Microsoft · AI-901
Text and speech apps
AI-901 · Lesson 8 of 10 · 50:36
YouTube loads only after you press play. This chapter opens at its timestamp in the full course. Watch on YouTube ↗
In this lesson
Prepare for the AI-901 exam (Microsoft Azure AI Fundamentals). In this lesson: text analysis by prompting a model vs Azure Language in Foundry Tools (detect_language, recognize_pii_entities), sending spoken prompts to an audio-capable model (input_audio, base64, Chat Completions), speech to text and text to speech with the Azure Speech SDK (SpeechConfig, SpeechRecognizer, SpeechSynthesizer, recognize_once vs continuous and batch), and real-time voice with Voice Live. Includes three original practice questions.
Preview the topics
- An insurer must store call notes with names and phone numbers masked, and auditors require identical output for identical input. What should the developer use?
- A voicemail app keeps only the first sentence of 90-second messages. It calls recognize_once. What’s the best fix?
- A hotel wants a lobby kiosk where guests talk naturally, can interrupt, and hear quick spoken answers, without chaining three services. What should the team use?