Microsoft AI-103: Azure AI App and Agent Developer Associate

Try the platform for free

• 0

  Lab 0: Deploy Your First Azure AI Model with Microsoft FoundryFree

  Deploy a GPT model in Microsoft Foundry and make your first successful call to it from a real Azure environment.

  Exam objective: Plan and manage an Azure AI solution

  • Microsoft Foundry
  • Model deployment
  • Azure OpenAI

  Run this lab free

  60 min

Portal-only orientation. No code, no downloads. Learn your way around Microsoft Foundry before Lab 2.

• 1

  Lab 1: Exploration of Microsoft Foundry

  Navigate every Foundry surface, create a project, and deploy gpt-5-mini, so you know where each capability lives before you build.

  Exam objective: Plan and manage an Azure AI solution

  • Microsoft Foundry
  • Projects and hubs
  • Model deployment

  45-60 min
• 2

  Lab 2: Day One at Project Horizon

  Deploy three models side by side and choose between them with evidence, using the Model Leaderboard, the comparison playground, and an LLM as judge script.

  Exam objective: Plan and manage an Azure AI solution

  • Model selection
  • Model leaderboard
  • LLM as judge
  • Cost and quality trade-offs

  90-120 min
• 3

  Lab 3: First Prototype

  Call models with the Azure OpenAI Responses API, keep conversation memory across turns, and stream replies so a chat app feels responsive.

  Exam objective: Implement generative AI and agentic solutions

  • Responses API
  • Conversation state
  • Streaming
  • Python SDK

  90-120 min
• 4

  Lab 4: Grounding in Reality

  Turn a raw model into a reusable Foundry agent with instructions and tools, and ground its answers in real data instead of guesses.

  Exam objective: Implement generative AI and agentic solutions

  • Foundry Agents
  • Grounding
  • Tool use
  • Agent instructions

  90 min
• 5

  Lab 5: Understanding Guest Reviews with Azure AI Language

  Apply Azure AI Language to real guest feedback: sentiment, key phrases, named entities, PII detection, and summarisation.

  Exam objective: Implement text analysis solutions

  • Azure AI Language
  • Sentiment analysis
  • Entity recognition
  • PII detection
  • Summarisation

  45-60 min
• 6

  Lab 6: The Stress Test

  Attack your own assistant with jailbreaks, prompt injection, and data leakage attempts, then close the gaps with a custom content filter and blocklist.

  Exam objective: Plan and manage an Azure AI solution

  • Content safety
  • Prompt injection
  • Guardrails
  • Blocklists
  • Responsible AI

  45-60 min
• 7

  Lab 7: Giving Solara Real Powers with MCP

  Connect an agent to a live Model Context Protocol server so it can take real actions against external systems, not just talk about them.

  Exam objective: Implement generative AI and agentic solutions

  • MCP
  • Tool calling
  • Foundry Agents
  • Integrations

  90-120 min
• 8

  Lab 8: Solara's Internal Knowledge Vault with Foundry IQ

  Build a knowledge base from internal documents with Foundry IQ and attach it to an agent so answers are grounded in company sources.

  Exam objective: Implement generative AI and agentic solutions

  • Foundry IQ
  • RAG
  • Knowledge bases
  • Grounded answers

  90-120 min
• 9

  Lab 9: The Quality Gate

  Run a production style evaluation with Foundry Evaluations: synthetic datasets, judge models, and the core quality and safety evaluators.

  Exam objective: Plan and manage an Azure AI solution

  • Foundry Evaluations
  • Judge models
  • Synthetic data
  • Quality metrics

  55-60 min
• 10

  Lab 10: Solara's Multi-Agent Team

  Design a team of specialist agents in the Workflows canvas and wire them into sequential and group chat patterns, then read the routing in traces.

  Exam objective: Implement generative AI and agentic solutions

  • Multi-agent workflows
  • Orchestration
  • Agent handoff
  • Tracing

  90-120 min
• 11

  Lab 11: What the Photos Say

  Use model vision to turn photographs into structured inspection reports and accessibility alt text, and learn where a visual answer stops being reliable.

  Exam objective: Implement computer vision solutions

  • Multimodal vision
  • Structured output
  • Alt text
  • Azure AI Vision

  60-90 min
• 12

  Lab 12: Making the Picture

  Generate images with gpt-image-2, edit what you generated rather than starting over, and turn a still into a short video with Sora 2.

  Exam objective: Implement computer vision solutions

  • Image generation
  • Image editing
  • Video generation
  • Multimodal

  60-90 min
• 13

  Lab 13: Reading the Paperwork

  Use Content Understanding to turn an invoice PDF and a contract into structured fields you can check against each other.

  Exam objective: Implement information extraction solutions

  • Content Understanding
  • Document intelligence
  • Structured extraction

  60-90 min