# Azure AI Cognitive Services

> Chapter 95 of the Courseiva AZ-900 curriculum — https://courseiva.com/learn/az-900/azure-cognitive-services

**Official objective:** 2.2 — Azure Architecture Services

## Introduction

This chapter covers Azure Cognitive Services, pre-built AI/ML capabilities. AZ-900 tests this under objective 2.2.

## AI Capabilities Without Building AI

Training an AI model from scratch is complex and expensive. Cognitive Services is like buying pre-trained AI capabilities off-the-shelf: vision, language, speech processing — ready to use without building the AI yourself.

## Core explanation

### What Cognitive Services are
Cognitive Services provide pre-trained, ready-to-use AI capabilities: computer vision (image recognition), natural language processing, speech recognition, translation — without needing to train models yourself.

### Common services
Vision (image analysis, OCR), Language (text analysis, sentiment), Speech (recognition, synthesis), Decision (anomaly detection, content moderation).

### No ML expertise required
Organizations without ML expertise can add AI capabilities to applications through simple API calls.

## Real-world context

An app adds image recognition without building an ML model: call Computer Vision API with an image, get recognized objects back.

## Exam focus

Objective 2.2 expects understanding of Cognitive Services as pre-built AI capabilities, not deep ML knowledge.

Common wrong answer: assuming Cognitive Services require ML expertise — they abstract that away.

Stable terms: **Cognitive Services**, **pre-trained**. Memory trick: Cognitive Services = AI capabilities, pre-trained, API-accessible, no model training needed.

## Step by step

1. **Identify needed capability** — Determine which Cognitive Service fits the need (vision, language, speech).
2. **Create a Cognitive Service resource** — A service instance is provisioned.
3. **Call the API** — Applications call the API to use the service.

## Common misconceptions

- **Misconception:** Cognitive Services require building ML models. **Reality:** They're pre-trained and ready-to-use via APIs.

## Key takeaways

- Cognitive Services provide pre-trained AI capabilities via APIs.
- Cover vision, language, speech, decision domains.
- No ML expertise or model training required.

## FAQ

**What are Azure Cognitive Services?**

They're pre-trained AI/ML capabilities (vision, language, speech) accessible via APIs, without requiring model training expertise.

---

Interactive version with quiz and diagrams: https://courseiva.com/learn/az-900/azure-cognitive-services