Domain 5 of 5 · Chapter 2 of 2

Azure GenAI Services

Unlock the complete study guide + 1,040 practice questions across 16 full exams.

Bundled into the existing Microsoft Azure AI Fundamentals premium course — no separate purchase.

Included in this chapter:

  • The four Azure generative AI building blocks
  • Which service for which need
  • Exam-pattern recognition: service-selection stems and distractors
ServiceWhat it isUse it when
Azure OpenAI ServiceEnterprise-secured access to OpenAI GPT, embedding, and DALL-E modelsYou need to call OpenAI-class generative models with Azure identity, security, and compliance
Azure AI Foundry (formerly Azure AI Studio)Platform to build/manage gen-AI apps: model catalog, prompt flow, project hubYou are building, comparing, evaluating, or operationalizing a generative AI solution
Microsoft CopilotPrebuilt generative AI assistants embedded in Microsoft productsUsers need a ready-made productivity or coding assistant, not a custom app
Azure AI Content SafetyDetects harmful content and prompt-injection; checks groundednessYou need responsible-AI guardrails around generative input and output

Decision tree

Goal is to filter harmful content, jailbreaks, or check groundedness? Yes Azure AI Content Safety harm filter + Prompt Shields No Need a ready-made assistant in a product, with no development? Yes Microsoft Copilot prebuilt, embedded in products No Building / comparing / evaluating a gen-AI app (catalog, prompt flow)? Yes Azure AI Foundry model catalog, prompt flow, project hub No: just call a model Azure OpenAI Service GPT / embeddings / DALL-E, Entra ID + RBAC Concept-level (what gen-AI is, tokens) lives in Generative AI Solutions; predictive ML uses Azure Machine Learning, not these gen-AI services.
Cheat sheet

Cheat sheet

  • Use Azure OpenAI to call OpenAI models under Azure security
  • Azure OpenAI has three AI-900 model families: GPT, embeddings, image
  • Azure OpenAI adds the Azure control plane over the public OpenAI API
  • Ground Azure OpenAI answers on your own data with RAG
  • Azure AI Foundry is the platform to build a whole gen-AI solution
  • Azure AI Studio is the old name for Azure AI Foundry
  • The AI Foundry model catalog browses and deploys models from many vendors
  • AI Foundry prompt flow orchestrates and evaluates multi-step prompt workflows
  • Microsoft Copilot is a family of ready-made, embedded AI assistants
  • Copilot is the answer when the scenario wants zero development
  • Microsoft 365 Copilot grounds on your files via Microsoft Graph, honoring permissions
  • Azure AI Content Safety is the guardrail layer for generative AI
  • Content Safety scores four harm categories: hate, sexual, violence, self-harm
  • Prompt Shields block jailbreak and prompt-injection attacks
  • Groundedness detection flags responses not grounded in the source material
  • On gen-AI questions, the wrong answers are usually non-generative services
  • OpenAI vs Foundry: let the verb in the stem decide
  • The AI Foundry playground is the no-code place to test and compare models
  • Prompt flow has three flow types: standard, chat, and evaluation
  • Deploy a model to an endpoint before you can call it
  • The Chat Completions API structures input as system, user, and assistant messages
  • DALL-E 3 offered vivid (default) vs natural style, but it is now retired
  • A harmful image prompt is blocked and returns no image
  • Content Safety severity maps to four display levels over a 0–7 scale
  • Copilot Studio connectors are API proxies that reach external services
  • Copilot Studio is the low-code platform to build, extend, and publish agents

Unlock with Premium — includes all practice exams and the complete study guide.

Also tested in

References

  1. What is Azure OpenAI Service?
  2. Azure OpenAI On Your Data
  3. What is Azure AI Foundry?
  4. Microsoft 365 Copilot overview
  5. What is Azure AI Content Safety?