Instiq
Chapter 5 · Implement Generative AI and Agentic Solutions·v2.0.0·Updated 6/3/2026·~10 min

What's changed: Deepened AI-102 Chapter 5 (ja figures; comparison tables/scenarios/FAQ/traps/deep paragraphs in all sections)

5.1Azure OpenAI, Prompts, and RAG

Key points

Understand Azure OpenAI Service for using large language models on Azure, prompt engineering, and RAG (retrieval-augmented generation) for answers grounded in your own data—the core of building generative AI.

Azure OpenAI Service lets you use large language models like GPT under enterprise security/compliance. Output quality is improved via prompts and grounding data (RAG).

5.1.1Generative AI building blocks

Diagram showing model deployments (GPT chat/completion, embeddings) in Azure OpenAI Service, an app calling with a system prompt + user prompt, RAG retrieving relevant data from Azure AI Search to ground the prompt, and Content Safety inspecting input/output.
Azure OpenAI, prompts, and RAG
  • Azure OpenAI: deploy and call GPT-family models—chat/completion, embeddings, images—via API.
  • Prompt engineering: set role/constraints with a system prompt and steer output with few-shot examples.
  • RAG: retrieve relevant data (vector/semantic search) and pass it to the prompt as grounding.
  • Parameters: tune output diversity/determinism with temperature, etc.
Exam point

Common on AI-102: enterprise GPT = Azure OpenAI Service, set role/constraints = system prompt, answers grounded in your data = RAG (retrieval + prompt), embeddings = vectorization (for similarity search). To reduce hallucinations, ground with RAG.

Continue reading — free sign-up

You're reading the free preview. Sign up free to read this section in full, plus every chapter (including 4+) and all questions.

Get started freeLog in