Explore how artificial intelligence turns spoken language into text and written text into speech. Intro to Speech AI is Part 4 of the eight-part Microsoft AI Fundamentals AI-901 series, building on the earlier parts with an introduction to speech technology and its use in Azure. Begin with speech recognition and speech synthesis. Learn how audio capture, preprocessing, acoustic models, language models, and phonemes contribute to speech-to-text. Then examine how text normalization, pronunciation, prosody, and waveform generation help text-to-speech produce spoken output. Connect these ideas to applications such as meeting transcription, accessibility, voice assistants, and reading content aloud.
Move into Azure Speech in Microsoft Foundry. Explore the available capabilities and the Speech playground, including transcription and voice selection. Review introductory Python Speech SDK examples to understand how an application configures authentication, accepts audio input, recognizes speech, and generates spoken responses. You will also be introduced to Voice Live and how speech-capable agents combine generative AI, tools, and real-time conversation, including handling interruptions. The course connects these technologies to responsible AI considerations, including fairness, privacy, consent, and handling language data. Quizzes, exercise guidance, a simulation exercise, and lab activities help reinforce the concepts and connect them to practical workflows.
No prior experience with speech AI is required. Basic AI and cloud knowledge is helpful, and the first three parts of this series provide useful background. Familiarity with Python can help you follow the code examples, but the conceptual lessons are accessible without programming experience. The optional browser speech exercise doesn’t require an Azure subscription; the live Azure lab does, along with access to the relevant services. A compatible browser, microphone, and speakers or headphones are useful for speech activities.
