: AI Foundry (with models), Blob Storage (for documents), and Azure AI Search (for retrieval).

Objectives

Create a Resource Group to organize all workshop resources
Deploy an AI Foundry hub and project with two models (GPT-4o for reasoning, text-embedding for search)
Create a Storage Account with a Blob container and upload the compliance KB documents
Set up Azure AI Search (Free tier) and create a knowledge index from the blob data
Verify the search index with a test query

Architecture of Azure Resources

graph TD
    RG[" Resource Group\ncompliance-agent-rg"]

    subgraph AIF[" AI Foundry Hub: Compliance-Sentinel"]
        GPT[" gpt-4o\nReasoning Model"]
        EMB[" compliance-embedding\ntext-embedding-ada-002"]
    end

    subgraph STO[" Storage Account: compliancekb"]
        BC[" Blob Container: kb-documents"]
        DOCS[" 12 Compliance\nMarkdown Files"]
        BC --> DOCS
    end

    subgraph AIS[" Azure AI Search: compliancesearch (Free)"]
        IDX[" Index: knowledgesource-index\n81 document chunks"]
    end

    RG --> AIF
    RG --> STO
    RG --> AIS

Step 1: Create a Resource Group

A Resource Group is a logical container for all the Azure resources you will create.

Option A: Azure Portal

Open the Azure Portal.
In the search bar at the top, type Resource groups and select it.
Click + Create.
Fill in the details:
- Subscription: Select your subscription (e.g., Visual Studio Enterprise Subscription – MPN)
- Resource group name: compliance-agent-rg
- Region: Select a region that supports AI Foundry (e.g., East US, East US 2, or Sweden Central)
Click Review + create → Create.

Option B: Azure CLI

# Set variables for the workshop (reused throughout this module)
RESOURCE_GROUP="compliance-agent-rg"
LOCATION="eastus2"

# Create the resource group
az group create --name $RESOURCE_GROUP --location $LOCATION

Tip

Use a consistent region for all resources (e.g., East US 2). Some AI models are only available in specific regions. On Windows CMD, replace $VARIABLE with %VARIABLE%; on PowerShell, use $Variable directly.

Step 2: Create an AI Foundry Hub and Project

AI Foundry is where you deploy the language model (GPT-4o) and embedding model that power the agent.

2.1 Create the AI Foundry Hub

Option A: Azure Portal

In the Azure Portal, search for Azure AI Foundry and select it.
Click + Create → Hub.
Fill in the details:
- Subscription: Your subscription
- Resource group: compliance-agent-rg
- Region: Same region as your Resource Group
- Name: Compliance-Sentinel
Click Review + create → Create.
Wait for deployment to complete, then click Go to resource.

Option B: Azure CLI

# Create the AI Foundry hub
az ml workspace create \
  --kind hub \
  --resource-group $RESOURCE_GROUP \
  --name "Compliance-Sentinel" \
  --location $LOCATION

2.2 Create a Project

Option A: Azure Portal

Inside the AI Foundry hub, click + New project.
Enter the project name: Compliance-Sentinel (or keep the default).
Click Create.

Option B: Azure CLI

# Create a project under the hub
az ml workspace create \
  --kind project \
  --resource-group $RESOURCE_GROUP \
  --name "Compliance-Sentinel" \
  --hub-id /subscriptions/<subscription-id>/resourceGroups/$RESOURCE_GROUP/providers/Microsoft.MachineLearningServices/workspaces/Compliance-Sentinel

Note

Replace <subscription-id> with your actual Azure subscription ID. You can retrieve it with:

az account show --query id -o tsv

2.3 Deploy the GPT-4o Model

This model will be used for reasoning — it analyzes retrieved compliance documents and generates structured reports.

Option A: Azure Portal

In the AI Foundry portal, navigate to Models in the left sidebar.
Click + Deploy model → Deploy base model.
Search for gpt-4o and select it.
Configure:
- Deployment name: gpt-4o
- Model version: Latest available
- Deployment type: Standard
Click Deploy.

Option B: Azure CLI

# Deploy GPT-4o model
az ml online-deployment create \
  --resource-group $RESOURCE_GROUP \
  --workspace-name "Compliance-Sentinel" \
  --endpoint-name "gpt-4o" \
  --name "gpt-4o" \
  --model "azureml://registries/azureml-meta/models/gpt-4o" \
  --instance-type "Standard"

Tip

Alternatively, you can deploy models through the AI Foundry portal under your project's Deployments section. The portal provides a visual interface for selecting model versions and configuring deployment parameters.

2.4 Deploy the Embedding Model

This model is used for vectorizing the compliance documents so Azure AI Search can perform semantic retrieval.

Option A: Azure Portal

Back in the Models section, click + Deploy model → Deploy base model.
Search for text-embedding-ada-002 (or text-embedding-3-small).
Configure:
- Deployment name: compliance-embedding
- Model version: Latest available
- Deployment type: Standard
Click Deploy.

