In this episode, Aaron Lee, Founder & CEO of Smith.ai joins me to discuss a variety of topics. First, we delve into the innovative strategies of tech giants like Google, and explore the future of AI in business. Then we journey back to the early days at Smith.ai, where AI was used to enhance productivity. Aaron shares a multitude of learnings from his time founding RedBeacon and then serving as CTO of Home Depot following its acquisition of RedBeacon. Next, we dig into the complexities of teaching machines to understand context, tone and nuance, and the potential for AI to develop emotional intelligence. We also explore the 'human in the loop' model, where AI efficiency is combined with human reasoning for maximum impact. Finally, we get to know Aaron a bit better through our Quantum Round of rapid-fire questions.

Highlights
[1:15] - Aaron walks us through his professional background, particularly digging into how his time at Google impacted him
[2:27] - We speak high-level about the idea for Smith.ai
[3:00] - Diving into how founding his first company impacted Aaron’s founding of smith.ai and how his customer learnings informed smith.ai
[4:34] - Technical challenges in building RedBeacon and what interested Aaron about starting a marketplace company when he did. Aaron shares how necessary it is to balance supply and demand at marketplace companies.
[6:12] - Considering verticals as well as going deep vs. wide when building marketplace companies, and then understanding what core tech will enable the marketplace you’ve decided to create
[8:37] - Matching components as core values in marketplace companies and how the quality of those matches is key to being successful. The matching in the beginning requires going deep in one vertical to start before expanding further.
[10:27] - The story of Home Depot acquiring RedBeacon and Aaron’s time as CTO at Home Depot plus the lessons he took away
[12:50] - How scale impacts your perspective on how multiple audiences are interacting with your software and the need for ease-of-use when you have a massive number of people needing to use your software
[15:06] - The origin story of Smith.ai and the core problem they were trying to solve as well as where they saw opportunities
[17:36] - In the early days of smith.ai (before all of the advances in LLMs), exploring how far were the team was able to get in removing challenges for customers using AI as a copilot vs. an actual participant in calls
[19:14] - What AI excels at vs. what humans excel at when it comes to using AI in customer service applications
[20:12] - Sharing thoughts on AI’s ability to develop human-like emotions/EQ and if there’s specific architecture that can help solve that problem
[23:02] - Exploring the human-in-the-loop idea and the power and limitations of AI systems as well as how to pitch software vs. services companies or companies that combine the two
[28:40] - Quantum Round of rapid-fire questions

Thank you to everyone who made Data Council 2025 Oakland an incredibly experience! This year's talk videos will be uploaded to YouTube by end of May 2025.

Subscribe to our newsletter to get updates for once they're ready to view!

The end of Prigozhin

The Rachman Review

Introduction

The AI-powered Podcast Player