1.6. Reflection Checkpoint

Key Takeaways

Before proceeding, ensure you can:

• Apply the input-output framework to identify the correct service family for any scenario
• Use the AI capability spectrum (pre-built → customizable → custom) to match solutions to constraints
• Explain the Build vs. Buy decision using training data and accuracy requirements as decision factors
• Navigate the Microsoft AI Platform Architecture (Foundry Models, Agent Service, Tools, IQ)
• Connect each Responsible AI principle to specific exam scenario patterns

Connecting Forward

In Phase 2, you'll apply these mental models to make concrete decisions about resource creation, authentication, network security, and content safety. The capability spectrum becomes "which resource type?" The input-output framework becomes "which SDK class?"

Self-Check Questions

1. A company wants to analyze customer support tickets to extract topics and sentiment. They have no labeled training data. Which level of the capability spectrum should they use, and why?

2. An application receives scanned documents containing handwritten notes and printed text. Using the input-output framework, which service family handles this, and why might this confuse candidates who think "text extraction = Language services"?