Option B: Azure CLI

# Deploy the embedding model
az ml online-deployment create \
  --resource-group $RESOURCE_GROUP \
  --workspace-name "Compliance-Sentinel" \
  --endpoint-name "compliance-embedding" \
  --name "compliance-embedding" \
  --model "azureml://registries/azureml-meta/models/text-embedding-ada-002" \
  --instance-type "Standard"

After both models are deployed, verify they are listed under Models in the AI Foundry portal. You should see both gpt-4o (reasoning) and compliance-embedding (text-embedding-ada-002).

Note

These two models will be used by the agent for document vectorization and reasoning respectively.

Step 3: Create a Storage Account and Upload Documents

3.1 Create the Storage Account

Option A: Azure Portal

In the Azure Portal, search for Storage accounts and select it.
Click + Create.
Fill in the details:
- Subscription: Your subscription
- Resource group: compliance-agent-rg
- Storage account name: compliancekb (must be globally unique — append random digits if needed, e.g., compliancekb1234)
- Region: Same as other resources
- Performance: Standard
- Redundancy: LRS (Locally Redundant Storage)
Click Review + create → Create.

Option B: Azure CLI

STORAGE_ACCOUNT="compliancekb"  # Must be globally unique; append digits if needed

az storage account create \
  --name $STORAGE_ACCOUNT \
  --resource-group $RESOURCE_GROUP \
  --location $LOCATION \
  --sku Standard_LRS \
  --kind StorageV2

3.2 Create a Blob Container

Option A: Azure Portal

Open the newly created storage account.
In the left sidebar, select Containers (under Data storage).
Click + Container.
Enter the container name: kb-documents.
Set Public access level to Private (no anonymous access).
Click Create.

Option B: Azure CLI

az storage container create \
  --name kb-documents \
  --account-name $STORAGE_ACCOUNT \
  --auth-mode login

3.3 Upload the Knowledge Base Documents

Option A: Azure Portal

Open the kb-documents container.
Click Upload.
Click Browse for files and navigate to the kb_markdown/ folder in your workshop repository.
Select all 12 Markdown files:
- DPDP_Act_India_2023.md
- EU_China_Data_Flows_2024.md
- Export_Control_US_EU_India_2025.md
- GDPR_Article_44_Transfers.md
- GDPR_Schrems_II_Post_2020.md
- Incident_Response_Best_Practices.md
- Mitigation_Templates_DPA.md
- RBI_Cross_Border_Transactions_Circular_2023.md
- RBI_Data_Localization_2018_Guidelines.md
- RBI_Vendor_Onboarding_Risk_2024.md
- Risk_Scoring_Framework.md
- SEBI_Insider_Trading_Policy.md
Click Upload.

Option B: Azure CLI

# Upload all 12 KB documents from the local kb_markdown/ folder
az storage blob upload-batch \
  --account-name $STORAGE_ACCOUNT \
  --destination kb-documents \
  --source ./kb_markdown/ \
  --auth-mode login

Verify the upload:

az storage blob list \
  --account-name $STORAGE_ACCOUNT \
  --container-name kb-documents \
  --auth-mode login \
  --query "[].name" \
  --output table

Expected output — all 12 documents listed:

Result
-----------------------------------------
DPDP_Act_India_2023.md
EU_China_Data_Flows_2024.md
Export_Control_US_EU_India_2025.md
GDPR_Article_44_Transfers.md
GDPR_Schrems_II_Post_2020.md
Incident_Response_Best_Practices.md
Mitigation_Templates_DPA.md
RBI_Cross_Border_Transactions_Circular_2023.md
RBI_Data_Localization_2018_Guidelines.md
RBI_Vendor_Onboarding_Risk_2024.md
Risk_Scoring_Framework.md
SEBI_Insider_Trading_Policy.md

Step 4: Create Azure AI Search and Build the Knowledge Index

Azure AI Search indexes the uploaded documents so the agent can retrieve relevant compliance policies when answering queries.

4.1 Create the Azure AI Search Resource

Option A: Azure Portal

In the Azure Portal, search for Azure AI Search and select it.
Click + Create.
Fill in the details:
- Subscription: Your subscription
- Resource group: compliance-agent-rg
- Service name: compliancesearch (must be globally unique)
- Location: Same as other resources
- Pricing tier: Free (sufficient for this workshop)
Click Review + create → Create.
Wait for deployment to complete.

Option B: Azure CLI

SEARCH_SERVICE="compliancesearch"  # Must be globally unique

az search service create \
  --name $SEARCH_SERVICE \
  --resource-group $RESOURCE_GROUP \
  --location $LOCATION \
  --sku free

Verify the search service is running:

az search service show \
  --name $SEARCH_SERVICE \
  --resource-group $RESOURCE_GROUP \
  --query "{Name:name, Status:status, Sku:sku.name}" \
  --output table

4.2 Create a Knowledge Base with Azure Blob as Source

Now link the AI Search to your Blob Storage to automatically index the compliance documents.

