AI, AI-900, Artificial Intelligence (AI), Microsoft Certification, Natural Language Processing (NLP)

Identify Features and Uses for Entity Recognition (AI-900 Exam Prep)

Where this fits in the exam

  • Exam domain: Describe features of Natural Language Processing (NLP) workloads on Azure (15–20%)
  • Sub-area: Identify features of common NLP workload scenarios
  • Key skill tested: Understanding what entity recognition is, what it’s used for, and which Azure service provides it

You are not expected to build or train models—only to recognize capabilities and use cases.

What Is Entity Recognition?

Entity recognition (also called Named Entity Recognition or NER) is an NLP capability that identifies and categorizes specific, real-world items mentioned in text.

These items (entities) typically fall into predefined categories such as:

  • People
  • Organizations
  • Locations
  • Dates and times
  • Numbers
  • Products
  • Email addresses, phone numbers, URLs

Simple example

Input text:

“Satya Nadella is the CEO of Microsoft, headquartered in Redmond.”

Extracted entities:

  • Person: Satya Nadella
  • Organization: Microsoft
  • Location: Redmond

Azure Service That Provides Entity Recognition

Entity recognition is provided by Azure AI Language, part of Azure’s AI services portfolio.

Key points for the exam:

  • Uses prebuilt models
  • No machine learning expertise required
  • Accessed via REST APIs or SDKs
  • Supports multiple languages

Types of Entity Recognition in Azure AI Language

For AI-900, you mainly need to understand the concept, but it helps to know the types at a high level.

1. Named Entity Recognition

Identifies common entity categories, such as:

  • Person
  • Location
  • Organization
  • Date
  • Quantity

2. Personally Identifiable Information (PII) Detection

Detects sensitive personal data, including:

  • Phone numbers
  • Email addresses
  • Social security numbers
  • Credit card numbers

This is often tested conceptually in the context of compliance and data privacy.

Common Use Cases for Entity Recognition

1. Information Extraction

Automatically pull important data from unstructured text such as:

  • Contracts
  • Emails
  • Reports
  • Support tickets

2. Search and Indexing

Improve search by identifying names, locations, or products mentioned in documents.

3. Data Classification and Tagging

Label documents based on recognized entities to:

  • Route support tickets
  • Organize content
  • Trigger workflows

4. Compliance and Security

Detect and flag PII to:

  • Prevent data leaks
  • Meet regulatory requirements
  • Mask sensitive data

Entity Recognition vs Other NLP Capabilities

This comparison is very exam-relevant.

CapabilityWhat it identifies
Entity recognitionSpecific items (names, places, dates)
Key phrase extractionMain topics and concepts
Sentiment analysisEmotional tone
Language detectionLanguage of the text

If the question asks “Who, where, or what specifically?” → entity recognition
If it asks “What is this text about?” → key phrase extraction

Key Features to Remember for the Exam

  • Uses pretrained models
  • Works with unstructured text
  • Supports multiple languages
  • Does not require labeled training data
  • Commonly used for information extraction and compliance

Responsible AI Considerations

Microsoft emphasizes responsible AI even at the fundamentals level.

Important considerations:

  • Entity recognition may misidentify entities due to ambiguity
  • Results should be reviewed before being used for critical decisions
  • Sensitive data detection should align with privacy and compliance policies

Exam Tips

  • Expect scenario-based questions, not code
  • Focus on matching the right NLP capability to the scenario
  • Look for keywords like:
    • names, addresses, dates, organizations → Entity recognition
    • topics, summaries → Key phrase extraction
    • opinions, feelings → Sentiment analysis

Go to the Practice Exam Questions for this topic.

Go to the AI-900 Exam Prep Hub main page.

Leave a comment