Skip to content

Analyze and Translate Text

This section of the Microsoft AI-102: Designing and Implementing a Microsoft Azure AI Solution exam covers analyzing and translating text with Azure AI Language and Azure AI Translator. Below are study notes for each sub-topic, with links to Microsoft documentation, exam tips, and key facts

Extract Key Phrases and Entities

📖 Docs: Key phrase extraction | Entity recognition

Overview

  • Azure AI Language can extract important information from text
  • Key phrase extraction identifies the main points
  • Entity recognition identifies people, places, organizations, dates, and more

Key Points

  • Supports multiple languages
  • Custom Named Entity Recognition (NER) available for domain-specific terms
  • Outputs structured JSON with identified entities and categories

Exam Tip

Watch for use cases: summarizing topics → key phrases, identifying people/places → entities

Determine Sentiment of Text

📖 Docs: Sentiment analysis

Overview

  • Sentiment analysis detects positive, neutral, or negative opinions in text
  • Opinion mining extracts targeted sentiment about specific aspects

Key Points

  • Returns sentiment scores for document, sentence, and aspect levels
  • Supports social media, customer reviews, and survey feedback
  • Multilingual support

Use Case

Analyzing customer reviews to detect satisfaction levels

Detect the Language Used in Text

📖 Docs: Language detection

Overview

  • Identifies the primary language of a text input
  • Supports over 100 languages

Key Points

  • Returns language name and ISO 639-1 code
  • Provides confidence score
  • Useful for routing text to the correct translation or processing service

Exam Tip

Keywords like detect language before translation → Language Detection API

Detect Personally Identifiable Information (PII) in Text

📖 Docs: PII detection

Overview

  • Detects and redacts sensitive data such as:
    • Names
    • Phone numbers
    • Credit card numbers
    • Social Security Numbers

Key Points

  • Can return redacted text or structured PII entities
  • Helps ensure compliance (GDPR, HIPAA, etc.)
  • Configurable for different entity categories

Exam Tip

Questions may ask how to mask sensitive data in text → answer: PII detection

Translate Text and Documents by Using the Azure AI Translator Service

📖 Docs: Translator service overview

Overview

  • Azure AI Translator provides real-time text translation
  • Supports over 100 languages
  • Modes:
    • Text-to-text
    • Document translation (PDF, Word, etc.)

Key Points

  • Uses neural machine translation
  • Can integrate with applications via REST API or SDKs
  • Document translation preserves layout and formatting

Use Case

Translating contracts from English to French while keeping original structure

Quick‑fire revision sheet

  • 📌 Key phrases = main points, Entities = names, places, dates
  • 📌 Sentiment analysis = positive/neutral/negative + opinion mining
  • 📌 Language detection returns ISO code + confidence score
  • 📌 PII detection redacts sensitive info for compliance
  • 📌 Translator = text + document translation with formatting preserved