Voice agents in Microsoft Foundry: Prototype to production in 45 mins | BRK244
Marvin Hsu, Matt Landis, Yiwen Rong and Archer Zhao present a practical look at moving voice agents from prototype to production, covering the capabilities available in Microsoft Foundry and what enterprise customers are running today.
Building and operating speech-native agents
The session focuses on voice-first agents: agents built around spoken interaction, rather than text agents with voice added afterward. It examines that distinction alongside the challenges of putting voice-based experiences into production.
Microsoft Foundry provides a single place to build, deploy and operate these agents. The session covers:
- Building: Creating a speech-native agent and understanding the capabilities available.
- Deployment: Delivering the agent through web, mobile, Microsoft Teams and telephony.
- Operations: Using audio-aware tracing and evaluation to assess voice interactions.
- Enterprise use: Examining what enterprise customers are running today.
Session resource
This Microsoft Ignite 2026 breakout session is part of the Agent Platform track and is classified as 200 — Intermediate.