Domain 4 of 5 · Chapter 2 of 2

Azure NLP Services

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:

  • The four Azure NLP services and what each one does
  • Which service for which NLP task
  • Exam-pattern recognition: stems, traps, and how to beat them

Azure NLP services: which one for the workload

ServiceBest for
Azure AI LanguageText understanding: sentiment, named entity recognition, key phrase extraction, language detection, PII detection, summarization, plus trainable conversational language understanding (CLU) and custom question answering
Azure AI SpeechAudio workloads: speech-to-text (transcription), text-to-speech (neural/custom voices), speech translation, and spoken-language identification
Azure AI TranslatorText-to-text neural machine translation across many languages, document translation that preserves formatting, and Custom Translator for domain terminology
Azure AI Bot ServiceBuilding, hosting, and channel-connecting conversational bots/assistants; delegates language understanding to Azure AI Language (CLU + question answering)

Decision tree

Is audio the input or the output? Yes (speech) Azure AI Speech STT, TTS, speech translation No (text in) Translate text language-A to written language-B? Yes Azure AI Translator text-to-text NMT, docs No Assembling a chatbot / multi-channel assistant? Yes Azure AI Bot Service hosting + channels (delegates to Language) No (analyse text) Azure AI Language sentiment, NER, CLU, question answering Anchor on modality: audio = Speech; written A to B = Translator; conversational front end = Bot Service; text-meaning = Language.
Cheat sheet

Cheat sheet

  • AI-900 NLP is a four-service selection drill
  • Pick the NLP service by the input/output modality
  • Azure AI Language is the single text-understanding service
  • Text Analytics, LUIS, and QnA Maker are now Azure AI Language
  • CLU predicts an utterance's intent and extracts its entities
  • Question answering returns the best stored answer from a knowledge base
  • Azure AI Speech handles audio in both directions
  • Translating spoken audio is Speech translation, not Translator
  • Identifying a spoken language is a Speech feature
  • Azure AI Translator is text-in, text-out neural machine translation
  • Azure AI Bot Service is the conversational framework, not an analysis service
  • A bot delegates language understanding to Azure AI Language
  • Recorded-call analysis chains Speech then Language
  • OCR and generated answers are not NLP services
  • NER returns confidence scores; key phrase extraction doesn't
  • Translator auto-detects the source language when 'from' is omitted
  • Transliteration changes the script, not the language

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

References

  1. What is Azure AI Language?
  2. What is the Speech service?
  3. What is Azure AI Translator?
  4. What is Azure AI Bot Service?