Microsoft Learn lab materials.

Detailed write-up coming in a later update.

← All projects

How it works — components

Language understanding

CLU / LUIS intents and entities map free text to structured actions — the theory is to separate "what the user wants" (intent) from "the values" (entities) so downstream code stays deterministic.

Document intelligence

Form Recognizer / OCR turns unstructured PDFs and images into key-value pairs and tables using layout models trained on spatial features, not just characters.

Vision & speech

Pretrained Computer Vision and Speech-to-Text services expose embeddings/transcripts via REST; the lab covers when to use prebuilt models vs. custom training.

Responsible AI

Content safety, confidence thresholds, and human-in-the-loop review — the exam emphasizes calibrating automation against the cost of a wrong prediction.