Getting Started with Azure AI Foundry

What is Azure AI Foundry?

Azure AI Foundry (formerly Azure AI Studio) is Microsoft's unified platform for building, evaluating, and deploying generative AI solutions. It brings together the best of Azure's AI capabilities into a single, cohesive experience.

Key Capabilities

• Model Catalog — Access 1,600+ foundation models from OpenAI, Meta, Mistral, Microsoft, and more
• Prompt Flow — Visual orchestration tool for building LLM-powered workflows
• Evaluation — Built-in evaluation tools for measuring model quality, safety, and groundedness
• Deployments — One-click deployment with managed endpoints and auto-scaling

Prerequisites

• An Azure subscription (create one free)
• Access to Azure AI Foundry portal at ai.azure.com

Step 1: Create a Hub and Project

A Hub is a top-level resource that provides shared infrastructure. A Project lives inside a hub and is where you do your actual work.

1. Navigate to ai.azure.com
2. Click + New project
3. Select or create a Hub (choose your Azure subscription and resource group)
4. Name your project and click Create

Step 2: Explore the Model Catalog

The Model Catalog is your starting point for choosing a foundation model.

1. In your project, go to Model catalog in the left nav
2. Browse models by provider, task, or license type
3. Click on a model to see benchmarks, pricing, and capabilities
4. Use Compare to evaluate models side by side

Step 3: Deploy a Model

1. Select a model (e.g., GPT-4o)
2. Click Deploy
3. Choose deployment type: Serverless API (pay-per-token) or Managed Compute
4. Configure rate limits and content filters
5. Once deployed, copy the endpoint URL and key

Step 4: Use Prompt Flow

Prompt Flow lets you build multi-step AI workflows visually.

1. Go to Prompt flow in the left nav
2. Click + Create and choose a template or start from scratch
3. Add nodes: LLM calls, Python code, embeddings, search, etc.
4. Connect nodes to create your orchestration logic
5. Test with the built-in chat interface

Step 5: Evaluate Your Solution

1. Go to Evaluation in the left nav
2. Create a new evaluation run
3. Choose built-in metrics: Groundedness, Relevance, Coherence, Fluency
4. Upload a test dataset or use AI-generated test data
5. Review results and iterate

Resources

• Azure AI Foundry Documentation
• Azure AI Foundry Quickstart
• Microsoft Learn: Azure AI Foundry Training

Video: Watch the official walkthrough on Microsoft Azure YouTube Channel — search for "Azure AI Foundry getting started"