At Memphis Agent Camp, I delivered an online session on the evolution of voice AI from rigid, multi-step pipelines to real-time conversational systems. The talk highlighted how audio-first models enable fluid, human-like interactions by reducing latency and supporting continuous, interruptible dialogue.
I walked through practical design patterns for building voice-enabled agents that integrate with enterprise systems, including tool usage, workflow orchestration, and secure data access. The session included a live demonstration of a real-time voice agent, providing attendees with actionable insights into building scalable and reliable conversational AI solutions.
Event details: https://www.meetup.com/memphis-technology-user-groups/events/312731040/
Event site: https://memphisagentcamp.github.io/
Additional materials (slides, video, code): https://udai.io/building-real-time-voice-ai-from-pipelines-to-conversations/