1.1 Azure AI Services Overview and Selection
Key Takeaways
- Azure AI services are organized into categories: Vision, Language, Speech, Decision, OpenAI, Search, Document Intelligence, and Content Safety.
- You can deploy Azure AI services as multi-service resources (single endpoint for multiple services) or single-service resources (dedicated endpoint per service).
- Azure AI Foundry is the unified portal for building, evaluating, and deploying AI solutions — it replaces the older Azure AI Studio.
- Each Azure AI service has specific pricing tiers (Free, Standard, Premium) with different throughput limits, features, and SLAs.
- Choosing the right service requires matching business requirements to service capabilities, considering latency, cost, compliance, and regional availability.
Azure AI Services Overview and Selection
Quick Answer: Azure AI services include Azure AI Vision, Azure AI Language, Azure AI Speech, Azure OpenAI Service, Azure AI Search, Azure AI Document Intelligence, Azure AI Content Safety, and Azure AI Translator. You can deploy them as multi-service or single-service resources. Azure AI Foundry is the unified portal for building and managing AI solutions.
The Azure AI Services Landscape
Azure AI provides a comprehensive set of services for building intelligent applications. Understanding which service to use for each scenario is a core AI-102 skill.
Service Categories and Capabilities
| Category | Service | Key Capabilities |
|---|---|---|
| Vision | Azure AI Vision | Image analysis, OCR, spatial analysis, object detection, image captioning |
| Vision | Azure AI Custom Vision | Train custom image classification and object detection models |
| Vision | Azure AI Face | Face detection, verification, identification, liveness detection |
| Vision | Azure AI Video Indexer | Video analysis, transcription, face identification, scene detection |
| Language | Azure AI Language | Sentiment analysis, NER, key phrases, PII detection, CLU, summarization |
| Language | Azure AI Translator | Text translation, document translation, custom translator |
| Speech | Azure AI Speech | Speech-to-text, text-to-speech, speech translation, speaker recognition |
| Search | Azure AI Search | Full-text search, AI enrichment, vector search, semantic ranking |
| Document | Azure AI Document Intelligence | Form extraction, receipt/invoice processing, custom document models |
| Safety | Azure AI Content Safety | Text and image moderation, prompt shields, groundedness detection |
| Generative | Azure OpenAI Service | GPT-4o, GPT-4, DALL-E, Whisper, embeddings, fine-tuning |
Multi-Service vs. Single-Service Resources
Azure offers two deployment patterns for AI services:
Multi-Service Resource (Azure AI Services)
- Single resource with one endpoint and one set of keys
- Access to multiple AI services (Vision, Language, Speech, etc.) through one resource
- Simplified billing — one invoice for all services
- Best for: Prototyping, small projects, cost management simplicity
- Limitation: Not available for Azure OpenAI Service (requires its own resource)
Single-Service Resource
- Dedicated resource for one specific AI service
- Separate endpoint and keys per service
- Granular billing and usage tracking per service
- Independent scaling and configuration
- Best for: Production workloads, compliance requirements, service-specific quotas
# Multi-service resource — one endpoint for Vision, Language, Speech
Endpoint: https://my-ai-services.cognitiveservices.azure.com/
# Single-service resources — separate endpoints
Vision: https://my-vision.cognitiveservices.azure.com/
Language: https://my-language.cognitiveservices.azure.com/
Speech: https://my-speech.cognitiveservices.azure.com/
On the Exam: When a question mentions needing a "single endpoint" for multiple cognitive capabilities, the answer is a multi-service Azure AI Services resource. When the question mentions granular access control or separate billing per service, choose single-service resources.
Azure AI Foundry Portal
Azure AI Foundry (formerly Azure AI Studio) is the unified portal for building, evaluating, and deploying AI solutions. It provides:
| Feature | Description |
|---|---|
| Model Catalog | Browse and deploy models from OpenAI, Meta, Mistral, and other providers |
| Prompt Flow | Visual workflow builder for designing AI pipelines with prompt engineering |
| Evaluation | Built-in tools to evaluate model quality, groundedness, and safety |
| Content Safety | Configure content filters and safety guardrails for AI applications |
| Deployments | Manage model deployments with version control and traffic routing |
| Data Management | Connect data sources for RAG, fine-tuning, and evaluation |
| Agent Builder | Build and test autonomous AI agents with tool-calling capabilities |
Azure AI Foundry Projects and Hubs
Azure AI Foundry organizes work into hubs and projects:
- Hub: A top-level container that manages shared resources (compute, storage, connections) across multiple projects. Think of it like an Azure resource group for AI work.
- Project: A workspace within a hub for a specific AI solution. Projects have their own deployments, data, and evaluation results.
Service Selection Decision Framework
When the exam presents a scenario requiring an AI service, use this decision tree:
Text Processing
- Extract sentiment, entities, key phrases → Azure AI Language
- Translate text between languages → Azure AI Translator
- Extract fields from invoices, receipts, forms → Azure AI Document Intelligence
- Generate text, answer questions, summarize → Azure OpenAI Service
- Build a custom intent/entity model → Azure AI Language (CLU)
- Moderate text for harmful content → Azure AI Content Safety
Image/Video Processing
- Analyze image content (tags, captions, objects) → Azure AI Vision (Image Analysis)
- Extract text from images → Azure AI Vision (OCR/Read)
- Train a custom image classifier → Azure AI Custom Vision
- Detect and verify faces → Azure AI Face
- Analyze video content → Azure AI Video Indexer
Audio Processing
- Convert speech to text → Azure AI Speech (STT)
- Convert text to speech → Azure AI Speech (TTS)
- Translate spoken language → Azure AI Speech (Translation)
- Identify speakers → Azure AI Speech (Speaker Recognition)
Search and Knowledge
- Full-text search with AI enrichment → Azure AI Search
- Build a knowledge base from documents → Azure AI Search + Knowledge Store
- RAG (Retrieval-Augmented Generation) → Azure AI Search + Azure OpenAI Service
Which Azure resource type provides a single endpoint for accessing multiple AI services such as Vision, Language, and Speech?
A company needs to extract fields from invoices AND analyze sentiment in customer reviews. They want simplified billing with one resource. Which deployment pattern should they use?
What is the primary purpose of Azure AI Foundry?