Ft Myers Beach - Good Neighbor

TECH TALK-From Buzzword To Blueprint: Choosing The Right AI For Your Business

• "Cabo" Jim Schaller • Season 5 • Episode 36

Share episode

Forget the sci‑fi headlines. We zero in on the kind of AI that actually ships value today and unpack how to choose it, pilot it, and prove ROI without blowing up your workflow. With Ken Johnson from USIT Systems, we sort hype from reality, cut through vendor noise, and focus on targeted wins that save time, reduce errors, and move the bottom line.

We start by clarifying the landscape: narrow AI that powers predictive text, chatbots, and smart thermostats is the workhorse of modern automation. Strong or self‑aware systems remain theory, while early “theory of mind” experiments are emerging in eldercare robotics. From there, we get practical. Ken walks through a simple framework: define the pain point, quantify cost versus value, choose whether your goal is direct monetization or better predictions, and run a small pilot before scaling. You’ll hear real cases, like a clinic enabling secure after‑hours bill pay and a dealership forecasting inventory and buyer segments to tune marketing and stocking.

We also tackle the pitfalls. Predictive systems can trap you in your past choices, narrowing research for tools and even healthcare providers. Vendor ecosystems push their own models, so we talk about comparing outputs, maintaining optionality, and avoiding lock‑in. Then we look at where AI meets robotics to make work safer and more consistent: fryer automation that reduces burns, robots handling hazardous tasks, and near‑term hospital helpers that free nurses to focus on human care. Through it all, the theme holds: AI augments people and process, but human judgment sets the guardrails.

If you’re ready to turn AI from a buzzword into a line on your P&L, this conversation gives you the blueprint. Subscribe for more practical tech strategy, share this with a teammate planning an automation pilot, and leave a review telling us the first task you’d automate.

USIT SYSTEMS
(407) 753-4499
Info@USITSystems.com
USITSystems.com

Ft Myers Beach-Good Neighbor
To Nominate Someone to be Featured
Or
To Subscribe to the Newsletter
CLICK HERE
https://cabowabojim.com/listen-to-podcasts/

Intro/Close:

Welcome to Tech Talk, your go-to guide for making technology work for you. Whether you're running a growing business or just trying to keep things running smoothly at home, we've got IT covered. US IT Systems brings the knowledge, experience, and solutions you need to stay connected, protected, and productive. It's Tech Talk Time. Here's your host, Cabo Jim.

Cabo Jim:

Welcome to Tech Talk again. We've got Ken Johnson from US IT Systems with us once again. And today we're going to talk a little bit about AI, right?

Ken Johnson:

Well, AI is the big buzzword. And after the last podcast, I put it out there, and that was the number one comment I got back was talk to us about AI and what a what AI is all about.

Cabo Jim:

So and there's there's a lot that AI can do, right?

Ken Johnson:

There's there's a lot it can do, and there's a lot that it can't do. And and so I want to clarify some of that, you know, uh as far as a lot of the misinformation out there. Yeah, so so that that's where I want to start.

Cabo Jim:

Okay, well, let's start off by I guess first addressing the types of AI that are out there, because there are a lot of different types of AI.

Ken Johnson:

Well, and and there's types by function, but let's talk about type by capability first. You've got you've got narrow AI, which is the the parts and pieces that um uh are there for specific functions, and then you've got uh much more intelligent AI or strong AI. And and currently that's still theoretical because because strong AI is is not implemented in in any way, shape, or form. The strong AI is where where you've got a machine making decisions, and and right now those decisions are still in human hands. The humans come along and they they say, all right, if this happens, then this happens. So basically, AI is nothing but a bunch of if-then-else statements. So if if you if this is true, then do this, or else go on to the next question. And so so that is what narrow AI is, and and narrow AI is is what we have today. And you see that with predictive text when you try to text somebody with your cell phone, and when you're you're doing that, the the narrow AI comes along and it it suggests what the next next thing is gonna be. Hey, every day at at uh you know two o'clock in the afternoon, you turn the temperature down two degrees. Would you like me to make that automatic for you? And so so that's that's narrow AI, and that's that's what we have right now. And then the the the next step is super intelligent AI, and superintelligent AI is is what everybody's afraid of, and and there's a lot of ethical and moral questions out there, but we don't have super intelligent AI out there. I I haven't uh I haven't seen Arnold Schwarzenegger walking down the street and going, I'd be back. But you know, it's it's the different levels there. So now if you'll humor me, we'll go on to to functionality and and let me grab my glasses so that I make sure that I talk intelligently. The by function, if we're just talking within narrow AI, which we have today, by function we have reactive AI, it responds to specific input, but lacks memory and learning capability. So uh a game that you have on your phone that you're playing against a computer, that's a reactive AI. Um uh words with friends when you when you do a a solo um game with with it, you're you're dealing with reactive AI because the AI is actually responding to your input and playing a game against you. Okay, limited memory AI learns from historical data. This is this is in past experience. This is your thermostat. This is self-driving cars, these are the chatbots that are on there, and and you can you can implement these chatbots in a business environment, you can implement these chatbots in in all kinds of environment, and and that limited memory AI, again, is a narrow AI, it's task specific. And then the next level it gets up is what's called theory of mind. Um, and and that's future AI models are expected to understand human emotions and intentions and interactions, but at this point, that's all still still very experimental. Um, a number of the Asian countries are out there and and doing that and implementing it because of the aging population, they're looking at AI robots basically that can understand the needs of the aging population and provide services for that aging population. So that that activity is is coming along quickly and and being implemented. And then the last piece of the puzzle is the self-aware AI, and that's still totally, totally theoretical.

