From Scripted to Smart: How to Systematically Humanize Your AI Sales Agent
Details
Over the past few years, we've focused on one core goal for our Sales AI agent: making interactions feel truly human. But what does “human” actually mean in the context of AI? For some, it’s about natural, fluent language. For others, it’s empathy, nuance, or the ability to understand intent—not just words.
In this session, we’ll take you behind the scenes of how we incorporated human reviewer feedback directly into our training loop.
We’ll share what worked well, what didn’t, and the surprising challenges we faced while trying to encode “humanness” into an AI system.
We’ll also dig into one of the hardest parts of the entire process: evaluation. How do you reliably test whether an AI response feels human? What metrics matter? How do you measure tone, empathy, or conversational flow?
Expect practical insights, real examples, and lessons learned you can apply to any AI product that aims to engage users in a more natural and human‑centric way.
Join us to explore the intersection of AI, human judgment, and the art of conversation.