1.1. The AI Capability Spectrum

💡 First Principle: Building AI is like cooking—you can order takeout (pre-built APIs), use a meal kit with some customization (fine-tunable services), or cook from scratch (custom ML). Each level trades convenience for control: pre-built is fastest but generic; custom is slowest but exact. The exam tests whether you can match a scenario's constraints (time, data, expertise) to the right level on this spectrum.

The following table summarizes the trade-offs at each level of the spectrum. Use this to quickly determine which approach fits your scenario based on available resources.

Azure AI Services Overview