5.2. Azure NLP Services

💡 First Principle: Azure AI Language handles text analysis (entities, sentiment, key phrases). Azure AI Speech handles audio conversion (speech-to-text, text-to-speech). Azure Translator handles text-to-text translation only. Knowing which service handles which modality prevents common exam errors.

What breaks without knowing these boundaries: The #1 NLP exam trap: "Which service translates spoken language in real-time?" If you think Azure Translator handles all translation, you'll pick wrong—Translator is TEXT-to-text only. Speech translation requires Azure AI Speech service. Similarly, language detection is a Language service feature, not a Speech feature.

Imagine these services like different translation professionals. Azure AI Language is a document analyst who reads and interprets written text. Azure AI Speech is an interpreter who converts between spoken and written forms. Azure Translator is a text translator who works only with written documents. For instance, consider a scenario: "A company needs to transcribe customer service calls and analyze sentiment." Which services? Speech (transcribe audio to text) THEN Language (analyze sentiment). What if they also need real-time translation of those calls? Add Speech translation—NOT Translator, because audio is involved. The exam tests whether you know who handles what modality.

Azure provides specialized services for NLP capabilities. Understanding which service handles which capability is essential for exam success.