[Live Webinar] A Founder's Playbook for Testing Voice Agents

​How do Bluejay's customers like Google and DoorDash ship AI agents fast without breaking them? The answer is not just better models. It is a better test suite.

​Most teams building voice agents face the same painful bottleneck: testing.

​When Rohan and his co-founder were building their first voice agent for restaurants, every update meant half a day of calling the agent themselves and testing 50 to 70 scenarios manually. And that was only one part of the problem. Real customers speak with different accents, interrupt mid-sentence, and can call from noisy environments. And once an agent goes live, most teams have limited visibility into whether it is actually working.

​That experience led them to build Bluejay, a simulation, evaluation, and production observability platform for voice agents. Backed by Y Combinator and Peak XV, the platform has analyzed over 10 million minutes of conversations and run 72 million AI evaluations.

​In this session, we interview Rohan Vasishth, CEO and Co-founder of Bluejay, in a 30 minute live conversation followed by 15 minutes of audience Q&A.

​What we will cover:

• ​Why manual testing breaks down at scale and what rigorous testing actually looks like

• ​How to simulate real-world complexity: accents, interruptions, and edge cases

• ​Pre-deployment simulation vs production monitoring and why you need both

• ​How to build testing into your pipeline so regressions never reach your users

​Come with your hardest testing questions.