The Crawl, Walk, Run Approach to AI Automation | Paul Kortman, Connex Digital

Adam (00:20) I'm not breaking any news when I say that AI is literally everywhere. It seems like every single process and document we dig into has an AI copilot of some sort attached to it, helping us optimize and create synergy and make us work smarter without working harder. But the reality is most leaders are paralyzed trying to figure out exactly where to start. There's this pressure to implement AI transformation or build agentic workflows and it can feel overwhelming. And that overwhelming feeling leads to inaction. That's why I was so excited to hear from Paul Cortman. He's the founder of connects digital and he has built out this framework all about crawl, walk, run. It's simply put, it's all about implementing AI and automation for your revenue organizations in a way that's actually sustainable and provides immediate impact without feeling overwhelming. Paul's advice cuts through the noise. Stop trying to eat the elephant in one bite. Start with the smallest pain point, automate that, and then build from there. Let's see what he has to say. Paul Kortman (01:37) Hey, Adam, thanks for giving me a call. I'm just returning your call, wanting to give a little bit more clarity into our crawl, walk, run approach. Basically, if you know anything about minimum viable product MVP in the startup world, it's very similar to that. of let's take something that's going to alleviate pain today. Say right now you've got somebody who's copying leads from Apollo and putting it into your CRM or into your email marketing tool. OK, well, that's like let's just start there when it comes to automation. Let's tackle the low hanging fruit as it were. And that's what we call kind of our crawl phase of like let's just start there. Do a little bit. Some people get really overwhelmed, even like with AI and like, how do I implement AI transformation and how do I do agentic work now? And it's like, start with the smallest thing, start small and then expand beyond that. And that's the whole concept of crawl, walk, run. I could get into all the details of each one, but right now we're doing this with agents for a lot of our clients of like, okay, Why would I use an agent versus like a make scenario or defined automation? And we start with the smallest, most simplest agent. I have an agent right now that files my call recordings for me. just puts it into a folder instead of going through the whole process of like, okay, now decide, do a sales coach on that, decide, did I do the right model, et cetera. That's going to be the next iteration of that agent. So instead of trying to eat the whole elephant in one bite, start small, start with a simple agent that does one thing and does it well, and then add onto it. Or maybe you divide that and and add a separate agent that does even more. So that's the whole concept with crawl, walk, run of just like start small, start simple. What's one thing today that you could automate that you could put on autopilot, whether it's through what we call deterministic, which is very, very well defined or through agentic, which is, you you let the agent figure it out. And there's a middle ground in there too, where you have mostly deterministic steps, but you have an AI prompt that says, hey, tell me whether this was a positive response or a negative response, and then take action based on that. So that's just kind of the real simple lay of the land of how to approach agents, AI, automation in 2026. Basically, Let's start with something. Paul Kortman (04:18) Don't try to map it all out. You're dealing with a whole lot of different features and functions and capabilities. so instead start with something small. The nice part in today's day and age is you're not building code. You're not building scripts. You're not having to think everything through because otherwise you wasted time and effort. Start with something today and build on that. That keeps it simple. It makes it a lot more fun of just like, can I get value and benefit out of an automation or an agent today? Well, there's this one thing that's annoying to me. Yeah, there's this huge thing that's annoying to me, but if I could get this one thing started, then we can work towards the end after that. there's a whole lot there. Feel free to call me back with any questions you have, but just kind of a basic framework of start small and even work on the deterministic scale of start with something you know, I want this done this way every time. And then you can start working your way into the agentic scale, which is having agents do stuff for you and hiring less employees. That's the kind of the name of the game here.