Voice Agents in 2026: What works, what doesn't, and how to fix them
Voice agents have moved from demo territory to production workloads. But the gap between a voice agent that impresses in a demo and one that holds up when real users start talking is still wide.
In this session, two senior product leaders from Agora and Murf AI will get into the specifics - the architecture decisions, the trade-offs, and the hard lessons that only come from shipping at scale.
No keynote slides. No polished talking points. Just a focused conversation between two practitioners about what it actually takes to build voice agents that work.
What you will walk away with
A clear mental model for thinking through latency, quality, and reliability trade-offs in a voice agent stack
An honest picture of what is working in production today and which use cases are proving out fastest
Practical engineering techniques to improve voice agent performance
Real answers on the hardest problems - interruption handling, noisy environments, multilingual deployments and more
Who should attend
Developers and engineers building voice agents
CTOs and technical leaders evaluating voice AI for their products
Product managers working on voice-first experiences
Founders building in the conversational AI space
Speakers
Patrick Ferriter, SVP Product, Agora
Patrick leads product at Agora, a global leader in real-time engagement infrastructure powering voice and video for developers building at scale. With a career spanning Polycom and Dolby Laboratories, he has spent years at the frontier of real-time audio technology and is now deep in the infrastructure that makes voice agents actually work in production.
Sahil Gupta, SVP Product, Murf AI
Sahil leads product at Murf AI, one of the world's leading AI voice platforms with over 10 million users across 195 countries. He brings close to a decade of product experience at Amazon, including Alexa and robotics, and now focuses on building the next generation of enterprise voice AI.
Register above to get the link and a reminder before the session.