Teach
Azure AI Fundamentals starts with choosing the right workload type before choosing services or models.
Core AI workload types
- Regression predicts a continuous value (for example, delivery time or sales amount).
- Classification predicts a category (for example, spam/not spam, fraud/legit).
- Clustering groups similar records without pre-labeled classes (for example, customer segments).
- Natural Language Processing (NLP) works with text and language understanding.
- Computer Vision (CV) works with images and video.
- Generative AI (Gen-AI) creates new text, code, images, or other content from prompts.
Azure-oriented solution patterns
- Use regression when business output is numeric and precision matters.
- Use classification when outcomes are discrete labels and decisions need thresholds.
- Use clustering when you need discovery and pattern mining before explicit labels exist.
- Use NLP for chat, summarization, translation, extraction, and sentiment.
- Use CV for detection, OCR, scene understanding, and visual quality checks.
- Use Gen-AI for copilots, draft generation, retrieval-augmented chat, and workflow automation.
Quick decision checkpoints
- What is the output type: number, label, group, text/vision understanding, or generated content?
- Is training custom or can prebuilt Azure AI capabilities solve it faster?
- Do you need explainability, low latency, or content controls from day one?
Pick the workload first; Azure service selection becomes much easier afterward.
Practice
Practice 1
Predicting monthly revenue amount from historical data is a:
Practice 2
Tagging an email as spam or not spam is usually:
Practice 3
Grouping customers into behavior-based segments without labels is:
Practice 4
A chatbot that drafts responses from prompts mainly uses:
Practice 5
Extracting text from scanned invoices is best mapped to: