Behind the feature: building in the age of AI
Every feature we ship starts with creating better experiences for our customers. The 750k developers running voice agents in production surface the problems, and our job is to listen, decide which signals matter most, and build. Teams deploying voice AI in production surface these problems. They need to know an agent works before it reaches a real caller. They need outbound that holds up at 10,000 contacts. And they need to pick the right models from a catalog that changes weekly.
This session goes behind three of our newest features: Simulations, AI testers that validate your agent before a real caller finds the bug; Campaigns, batch outbound rebuilt for enterprise-scale calling; and Model Intelligence, curated presets and performance metrics measured on live Vapi calls. We'll start with a short presentation on what each feature does and the problem it solves, then run a live demo of all three. From there, it's a live Q&A panel with the three engineers most pivotal in taking each one from an idea to a feature customers use every day, covering how they thought about building them, what worked, what didn't, and what they learned along the way.
What you'll learn:
How Vapi turns customer signal into a feature roadmap
What it means to build AI features when the underlying models keep improving
What Simulations, Campaigns, and Model Intelligence do, with a live demo of each
What worked, what didn't, and what the engineers would do differently
Who should attend:
Teams running voice agents on Vapi who want depth on the newest features
Engineers and PMs building their own AI features who want to compare notes
Product leaders who want to see how customer feedback becomes shipped product
Anyone evaluating whether these features fit their production workflows