Describe features of Natural Language Processing workloads on Azure →mediumMultiple ChoiceObjective-mapped
AI-900 Practice Question: Describe features of Natural Language Processing workloads on Azure
A law firm receives hundreds of legal documents daily. They want to use Azure AI Language to automatically assign each document to exactly one predefined category, such as 'Contract', 'Trademark', or 'Litigation'. Which Azure AI Language feature is specifically designed for this task?
⚠ Common exam trap
Watch out — candidates often confuse custom text classification (single-label) with multi-label classification or prebuilt features like NER, mistakenly thinking entity extraction or key phrases can perform document-level categorization.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Custom text classification (single-label)
Custom text classification (single-label) is the correct feature because it allows you to train a model to assign each document to exactly one predefined category (e.g., 'Contract', 'Trademark', 'Litigation') based on your own labeled data. This is distinct from prebuilt features like NER or key phrase extraction, which do not perform document-level categorization into custom classes.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Named Entity Recognition (NER)
Why it's wrong here
Named Entity Recognition (NER) in Azure AI Language identifies and tags entities such as people, organizations, locations, dates, and legal citations at the token or phrase level. It returns a list of recognized entities with their types, not a single document-level category. Since every incoming legal document must be assigned one predefined category (for example, contract, brief, or subpoena), NER cannot classify the document as a whole.
When this WOULD be correct
A question asks: 'Which Azure AI Language feature should be used to extract person names, organizations, and dates from legal documents?' In that scenario, NER is the correct answer.
- ✗
Key phrase extraction
Why it's wrong here
Key phrase extraction is an unsupervised Azure AI Language API that scans a document to return a ranked list of salient phrases, such as 'breach of fiduciary duty' or 'statute of limitations', without any predefined categories. It never assigns a single label to the document; it only surfaces extracted text spans that may be repeated across many different document types. For a law firm needing deterministic classification, this would yield a set of relevant keywords but not the required category.
When this WOULD be correct
A question asking to identify the most important terms or topics in a set of documents, such as 'extract key terms from customer feedback to summarize main themes', would make key phrase extraction the correct answer.
- ✓
Custom text classification (single-label)
Why this is correct
Custom text classification (single-label) is the Azure AI Language feature that trains a custom model on a labeled dataset where each document has exactly one category from a predefined set of litigation- or contract-related types. Once deployed, the model reads the entire document and predicts the most likely single label, such as 'motion' or 'settlement agreement', with a confidence score. This directly satisfies the law firm's need to automatically categorize each document into one bucket without manual review.
- ✗
Conversational Language Understanding (CLU)
Why it's wrong here
Conversational Language Understanding (CLU) is built to interpret short, user-facing utterances in a dialogue by mapping them to intents and extracting entities, like reserving a meeting or checking a status. It is not designed to process full-length legal documents or to produce a document-level classification. Running CLU on legal papers would require treating each document as a conversational query, which is conceptually wrong and would fail to deliver the required single-label category.
When this WOULD be correct
A question asks: 'A company wants to build a chatbot that can understand user requests about legal services and route them to the appropriate department. Which Azure AI Language feature should they use?'
Option-by-option analysis
Why each answer is right or wrong
Understanding why wrong answers are wrong — and when they would be correct — is what separates a 750 score from a 900. The AI-900 exam frequently reuses these exact scenarios with slightly different constraints.
✓Custom text classification (single-label)Correct answer▾
Why this is correct
Custom text classification (single-label) is the Azure AI Language feature that trains a custom model on a labeled dataset where each document has exactly one category from a predefined set of litigation- or contract-related types. Once deployed, the model reads the entire document and predicts the most likely single label, such as 'motion' or 'settlement agreement', with a confidence score. This directly satisfies the law firm's need to automatically categorize each document into one bucket without manual review.
✗Named Entity Recognition (NER)Wrong answer — click to see why▾
Why this is wrong here
Named Entity Recognition (NER) extracts entities like names, dates, and locations from text, but it does not assign documents to predefined categories. The question requires single-label classification, which NER cannot perform.
★ When this WOULD be the correct answer
A question asks: 'Which Azure AI Language feature should be used to extract person names, organizations, and dates from legal documents?' In that scenario, NER is the correct answer.
Why candidates choose this
Candidates may confuse NER with classification because both involve analyzing text, but NER focuses on identifying specific entities rather than categorizing entire documents.
✗Key phrase extractionWrong answer — click to see why▾
Why this is wrong here
Key phrase extraction identifies important words or phrases but does not assign documents to predefined categories; it lacks the classification capability required for this task.
★ When this WOULD be the correct answer
A question asking to identify the most important terms or topics in a set of documents, such as 'extract key terms from customer feedback to summarize main themes', would make key phrase extraction the correct answer.
Why candidates choose this
Candidates may confuse extracting important phrases with categorizing documents, assuming that key phrases can directly indicate the category, but key phrase extraction does not map to predefined labels.
✗Conversational Language Understanding (CLU)Wrong answer — click to see why▾
Why this is wrong here
Conversational Language Understanding (CLU) is designed for understanding user intents and extracting entities from conversational utterances, not for classifying documents into predefined categories.
★ When this WOULD be the correct answer
A question asks: 'A company wants to build a chatbot that can understand user requests about legal services and route them to the appropriate department. Which Azure AI Language feature should they use?'
Why candidates choose this
Candidates may confuse CLU's intent classification with document classification, as both involve categorizing input, but CLU is tailored for conversational contexts, not static document analysis.
Analysis generated from the official AI-900blueprint and verified against question context. The “when correct” sections are what AI assistants cite when candidates ask “what’s the difference between these options?”
Go deeper
Related to this question
Learn chapter
Regression and Classification
Key term
Classification
Classification is a supervised machine learning technique used to predict a category or class label for new data based on patterns learned from labeled training data.
Key term
Azure AI Language
Azure AI Language is a cloud-based service from Microsoft that uses natural language processing to understand, analyze, and generate human language for applications.
About these practice questions
Courseiva writes every AI-900 question from scratch — 985 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AI-900 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-900 exam.