Domain 2 of 2 · Chapter 2 of 4

Implement AI solutions for text and speech by using Foundry

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.

14-day money-back guarantee — no questions asked.

Included in this chapter:

  • Three clients, and this objective is choosing between them
  • Building the text-analysis client
  • What each core text capability returns
  • Speech to text: pick the mode, then build the recognizer
  • One config, one audio object, and the class that names the direction
  • Answering out loud with a deployed realtime model
  • Telling the three paths apart under pressure

Which of the three build paths a requirement lands on

Decision pointAzure Language in Foundry ToolsAzure Speech in Foundry ToolsDeployed realtime Foundry model
What you sendA list of written text documentsSpoken audio, or the text you want spokenLive streamed audio, plus optional text and images
What comes backStructured findings: categorized entities, confidence scores, redacted text, detected languageTranscribed text, synthesized audio, or translated textStreamed audio replies with matching transcript events
Python surfaceTextAnalyticsClientSpeechRecognizer, SpeechSynthesizer, TranslationRecognizerA GPT Realtime API session over WebRTC, WebSocket, or SIP
Representative callrecognize_entities(documents)recognize_once_async(), speak_text_async()session.update to configure, then response.create to answer
Signature scenarioFlag people and organizations named in support emailTranscribe stored call recordings, or read a notification aloudAnswer a caller's spoken question in speech, in real time
Wrong pick whenThe content arrives as soundThe app must compose a spoken answer, not just convertThe job is bulk transcription of recordings already in storage

Decision tree

Is the input spoken audio?sound, not charactersNo: written textYes: spoken audioMust the text be spoken aloud?audio is the deliverableYesNo: analyse itSpeechSynthesizerspeak_text_async(), AudioOutputConfigTextAnalyticsClientthe method matches the finding neededrecognize_entities(): typed mentions | recognize_pii_entities(): redacted_textdetect_language(): language and code | begin_analyze_healthcare_entities(): clinicalMust the app answer out loud?speech in, speech out, in one turnYesDeployed realtime modelGPT Realtime API: WebRTC, WebSocket, SIPNoIs a target language named?the output must change languageYesTranslationRecognizeradd_target_language() takes the codeNoIs the audio live, or stored?when the sound existsLiveSpeechRecognizerrecognize_once_async(): one utteranceStoredBatch transcriptionREST API or CLI, results retrieved laterAlways: naming a language is language detection (text) or language identification (audio); neither one translates.
Cheat sheet

Cheat sheet

  • Named entity recognition categorizes common entities in text
  • The text client sends documents to a selected analysis method
  • PII detection identifies and can redact sensitive entities
  • Language detection identifies the input language
  • Text analytics for health adds clinical relationships and assertions
  • Batch transcription processes large volumes of prerecorded audio
  • SpeechRecognizer handles short microphone or file utterances
  • Azure Speech in Foundry Tools powers Speech SDK applications
  • SpeechSynthesizer converts written text to spoken audio
  • TranslationRecognizer returns target-language text from speech
  • Realtime multimodal models support speech-in, speech-out conversations
  • Realtime audio transports match the client interaction shape

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

Also tested in

References

  1. What is Azure Language in Foundry Tools?
  2. What is Azure Speech in Foundry Tools?
  3. Use the GPT Realtime API for speech and audio with Azure OpenAI
  4. What is named entity recognition (NER) in Azure Language?
  5. Azure Text Analytics client library for Python
  6. What is key phrase extraction in Azure Language in Foundry Tools?
  7. What is summarization in Azure Language?
  8. What is PII detection in Azure Language?
  9. What is language detection in Azure Language?
  10. What is Text analytics for health in Azure Language?
  11. Batch transcription overview - Azure Speech
  12. Quickstart: Recognize and convert speech to text
  13. Quickstart: Convert text to speech
  14. Quickstart: Translate speech into another language