The complete preparation pathway for Microsoft's Azure AI Apps and Agents Developer – Associate exam (AI-103). Microsoft Foundry, RAG and agents, computer vision, text analysis, and information extraction.
Azure AI-103: Exam Overview
What AI-103 tests, five domain weightings, six production scenarios, scoring (700/1000 to pass), and how to prepare.
Microsoft Foundry Naming, Explained
Azure AI Studio → Azure AI Foundry → Microsoft Foundry — the rebrand timeline, and Foundry Tools vs Foundry Agent Service vs Foundry IQ.
Embeddings & Vector Search: Complete Guide
How embeddings power retrieval — foundational for choosing indexing and retrieval methods in Foundry.
Human-in-the-Loop: When to Let an Agent Run
Approval workflows, oversight modes, and tool-access controls — the responsible AI instrumentation AI-103 expects you to configure.
RAG Context Injection Pipeline Design
Chunking, retrieval orchestration, and context assembly — implementing RAG in a Foundry application.
RAG vs Agentic RAG
When retrieval-augmentation beats agent loops, and multi-step retrieval for Foundry Agent Service workflows.
Multi-Agent Orchestration Patterns
Sequential handoff, planner/router, and parallel agent patterns — a core AI-103 Domain 2 competency and common exam trap.
How to Build Your First Agent Loop
Tool invocation cycles, role/goal/tool-schema definitions, and stopping conditions for Foundry agents.
Context Engineering for RAG Systems
Assembling retrieved documents, metadata, and system instructions for Foundry-based generative apps.
Semantic vs Vector vs Hybrid Search
The single most confusable topic on AI-103 — what each retrieval approach does and how Azure AI Search implements all three.
Document OCR & Field Extraction
OCR, layout analysis, and field extraction as distinct pipeline stages — the Content Understanding pattern for information extraction.
What Is Multimodal AI?
Visual understanding, captioning, and multimodal reasoning — the foundation for AI-103's computer vision and speech domains.
Structured Output & JSON Schema Enforcement
Structured JSON extraction via generative prompting — a core text-analysis and information-extraction competency.
Bias in AI: Types, Examples, and Mitigation
Responsible AI for multimodal content — unsafe content filters, indirect prompt injection via embedded image text, and content moderation.
Practice exam
3 timed mock exams with shuffled questions, instant scoring, and per-question explanations. Pass score: 700/1000. The fastest way to find your weak domains before exam day.
Microsoft's Associate-level certification for Azure AI engineers who build, manage, and deploy generative AI apps and agents using Microsoft Foundry and Python. The 120-minute exam covers five domains with a minimum passing score of 700 out of 1,000 (the one officially published number — the exact question count is not published, so we estimate ~40 for pacing).
Python developers and AI engineers building on Microsoft Foundry — RAG, agents, computer vision, text analysis, and information extraction pipelines. Familiarity with general AI and generative AI concepts is expected.
Microsoft's AI platform has rebranded twice in about two years: Azure AI Studio → Azure AI Foundry → Microsoft Foundry (current, as of Jan 2026). Foundry Tools (prebuilt APIs) is easily confused with Foundry Agent Service (the agent runtime) — this pathway has a dedicated article untangling it.
14 articles across all five exam domains, approximately 13 hours of study. The pathway mirrors exam weighting: heaviest on Implement generative AI and agentic solutions (Domain 2 at 33%, the largest single domain).
The pathway includes scenario-based quiz questions throughout, including the semantic-vs-vector-vs-hybrid-search distinction flagged as the exam's most confusable topic. After completing it, use the mock tests at /tests/azure-ai-apps-agents-developer — timed, full-length practice exams with shuffled questions and explanations.
Comfort reading and writing Python, familiarity with REST APIs, and basic AI/ML concepts. The Building AI Agents and MCP: Model Context Protocol pathways on this platform cover prerequisite knowledge if you need to build up first.
Understand what AI actually is — tokens, transformers, agents, and the landscape. Start here if you're new.
11 articles · ~4h →Go from vague requests to precise, reproducible AI outputs. The skill that underpins everything.
12 articles · ~5h →Go from zero to productive with Claude Code — the terminal AI coding agent that ships real projects.
13 articles · ~7h →Understand and build the loops, harnesses, and protocols that make AI agents reliable and autonomous.
14 articles · ~6h →Practical AI adoption for your specific function — marketing, engineering, HR, finance, and more.
10 articles · ~4h →Navigate the crowded model market — Claude, GPT, Gemini, open-source — and understand the tradeoffs.
10 articles · ~6h →
