Voice Agents Forum
Hosted by the Agentic AI Foundation
Voice agents are being deployed across customer support, healthcare, financial services, enterprise operations, and consumer applications. As they move into production, teams need to manage latency, turn-taking, interruptions, handoff, reliability, and the operational constraints that come with live voice systems.
Users notice latency quickly. Long pauses can make a system feel unresponsive. Missed interruptions disrupt conversation. Poor handoffs create frustration.
Voice Agents Forum is a one-day event for teams building and deploying voice agents in production.
The event brings together engineers, founders, product leaders, researchers, and operators working on the technical and operational challenges behind voice AI.
About the event
Voice systems combine speech, models, orchestration, infrastructure, and user experience in ways that make failures difficult to isolate and evaluate. Decisions in one part of the stack can affect latency, reliability, conversation quality, and how easily a system can be monitored or improved.
Voice Agents Forum is designed to give teams a chance to compare how others are approaching those problems, hear where different technical choices create tradeoffs, and discuss the issues that become more important as voice agents move from prototypes into deployed systems.
The focus will be on useful technical and operational lessons around architecture, evaluation, observability, orchestration, reliability, compliance, and human handoff.
Hosted by the Agentic AI Foundation and brought together with the MLOps Community.
Agenda
9:00 AM - 9:45 AM Registration
9:35 AM - 10:25 AM Opening Keynotes
10:25 AM - 11:00 AM Break
11:00 AM - 12:15 PM Talks
12:20 PM - 2:00 PM Lunch & Expo Hall
2:00 PM - 3:00 PM Hands-On Workshops (Limited Availability)
3:00 PM - 3:20 PM HOT TAKES
3:25 PM - 3:55 PM Lightning Talks
3:55 PM - 4:45 PM Closing Keynote
4:45 PM - 5:30 PM Booth Crawl Happy Hour (TBD)
Theme we expect to cover
The program is still being shaped, but the event will focus on the practical problems teams face when voice agents move into production.
Expected themes include:
Latency and real-time user experience
Turn-taking, interruptions, and barge-in
Orchestration and dialogue management
Evaluation, observability, and debugging
Human handoff and escalation
Compliance, governance, and trust
Architecture choices across speech, LLMs, and tooling
The focus will be on practical lessons from teams building and operating voice systems in the real world, including what worked, what broke, and what changed after launch.
What you’ll leave with
Practical architecture patterns for production voice agents
Better approaches to latency, turn-taking, interruptions, and barge-in
Evaluation ideas that go beyond demo performance and simple benchmarks
Lessons from real deployments, including failures, fixes, and tradeoffs
More concrete ways to think about orchestration, observability, and infrastructure choices
A better understanding of where handoff, compliance, and trust break down in production
Confirmed Speakers
Kaan Karakas (Engineer, Delphi),
Vyoma Gajjar (AI Architect, Service Now)
Brooke Hopkins (Founder, Coval)
Pete Warden, (CEO, Moonshine AI)
And more to be announced...
Imagine being in a full room of the best engineers, each one hyper focused on their craft.
Hands-On Workshops
Masterclasses led by experts building voice AI systems at scale.
Bring your laptop.
Leave with frameworks, code, and practical workflows you can implement immediately.
Networking & Community
250+ AI builders, engineers, founders, researchers, and operators. Great food. Meaningful conversations.
20+ Booths
Featuring teams that are actually building production-grade AI agents. Come chat with engineers directly, one-on-one. Ask questions and get the answers you really want.
Interested in a booth? See our prospectus here!
Or email sponsorships@aaif.io