We learn to talk before we can read, write, or type, making voice humanity's most natural form of communication. Yet for decades, we have forced ourselves to interact with machines through rigid code and text queries. Today, technology is finally catching up. By combining speech-to-text, large language models, and text-to-speech, we can now build seamless conversational interfaces that solve real-world problems. To prove it, I built an AI voice assistant to help me navigate the incredibly complex rules, mechanics, and card interactions of Magic: The Gathering. Plus, I'll walk you through a bunch of other use cases, and how you can apply this to your applications.

In this talk we'll cover:

- Why voice is the natural progression of AI development and human-machine interaction
- How to combine speech-to-text, LLMs, and text-to-speech into a seamless loop
- Live demo
- Limitations of voice tech and how to get around these
- Best practices for managing latency, context, and user experience in conversational apps
- A walk-through of the codebase and the existing tools you can use to start experimenting
- Other practical ways teams can leverage voice AI for productivity and accessibility

This demo is practical, fun, and showcases how developers can blend modern AI tools with natural language to turn spoken words into delightful, real-world applications.