Domain 4 of 5 · Chapter 1 of 2

NLP Scenarios

Unlock the complete study guide + 1,040 practice questions across 16 full exams.

Bundled into the existing Microsoft Azure AI Fundamentals premium course — no separate purchase.

Included in this chapter:

  • Text analytics: the four extraction tasks the exam keeps re-asking
  • Translation, speech (STT/TTS), and conversational AI
  • Exam-pattern recognition: scenario → NLP capability, and the distractor traps
NLP taskWhat it doesExample scenario
Sentiment analysisClassifies emotional tone as positive, negative, or neutral with a 0–1 confidence scoreFlag negative product reviews to gauge customer mood
Entity recognitionIdentifies and categorizes mentioned things (person, place, organization, date, quantity)Pull all people and locations named in a news article
Key phrase extractionReturns the main talking points / themes of a documentSummarize what survey responses are mostly about
Language detectionIdentifies the language of text and returns name, ISO code, and confidenceRoute incoming messages to the right-language support queue
TranslationConverts text or speech from a source language into one or more target languagesTranslate multilingual support tickets into English
Speech (STT / TTS)Speech-to-text transcribes spoken audio to text; text-to-speech synthesizes audio from textCaption a recorded meeting (STT); voice a navigation app (TTS)

Decision tree

Spoken audio in or out (not just text)? Yes No Speech (STT / TTS) transcribe or synthesize audio Convert between languages? Yes No Translation source → target language Gauge opinion / mood of the text? Yes No Sentiment Analysis positive / negative / neutral score Pull names / places / orgs / dates? Yes No Entity Recognition categorize named things in text Key Phrase Extraction main talking points / themes Conversational AI (bots / virtual agents) composes several of these tasks rather than picking one.
Cheat sheet

Cheat sheet

  • Sentiment analysis scores mood as positive/neutral/negative, 0-1
  • Sentiment is evaluated at both document and sentence level
  • Opinion mining ties a sentiment to a specific aspect
  • Key phrase extraction returns a document's main talking points
  • NER identifies and categorizes the entities named in text
  • Entity linking disambiguates a mention to a known reference
  • PII detection identifies, classifies, and redacts personal data
  • Language detection identifies the language; it never translates
  • Translator does neural machine translation, text in to text out
  • Speech-to-text turns spoken audio into written text
  • Text-to-speech synthesizes audible speech from text
  • Speech translation recognizes and translates spoken audio
  • Conversational AI leans on language understanding plus question answering
  • Conversational AI is the whole-scenario, back-and-forth answer
  • Reading text from an image is computer vision (OCR), not NLP
  • Sentiment returns a 0-1 confidence for positive, neutral, and negative
  • Tokenization, stemming, frequency, and n-grams are statistical text primitives

Unlock with Premium — includes all practice exams and the complete study guide.

Also tested in

References

  1. Sentiment analysis and opinion mining - Azure AI Language
  2. Key phrase extraction - Azure AI Language
  3. Named Entity Recognition (NER) - Azure AI Language
  4. Language detection - Azure AI Language
  5. What is Azure AI Translator?
  6. What is speech to text? - Azure AI Speech
  7. What is text to speech? - Azure AI Speech
  8. What is speech translation? - Azure AI Speech
  9. What is question answering? - Azure AI Language