6.2.1. Azure AI Foundry
Azure AI Foundry (formerly Azure AI Studio) is Microsoft's unified platform for building AI applications. Think of it as a one-stop shop for AI development—rather than jumping between different Azure services and portals, Foundry brings everything together in a single environment.
Why Azure AI Foundry exists: Before Foundry, building AI applications required navigating multiple Azure services, each with its own portal, API, and learning curve. Foundry solves this by providing a unified interface where you can access models, build prompts, test applications, and deploy solutions—all in one place.
Key features:
- Central hub for AI development - Single workspace for all AI projects
- Access to multiple model providers - Azure OpenAI, open-source models, and third-party models
- Prompt engineering tools - Test and refine prompts before deployment
- Enterprise security and governance - Built-in compliance and access controls
- Model evaluation and comparison - Test different models on your data
What Azure AI Foundry provides:
- Unified development environment - Code, test, and deploy from one interface
- Access to Azure OpenAI models - GPT-4, GPT-3.5, DALL-E, Embeddings
- Access to third-party models - Open-source models from Hugging Face, Meta, and others
- Prompt flow - Visual tool for orchestrating AI workflows and chains
- Responsible AI tools - Built-in content filters and safety guardrails
- Deployment options - Deploy to Azure endpoints with monitoring
Azure AI Foundry vs Azure OpenAI Service:
| Feature | Azure AI Foundry | Azure OpenAI Service |
|---|---|---|
| Purpose | Development platform | Model hosting service |
| Models | Multiple providers | OpenAI models only |
| Interface | Visual studio + API | API primarily |
| Use case | Build and experiment | Production deployment |
⚠️ Exam Tip: Azure AI Foundry is the PLATFORM for AI development. Azure OpenAI Service is one of the model PROVIDERS accessible through that platform. Questions about "unified AI development" or "single hub" point to Foundry.
Prompt Flow in Azure AI Foundry: Prompt Flow is a visual tool for building AI workflows that chain multiple steps together:
| Component | What It Does |
|---|---|
| Inputs | Define what data flows into the workflow |
| LLM nodes | Call language models with prompts |
| Python nodes | Run custom code for data processing |
| Outputs | Define what the workflow returns |
Example workflow:
- User asks a question
- Python node searches a knowledge base
- LLM node generates response using retrieved context
- Output returns the grounded answer
Model evaluation capabilities: Azure AI Foundry helps you choose the right model by:
- Running the same prompts across multiple models
- Comparing response quality and accuracy
- Measuring latency and cost differences
- Evaluating safety and responsible AI metrics
Deployment from Foundry: Once you've built and tested your AI application:
- Create a deployment endpoint
- Configure scaling settings
- Enable monitoring and logging
- Integrate with your applications via API