Navigate to your AI Foundry project in the AI Foundry portal.
In the left sidebar, select Knowledge → + New knowledge base.
Fill in the Basics:
- Name: compliance-kb (or keep the auto-generated name)
- Description: Compliance regulatory knowledge base for vendor risk, data protection, and cross-border operations
Under Knowledge sources, click + Create new.
In the Create knowledge source popup, you will see multiple options (Azure blob, Search index, SharePoint, etc.). Select Azure blob (Indexed).
Configure the Azure Blob source:
- Name: Keep the auto-generated name (or enter compliance-kb-source)
- Subscription: Your subscription
- Storage account: compliancekb
- Blob container: kb-documents
- Content extraction Mode: Minimal
Scroll down and check Enable text vectorization if available, and select your embedding model deployment (compliance-embedding).
Click Create.

Note

The indexing process may take 2–5 minutes to complete. The system will:

Connect to your Blob Storage
Extract text from each Markdown file
Chunk the text into smaller segments
Generate vector embeddings using the embedding model
Create a searchable index

4.3 Verify the Search Index

Navigate back to your Azure AI Search resource in the Azure Portal.
Click on Indexes in the left sidebar.
You should see a new index (e.g., knowledgesource-1773330398864-index).
Click on the index to open it. Verify the statistics:
- Documents: ~81 (chunks from 12 documents)
- Total storage: ~3.5 MB
- Vector index quota usage: ~743 KB

4.4 Test the Retrieval

In the index page, click the Search explorer tab.

In the search box, enter a test query:

RBI data localization payment data India

Click Search.
Verify that the results return relevant chunks from RBI_Data_Localization_2018_Guidelines.md.

Tip

Try additional test queries:

GDPR cross-border transfer safeguards SCCs
vendor onboarding risk assessment
insider trading compliance SEBI

Each should return relevant document chunks from the Knowledge Base.

Step 5: Verify All Resources

Option A: Azure Portal

Open your Resource Group (compliance-agent-rg) in the Azure Portal and verify that all resources are present:

Resource	Type	Purpose
`Compliance-Sentinel`	AI Foundry Hub	Hosts models and agent
`gpt-4o`	Model Deployment	Reasoning model
`compliance-embedding`	Model Deployment	Embedding model for vectorization
`compliancekb`	Storage Account	Stores compliance documents
`compliancesearch`	Azure AI Search	Indexes and retrieves documents

Option B: Azure CLI

# List all resources in the resource group
az resource list \
  --resource-group compliance-agent-rg \
  --query "[].{Name:name, Type:type, Location:location}" \
  --output table

Verify blob documents were uploaded:

az storage blob list \
  --account-name $STORAGE_ACCOUNT \
  --container-name kb-documents \
  --auth-mode login \
  --query "length(@)" \
  --output tsv

Expected output: 12 (the total number of compliance documents).

Prerequisites Checklist

Before moving to Module III, confirm:

Resource Group compliance-agent-rg created
AI Foundry hub Compliance-Sentinel created with a project
GPT-4o model deployed
Embedding model deployed
Storage Account created with kb-documents container
All 12 Markdown documents uploaded to the container
Azure AI Search resource created (Free tier)
Knowledge index created and populated (~81 documents)
Test query in Search Explorer returns relevant results

Key Takeaways

Azure AI Foundry serves as the brain of the system — it provides both the reasoning model (GPT-4o) and the embedding model for vectorization.
Azure Blob Storage acts as the document repository — a simple, scalable way to store compliance policies.
Azure AI Search bridges the gap between raw documents and intelligent retrieval — it chunks, vectorizes, and indexes the content for fast semantic search.
The Free tier of Azure AI Search is sufficient for prototyping RAG agents with small document sets.

Click Next to proceed to Module III: Designing the Agent in Foundry Toolkit for VS Code.

FilesExpand file tree

02_Azure_Resources.md

Latest commit

History

02_Azure_Resources.md

File metadata and controls

Module II — Provisioning Azure Resources

Objectives

Architecture of Azure Resources

Step 1: Create a Resource Group

Option A: Azure Portal

Option B: Azure CLI

Step 2: Create an AI Foundry Hub and Project

2.1 Create the AI Foundry Hub

Option A: Azure Portal

Option B: Azure CLI

2.2 Create a Project

Option A: Azure Portal

Option B: Azure CLI

2.3 Deploy the GPT-4o Model

Option A: Azure Portal

Option B: Azure CLI

2.4 Deploy the Embedding Model

Option A: Azure Portal

Option B: Azure CLI

Step 3: Create a Storage Account and Upload Documents

3.1 Create the Storage Account

Option A: Azure Portal

Option B: Azure CLI

3.2 Create a Blob Container

Option A: Azure Portal

Option B: Azure CLI

3.3 Upload the Knowledge Base Documents

Option A: Azure Portal

Option B: Azure CLI

Step 4: Create Azure AI Search and Build the Knowledge Index

4.1 Create the Azure AI Search Resource

Option A: Azure Portal

Option B: Azure CLI

4.2 Create a Knowledge Base with Azure Blob as Source

4.3 Verify the Search Index

4.4 Test the Retrieval

Step 5: Verify All Resources

Option A: Azure Portal

Option B: Azure CLI

Prerequisites Checklist

Key Takeaways