AI-103T00 is Microsoft's new intermediate-level developer course titled "Develop AI Apps and Agents on Azure" — the 2026 successor to AI-102T00. It is designed for software developers who want to build production-ready AI solutions on Microsoft Azure AI Foundry. The course covers the full spectrum of modern AI development: deploying and optimizing generative AI models, implementing RAG pipelines, building autonomous AI agents with tool integration and multi-agent orchestration, and working with natural language, speech, computer vision, and document intelligence capabilities.

The course is organized into 4 Learning Paths. Each learning path maps directly to a skill domain of the AI-103 exam (Microsoft Certified: Azure AI Apps and Agents Developer Associate) — a brand-new certification launching in 2026. Students build real, deployable applications using Python, the Azure AI Projects SDK, Semantic Kernel, Azure OpenAI Service, and Azure AI Foundry.

AI-103T00 Course Summary:

Course Title	Develop AI Apps and Agents on Azure
Course Code	AI-103T00
Associated Certification	Microsoft Certified: Azure AI Apps and Agents Developer Associate
Replaces	AI-102T00: Designing and Implementing a Microsoft Azure AI Solution
Duration	4 Days
Level	Intermediate (Developer)
Structure	4 Learning Paths
Delivery	Instructor-Led Training (ILT) / Online
Exam Price	$165 (USD)
Passing Score	700 / 1000
Schedule Exam	Pearson VUE
Prerequisites	Proficiency in Python; experience using REST APIs and SDKs; familiarity with Azure (AI-901 or equivalent recommended)
Target Role	Azure AI App and Agent Developer
Official Course Page	learn.microsoft.com/en-us/training/courses/ai-103t00

Learning Path 1: Develop Generative AI Apps in Azure

Build generative AI applications that use language models to interact with users, grounded in enterprise data. Covers the full pipeline from model deployment to optimized, responsible production applications using Azure AI Foundry.

Topic	Details
Get Started with Azure AI Foundry	- Overview of Azure AI Foundry as the unified AI development platform - Creating projects and hubs in Azure AI Foundry - Navigating the Azure AI Foundry portal and SDK - Deploying models from the model catalog (GPT-4o, Llama, Mistral, Phi) - Managing API keys, endpoints, and connections - Azure AI Projects SDK (azure-ai-projects) for programmatic access
Develop a Generative AI App	- Building chat-completion apps with Azure OpenAI Service (GPT-4o) - Using the Azure OpenAI Python SDK and REST API - System prompts, message history, and multi-turn conversations - Streaming responses for real-time user experiences - Responsible AI: Azure AI Content Safety integration - Deploying and testing apps in Azure AI Foundry playground
Prompt Engineering & Model Optimization	- Prompt engineering techniques: zero-shot, few-shot, chain-of-thought - System prompt design for reliable, consistent outputs - Parameter tuning: temperature, top_p, max_tokens, stop sequences - Comparing prompt engineering vs RAG vs fine-tuning — when to use each - Evaluating model output quality in Azure AI Foundry - Fine-tuning models for domain-specific behavior
Implement RAG (Retrieval Augmented Generation)	- RAG architecture: retriever + generator pattern - Creating and populating an Azure AI Search vector index - Generating embeddings with Azure OpenAI Embeddings models - Connecting Azure AI Search as a knowledge source in Azure AI Foundry - Grounding LLM responses with enterprise data (PDFs, docs, databases) - Implementing RAG pipelines using LangChain on Azure - Evaluating and improving retrieval quality
Multimodal & Image Generation	- Using GPT-4o for multimodal (text + image) inputs - Image generation with DALL-E 3 via Azure OpenAI - Generating code with GitHub Copilot models on Azure AI Foundry - Vision capabilities: image analysis in generative AI workflows

Learning Path 2: Develop AI Agents on Azure

Understand, design, and build autonomous AI agents using Microsoft Foundry Agent Service, Semantic Kernel, and the Microsoft Agent Framework. Covers single-agent and multi-agent systems with tool integration and memory.

Topic	Details
Introduction to AI Agents	- What is an AI agent? Agentic AI vs standard AI apps - When to use AI agents vs a simple LLM call - Agent anatomy: LLM + tools + memory + planning loop - Azure AI Foundry Agent Service overview - Comparing agent frameworks: Semantic Kernel, AutoGen, LangGraph
Build Agents with Microsoft Foundry Agent Service	- Creating and configuring agents in Azure AI Foundry - Defining agent instructions (system prompts) and personas - Adding tools to agents: code interpreter, file search, custom functions - Function calling: connecting agents to external APIs and services - Agent threads, runs, and messages lifecycle - Handling tool call responses and multi-step reasoning
Build Agents with Semantic Kernel	- Semantic Kernel SDK for Python: kernels, plugins, and planners - Creating native plugins (functions) and semantic plugins (prompts) - Auto function invocation and automated planning - Memory and context management in Semantic Kernel - Integrating Semantic Kernel with Azure OpenAI - Building chat agents with Semantic Kernel chat history
Multi-Agent Orchestration	- Multi-agent architecture: orchestrator + specialist agents - Agent handoff patterns and delegation - Building multi-agent workflows with Microsoft Agent Framework - Shared memory and context across agents - Error handling, retry logic, and guardrails in agentic systems - Responsible AI safeguards: content filtering and grounding for agents
Knowledge & Tool Integration for Agents	- Connecting agents to Azure AI Search (RAG for agents) - Using file search and vector stores in agent tool belt - Connecting agents to external APIs and databases - Integrating agents with Microsoft 365 and Power Platform - Evaluating agent performance and reliability

