One-Pager · 8 Services

Azure AI & Machine Learning at a Glance

Every Azure AI and ML service on one page — generative AI, pre-built models, custom training, and intelligent search. A quick reference for architects building AI-powered solutions.

Azure OpenAI Service

Access GPT-4o, GPT-4, GPT-3.5, DALL-E, and embedding models via Azure-hosted API. Enterprise security, private networking, content filtering, regional data residency.

Chat & copilotsContent generationGPT-4oEnterprise

Azure AI Services

Pre-built cognitive APIs for vision, speech, language, and decision. No ML expertise needed. OCR, sentiment analysis, translation, speech-to-text, and more.

Vision & OCRSpeech & languagePre-builtREST API

Azure Machine Learning

End-to-end MLOps platform. AutoML, designer, notebooks, managed endpoints, model registry, responsible AI dashboard. Train on GPU clusters, deploy anywhere.

Custom modelsMLOpsAutoMLGPU training

Azure AI Search

Enterprise search with vector, semantic, and hybrid retrieval. The retrieval layer for RAG architectures. Built-in AI enrichment pipeline, integrated vectorization.

RAG retrievalEnterprise searchVector + semanticAI enrichment

Document Intelligence

Extract text, key-value pairs, tables, and structure from documents. Pre-built models for invoices, receipts, IDs, W-2s. Custom model training for domain-specific forms.

Invoice processingForm extractionOCRCustom models

Databricks (ML)

Unified analytics and ML on Apache Spark. MLflow for experiment tracking, Delta Lake for feature stores, distributed training on large datasets. Shared with data engineering.

Large-scale MLFeature engineeringSparkMLflow

Bot Service

Framework for building conversational AI bots. Power Virtual Agents for low-code, Bot Framework SDK for custom. Deploy to Teams, web, Slack, and other channels.

ChatbotsTeams integrationMulti-channelLow-code

Content Understanding

Multimodal AI for analyzing documents, images, audio, and video. Extracts structured fields from unstructured content. Pre-built and custom analyzers.

Multimodal analysisContent extractionDocumentsVideo/audio

Quick Comparison

Service	Category	Best For	ML Expertise	Min Cost
Azure OpenAI	Generative AI	Chat, content gen, reasoning	Low	Per 1K tokens
AI Services	Pre-built AI	Vision, speech, language tasks	None	Free tier (5K txn/mo)
Machine Learning	Custom ML	Train & deploy custom models	High	Free tier (compute costs)
AI Search	Search & RAG	RAG retrieval, enterprise search	Low	~$75/mo (Basic)
Document Intelligence	Document AI	Invoice/form extraction, OCR	Low	Free tier (500 pages/mo)
Databricks ML	Platform ML	Large-scale training, feature eng	High	~$0.40/DBU
Bot Service	Conversational	Chatbots, virtual assistants	Low–Medium	Free (Standard channel)
Content Understanding	Multimodal AI	Analyze docs, images, video	Low	Per transaction

Common Architecture Patterns

Pattern	Services	Description
RAG (Retrieval-Augmented Generation)	OpenAI + AI Search + Blob Storage	Ground LLM responses in your own data. Vectorize documents, retrieve context, generate answers.
Intelligent Document Processing	Document Intelligence + OpenAI + Cosmos DB	Extract structured data from forms and invoices, enrich with LLM, store in database.
Custom ML Pipeline	Machine Learning + Databricks + ADLS	Feature engineering in Databricks, model training in AML, deploy to managed endpoints.
Enterprise Copilot	OpenAI + AI Search + Bot Service + Teams	Internal knowledge assistant grounded in company data, deployed to Microsoft Teams.
Real-time Vision	AI Services (Vision) + Functions + Event Grid	Process uploaded images for classification, OCR, or object detection in serverless pipeline.