Own The AI Roadmap

AI has a weird way of failing: not with a loud crash, but with quiet drift. You start with a promising idea, then weeks go by as everyone keeps “making it better” without a clear finish line, a real workflow target, or a definition of success. We’re talking about the pattern we’re seeing across companies that actually get results from AI and what the best leaders do differently from day one.

We share a hands-on story from a small business with 20 years of history, loads of data, steady inbound demand, and processes that must be documented with care. The breakthrough isn’t a magic prompt or a trendy tool. It’s the owner stepping up like a program manager: guiding development, testing what’s built, and giving consistent feedback so the system matches the messy reality of how work gets done. We also explain why the classic “I want this tool, tell me when it’s ready” approach almost guarantees disappointment with AI implementation.

We dig into why custom AI tools are more achievable now, how to think about frontier models versus closed or proprietary models, and how to keep your voice at the center so you’re steering the ship. If you’re serious about AI strategy, AI transformation, and practical business automation, this is a mindset shift worth stealing. Subscribe, share this with a teammate, and leave a review with the workflow you most want AI to improve.

Enjoying the show? Subscribe, share this with a leader in your organization, and leave a review with the biggest AI question your team is wrestling with.

I'm Craig Francisco, CEO of AI23. I help privately held and family businesses turn their hard-earned domain intelligence into an advantage AI can't hand your competitor down the street. If your team is wrestling with where to start, the best place to begin is a conversation: www.craigfrancisco.com 

Transcript

Copy Transcript Copied

Why AI Efforts Drift

SPEAKER_00 0:12

Hello everyone. I wanted to take a minute to talk about a success that I'm seeing across a couple companies here recently that I believe are it's worth noting and worth others to consider as they go forward with their own AI journey. You see, there is there's so much power, as we all know, within AI that it's so easy to get sucked into the possibilities and get lost, continuing to just try to improve, improve, and make what you have better and better and better with no real clear guidelines or outcome goal you're trying to achieve. There is a client that we're working with

A Small Business AI Win

SPEAKER_00 0:55

at AI-23 right now that has a business, smaller business, been around 20 years, a lot of data, a lot of data, a lot of outreach, a lot of inbound, um, a lot of processes that need to be documented extremely well. Uh just you know, can be can be complex and overwhelming for sure. And so what this gentleman has done, the owner of this business that's that's been fascinating for me to watch and is really, really working, he's almost taken the lead as the the program manager, right? The the person that's super curious about what is possible with AI. So as we build out this kind of narrow and deep custom approach to leverage all of this incredible knowledge and documentation that he has to make this this whole you know step one through 15 as smooth as possible, he's guiding, he's guiding the development, which is so different than a lot of companies that are like, hey, I want this tool, I want it to do this, let me know when you're done and ready to deploy. Like that type of a mindset, I can almost guarantee you will fail. It's just gonna fail. It's that's just not how AI works.

The Owner As Program Manager

SPEAKER_00 2:16

So, what he's doing is the opposite of that. He's providing an incredible amount of feedback, an incredible amount of curiosity is going into his thoughts as he's working with what has been developed. And, you know, in the past, anything custom, as we all know, around software, uh, you know, systems, all of that was just a word that you didn't bring up because it was going to be very expensive. The developer wouldn't want to do it because they were gonna try to deploy the same thing to 10,000 other customers. But now,

Why Hands Off Deployments Fail

SPEAKER_00 2:50

what I would recommend all of you do is realize that you know, there so much is possible. I kind of hesitate a little bit, like anything's possible. I've so far we have seen just about that. I'm sure there's gonna be some limitations, but don't, you know, don't answer it for yourself. Push it, be curious, ask the questions, make sure whoever you're working with has the ability to take that knowledge and that information in your request and work with that. See if the AI model, whatever frontier model you're using, whether it's your own proprietary model, um, a closed model, whatever that looks like,

Custom AI Is Finally Practical

SPEAKER_00 3:32

make sure that you are being heard, that your voice is heard, that you are steering the ship. It doesn't take a ton of time. And remember, this is your business, this is your future. So if the AI model, if the technology can do amazing things and what you want, why not ask for it? Why not push for it? Because your business, if it can transform and you then can amplify what you're good at and what your team's good at to make the business even better, whether that you're trying to grow the business, improve your quality, improve your service, whatever that looks like, go for it. Make sure your voice is heard. So that's that's the message for today's podcast. Make sure your your voice is heard, do what you need to do, spend the time, think, be curious.