Learning Path 3: Develop Natural Language Solutions in Azure

Use Azure AI Foundry and Azure AI Language / Speech / Translator services to build apps that understand and generate text, transcribe and synthesize speech, and translate languages.

Topic	Details
Analyze Text with LLMs & Azure AI Language	- Using LLMs (via Azure OpenAI) for semantic text analysis - Azure AI Language service: sentiment analysis, NER, key phrase extraction - PII detection and redaction in text - Language detection and text classification - Summarization (extractive and abstractive) with Azure AI Language - Conversational language understanding (CLU) models - Question answering projects with custom knowledge bases
Process Speech with Azure AI Speech	- Real-time speech-to-text and batch transcription - Text-to-speech with neural voices (SSML customization) - Custom Speech models for domain-specific vocabulary - Keyword recognition and voice activity detection - Speaker diarization (who said what) - Speech translation (speech-to-translated-text) - Integrating Azure AI Speech SDK in Python applications
Translate Language with Azure AI Translator	- Text translation across 100+ languages - Document translation (batch) - Custom Translator: training domain-specific translation models - Language detection and transliteration - Integrating translation in agentic and app workflows
Build Conversational AI Solutions	- Designing conversational AI with intents, entities, utterances - CLU (Conversational Language Understanding) on Azure AI Foundry - Building and deploying bots with Azure Bot Service - Connecting bots to Teams, web chat, and other channels - Integrating generative AI (LLMs) into bot conversations

Learning Path 4: Extract Insights from Visual Data on Azure

Use generative AI, Azure AI Vision, and Azure AI Content Understanding to extract insights from images, videos, and documents — powering vision-based agents and information-extraction pipelines for RAG.

Topic	Details
Analyze Images with Azure AI Vision & GPT-4o	- Azure AI Vision: image analysis, tagging, object detection, captioning - Using GPT-4o multimodal for visual Q&A and image description - Azure AI Custom Vision: training custom image classifiers - Optical character recognition (OCR) with Azure AI Vision Read API - Azure AI Face service: detection, verification, identification - Spatial analysis: detecting people and movement in video feeds
Process Video with Azure AI Video Indexer	- Indexing videos for transcription, speaker identification, and key topics - Extracting text overlays and labels from video frames - Searching video content semantically - Integrating video insights into agentic workflows
Extract Data with Azure AI Document Intelligence	- Prebuilt models: invoices, receipts, business cards, ID documents, W-2 - Custom extraction models: training on your own document layouts - Composed models for multi-document workflows - Extracting tables, key-value pairs, and structured fields from PDFs - Using Document Intelligence output as input for RAG pipelines - Azure AI Content Understanding for complex, unstructured documents
Build AI-Powered Search & Knowledge Mining	- Azure AI Search: creating indexes, indexers, data sources, and skillsets - Built-in cognitive skills: OCR, key phrase extraction, NER, image analysis - Vector search and hybrid search (keyword + semantic + vector) - Semantic ranker for improved relevance - Custom skills: calling external APIs in an enrichment pipeline - Knowledge store: persisting enriched output to blob/table storage - Combining Azure AI Search with Azure OpenAI for enterprise RAG

Key Learning Outcomes

• Deploy, configure, and call models from Azure AI Foundry using the azure-ai-projects SDK
• Build production-grade generative AI applications with Azure OpenAI Service (GPT-4o, DALL-E)
• Implement RAG pipelines combining Azure AI Search and Azure OpenAI for grounded responses
• Build autonomous AI agents using Foundry Agent Service, Semantic Kernel, and function calling
• Design and deploy multi-agent systems using the Microsoft Agent Framework
• Analyze text, detect entities, and build conversational apps with Azure AI Language
• Transcribe, synthesize, and translate speech with Azure AI Speech
• Analyse images and video using Azure AI Vision and GPT-4o multimodal
• Extract structured data from documents with Azure AI Document Intelligence
• Implement responsible AI: content safety, evaluation, grounding, and safeguards

To ensure success in the AI-103 exam and real-world AI development, we recommend completing the official Microsoft Learn learning paths, hands-on labs in Azure AI Foundry, and practice tests aligned to the Microsoft Certified: Azure AI Apps and Agents Developer Associate certification.