How Bethany Crystal Founder CEO of Build First Built an AI-Powered Marketing Operating System

Introduction to AI-Powered Marketing Operating Systems

(0:00) If you're wondering what it takes to build a marketing operating system with AI from the ground up, you're in the right place. Today, we have Bethany Crystal, CEO and founder of Build First. She's going to share what she's been building and how she's been building it. It is fascinating. You're going to get a lot out of this. Buckle up. You're in for a good one.

Welcoming Bethany Crystal

(0:19) Bethany, thank you much for joining us here today.

Bethany's Introduction and Build First Overview

(0:22) Hey, happy to be here. Thanks for having me. (0:24) Can you tell us a little about yourself, about Build First, and what you're up to these days? (0:30) I'm Bethany Crystal, founder and CEO of Build First. We build what we call AI-powered marketing operating systems. We help companies build their own custom internal systems to manage their marketing efforts. You can think of it like a smart marketing brain. We combine tools like Airtable, Make, Webflow, and large language models like Claude, ChatGPT, to build something unique and completely custom for our clients. This allows them to automate anywhere from 40 to 80% of repetitive marketing tasks. We can also train the LLMs on their specific brand voice and internal data so they can turn an LLM into an intelligent teammate. We've built everything from full social media operating systems, to a system for a large-scale product launch for one of our largest clients, which was a very intensive and complex project. That was fascinating. I'm also proud to say that we built our entire business, the Build First business, using the same AI marketing operating system that we build for our clients. We dog food our own product. We've been operating this way for about a year and a half now, even before the big explosion of AI. It's something we've been practicing for a long time. Prior to Build First, I worked for about 15 years as a director of marketing at various startups, tech companies. I worked at a couple of agencies, one of which was a boutique agency that served venture-backed startups. I've got a lot of experience in the marketing world. I'm excited to bring this new technology to marketers.

Building Before GPT-4

(2:49) Wow. That's incredible. A year and a half ago, this would be before ChatGPT-4. You were still building with ChatGPT-3.5, or what were you building with back then? (3:06) We were building with the models that were available at the time. OpenAI, Google, there were models out there. The bigger thing at the time for us was coding AI. We were focused on using coding AI to build the systems. That was how we started our journey. When the bigger LLMs exploded, we were able to bring those into the system. That's what we started to build out for our clients.

Defining an AI Operating System vs. Chatbot

(3:40) Wow. Tell us a little about how you define an operating system in this context. How does that differ from using a chatbot? (3:51) That's a great question because I think a lot of people think of AI as just a chatbot. For us, an operating system is much more than that. A chatbot is an input and an output. You put something into it, it gives you something back. An operating system is an entire system of tools that are talking to each other, they're integrated. That's where you get the power of automating those repetitive marketing tasks. For us, that looks like Airtable. We use Airtable as the database. Then we use a tool like Make, which is a visual automation builder to integrate all the different pieces together. Then we use the LLMs, which are the smart brain that can connect everything together and automate all the different parts. It's the sum of all the different parts. It's not just an input-output model, it's an entire system of tools that are integrated and talking to each other.

The Iceberg Analogy for AI Operating Systems

(4:54) Okay. It's what happens if marketers stop thinking about it as a chatbot and start thinking about it like a buildable operating system? That's what you're getting at, correct? (5:03) Yeah. I like to think of it like an iceberg. Most people, when they think of AI, are thinking about the chatbot, which is the tip of the iceberg. The operating system is everything underneath the surface. It's all the pieces working together, all the integrations, all the data sources that are connected, the LLM itself, which is the engine, the brain, and how you're training that LLM. It's all of those pieces working together that makes it an operating system.