AI for Small Business: Automate, Audit, Adapt | Bryan Barnhill | Weekly Blitz
AGI - Advance, Grow, Innovate with AI
Who’s talking
2 speakers1Transcript79 turns · 39 min
Hey, welcome back to another edition of the AGI Podcast Weekly Blitz. I have a couple surprises for you today. One, we are not going to talk about any of this bullshit about jailbreaks or existential threats. like every other podcast has been for the last two weeks. And two, Julie's not here. I have a co-host, Brian Barnhill. He's a new friend of mine. We've met through some of these in-person AI meetings, so I'm going to let him introduce himself, and then we're going to talk to you about what AI can do for you, not how it's going to end the world. So with that, Brian, go to 11.
Code 11. Thanks, Jason. I'm glad to be here and get my podcast hosting chops up to par. Looking forward to a little time with you and talking about the stuff that we seem to be involving ourselves in more and more every day. My background, I'm a retired healthcare IT, mostly IT, spent the last several years in the healthcare space. But I'm a techie at heart. So when AI started coming around a few years ago, I didn't have as much time to delve into it as I had hoped or I would have liked to at the time. So now I've got some time and I'm getting my feet wet and learning things and trying to catch up with some of the smarter folks.
Yeah, I don't know about smarter, but more exploratory. That is a very interesting dynamic that I find with AI. Like, if you look at Silicon Valley, which I do watch them a lot, right? I think that their definition of the ideal employee today is kind of part project manager, part designer, part developer, right? And I would say in that culture, I can see that. In the Midwest and in, like, small to medium-sized business, I think it's part... domain expert that understands the business case of whatever you're trying to do, part beginner's mindset, like exploratory, just like the world could be our oyster, and then part computer science background. Because I do think that there's still a ceiling. That ceiling seems to be rising all the time with just how much you can get done without having a computer science background. But I still, as a person who doesn't have that, see where my limitations end. And people who are developers or coders or have a tech background, they just understand some things that would make my life a lot easier if I had that knowledge. Does that make sense to you? Do you feel that way?
I think it does, but I look at it also the other direction is we're constrained. I find myself wanting to give too much. I want to create parameters. I want to create processes. That's not what these LLMs are geared toward. They're geared toward taking Jason without that background saying, I want to do this. How do we get there?
yeah that is so true that is where i'm at right now that that is actually some of the debates that i'm having with people like this idea of harnessing because i think the foundation model companies that we just saw muse and grokbot come out with like these you know give it access to everything and it can do your email it can be like it can do a whole lot it is that is because it's pre-harnessed really well for a lay person to be able to just tell it what they want to get done and it goes and does it. And I think that some
2Episode notes
AI should take work off your plate. So why does choosing models, checking results and keeping up sometimes feel like another job?
Jason Padgett is joined by guest co-host Bryan Barnhill for a practical Weekly Blitz conversation about small-business AI, useful automation and the work of checking what your agents actually did.
Bryan is a retired IT professional whose later career included healthcare. He describes his interest in helping owners think about AI before a business transition, building an app that lets customers buy friends a drink for later collection, and learning from a multi-agent fantasy-football project that did not generate sales.
WHAT YOU’LL TAKE AWAY
Compare model quality with the quality of your prompts and feedback.
Give every automation a way to be checked and updated.
Start with the customer’s problem before sending agents to fix things.
Treat an unsuccessful product experiment as evidence you can learn from.
TRY THIS
Choose one recurring task. Define the useful result, decide how you will check it and put a review date on the calendar.
CONNECT WITH BRYAN BARNHILL
Bryan is based in Indianapolis. He is exploring AI uses for local small businesses and building a drink-gifting app.
https://www.linkedin.com/in/bryan-barnhill-4698b16/
RESOURCES MENTIONED
Zen Mind, Beginner’s Mind. Jason discusses its beginner’s-mind idea during the