Skip to content

Quick Deploy Guide

Prerequisites

  • Azure subscription with Contributor access & Role Based Access Control access
  • VS Code, Azure Developer CLI (aka.ms/azd), Python 3.10+, Git
  • For Fabric deployment: Microsoft Fabric workspace (F4+ capacity) with admin permissions

Choose Your Development Environment

Local Visual Studio Code: Open Visual Studio Code. From the File menu, select Open Folder and choose the folder where you want to deploy the workshop.

Or choose one of the options below:

Open in GitHub Codespaces Open in VS Code Web

Note: Please use this prompt if you would like to use GitHub Copilot to run the workshop: Can you please follow the step by step in https://microsoft.github.io/agentic-applications-for-unified-data-foundation-solution-accelerator/deployment-guide/ and follow Option A for me. My Fabric Workspace id = <YOUR_FABRIC_WORKSPACE_ID>. Important instructions: Do NOT make any code changes to the repository files. Only follow the deployment guide instructions exactly as documented. Run the commands step by step and wait for each to complete before proceeding. If I encounter any errors or issues, help me troubleshoot and resolve them before continuing. Explain what each step does before running it. If a step fails, suggest solutions based on the error message.

Option A: Full Deployment (Fabric + Foundry)

1. Configure Fabric workspace

Create a new Fabric workspace.

2. Clone the repository

git clone https://github.com/microsoft/agentic-applications-for-unified-data-foundation-solution-accelerator.git

cd agentic-applications-for-unified-data-foundation-solution-accelerator

cp .env.example .env # or: copy .env.example .env

2.1 Get Fabric workspace Id

Open .env and set FABRIC_WORKSPACE_ID from Microsoft Fabric URL

Setting Where to find it
Workspace ID URL after /groups/
Workspace name Workspace settings

3. Deploy Azure resources

azd auth login

azd up

When you start the deployment, you will need to set the following parameters:

Setting Description Default value
Azure Region The region where resources will be created. (empty)
Environment Name A unique 3–20 character alphanumeric value used to prefix resources, preventing conflicts with others. env_name
AI Model Location The region where AI model will be created *(empty)

Different tenant? Use: azd auth login --tenant-id <tenant-id>

4. Setup Python environment

python -m venv .venv

.venv\Scripts\activate   # or: source .venv/bin/activate

pip install uv && uv pip install -r scripts/requirements.txt

5. Build the solution

az login

VS Code Web users: Use az login --use-device-code since browser-based login is not supported in VS Code Web.

python scripts/00_build_solution.py --from 02

Note: Press Enter key to start or Ctrl+C to cancel the process.

6. Test the agent

python scripts/08_test_agent.py

Sample questions to try:

  • "How many tickets have priority = 'High'?"
  • "What is the average score from inspections?"
  • "What are the requirements for handling customer tickets?"
  • "Are we meeting our resolution targets for high priority tickets according to our Ticket Management Policy?"

7. Test the Fabric Data Agent

  1. Go to your Microsoft Fabric workspace
  2. Select "New item" → Search for "Data Agent" → select data agent, provide a name and click create
  3. Add data source → Select your Ontology resource for this workshop
  4. Click Agent instructions from top menu and add the below agent instructions: 
    You are a helpful assistant that can answer user questions using data.
Support group by in GQL.
  5. Click Publish from the top menu and select Publish.

Note: The Ontology set up may take a few minutes so retry after some time if you don't see good responses.

Sample questions to try:

  • "How many tickets have priority = 'High'?"
  • "What is the average score from inspections?"
  • Show tickets grouped by status.

8. Deploy and launch the application

azd env set AZURE_ENV_DEPLOY_APP true

azd up

9. Set up app permissions

python scripts/00_build_solution.py --from 09

Note: Press Enter key to start or Ctrl+C to cancel the process.

After the agent configuration & API permission set up completes, open the app URL shown in the output.

10. Customize for Your Industry (Optional)

Follow steps in this page to Customize for your use case.

Option B: Azure-Only Deployment

1. Clone the repository

git clone https://github.com/microsoft/agentic-applications-for-unified-data-foundation-solution-accelerator.git

cd agentic-applications-for-unified-data-foundation-solution-accelerator

2. Enable Azure-only mode

azd env set AZURE_ENV_ONLY true

3. Deploy Azure resources

azd auth login

azd up

Choose environment name and region. Different tenant? Use: azd auth login --tenant-id <tenant-id>

4. Setup Python environment

python -m venv .venv

.venv\Scripts\activate   # or: source .venv/bin/activate

pip install uv && uv pip install -r scripts/requirements.txt

cp .env.example .env # or: copy .env.example .env

5. Build the solution

az login

VS Code Web users: Use az login --use-device-code since browser-based login is not supported in VS Code Web.

python scripts/00_build_solution.py --from 04

Note: Press Enter key to start or Ctrl+C to cancel the process.

6. Test the agent

python scripts/08_test_agent.py

Sample questions to try:

  • "How many outages occurred last month?"
  • "What's the average resolution time?"
  • "What are the policies for notifying customers of outages?"
  • "Which outages exceeded the maximum duration defined in our policy?"

7. Deploy the application

azd env set AZURE_ENV_DEPLOY_APP true

azd up

8. Set up app permissions

python scripts/00_build_solution.py --from 09

Note: Press Enter key to start or Ctrl+C to cancel the process.

After the agent configuration & API permission set up completes, open the app URL shown in the output.

9. Customize for Your Industry (Optional)

Follow steps in this page to Customize for your use case.

Repository: github.com/microsoft/agentic-applications-for-unified-data-foundation-solution-accelerator