Cabo Jim:

So as an individual, as a business owner, how do I decide where and how or what type of AI to use?

Ken Johnson:

Well, if you listen to the last podcast, I'm gonna tell you that I am not an expert in all these fields. Nobody is, but I've got access to uh little over 400 experts out there. So what you're gonna do is you're gonna decide what you're trying to achieve. You've got to define your pain point, you know, what it is in your business that you want to implement and automate, and and what those points are. Oh, I mean, everybody wants to call in at seven o'clock at night to to pay their bill because they got home from work and I've got a small medical office, and and I don't want to staff at at uh eight o'clock at night. I was talking to a doctor yesterday that I'm I'm working with, and and the the doctors told me that in in town there are five independent medical offices left. All the rest of them have been bought up and bought out and are part of a major hospital group, and that hospital group implements all the technology. And he says, How do I implement some of this technology to compete and stay independent? And and so that's that's what I did is I set a meeting with him to sit down and identify what he wants to automate, and that's that's the first piece you've got to do is decide what you want to automate. Okay, second piece of the puzzle is to quantify the cost versus the value, and that that value can be a complete return on investment where you you turn around, you say it's gonna cost X number of dollars to implement it, it's gonna save me X number of man hours a month or a year or a week or whatever, and that is gonna be a labor savings of such and such, or it's gonna reduce errors in data entry or errors in in implementation. So, what are you trying to fix? And then what's the cost versus the value? The next piece of it is is the hard piece, and that's that's to choose your model. Are you trying to directly monetize whatever you're implementing, or are you trying to just collect data for independent activity, like you know, sales information and predictive sales information? And AI is excellent at doing the predictive sales. Uh, I've got a car dealership that I've been working with over the last uh 18 months or so, and and that car dealership is looking at what cars are selling in new and used. They are also looking at who's buying in new and used, and they're looking at those demographics, and they're making predictions based on the demographics, so who they're gonna market and advertise to. They're also making predictions on what vehicles they need in both the new and used markets to have on the lot in order to get people to sell, and car buying has has changed. I mean, uh when you and I were kids, we went out and and we went to a car lot and we walked around and looked at what cars were there. Well, that's not the game anymore. Now everybody goes online and they look at what cars are available at what dealership, and then they decide what dealership to go to and put their hands on it. So the entire experience has totally changed, and AI can predict that experience and predict what people are doing and predict what you need to have as a retailer to get the job done.

Cabo Jim:

Interesting. So so now, as a you know, as a smaller business, for uh one of your examples, you can now more or less compete with some of the bigger people because you've got some of that technology, you've got some of that knowledge without having to go out and spend money on a whole new department, right?

Ken Johnson:

Correct. And and uh I'm gonna I'm gonna say anything, anything that you're implementing with AI, I'm gonna tell you start small, do a pilot operation and start very, very small. Do not try to implement this broad sweeping change throughout your operation. Start small. I want to process, I want I want to change my phone system up, I want to have it set up so that when people call in, that they can pay their bill directly over the phone. Boom, that's it. I don't want to do anything else, I just want to do that. Next step. Okay, what time of day do people pay their bill? Well, they pay their bill at 7 42 p.m. Um, 80% of the people pay their bill then. Okay, that's great. Now I want to send out my invoices at 6 45 at night to remind them that they got to pay their bill, and they'll go pay their bill in an hour. So you can start off with, I want to do one thing, then I want to implement another thing, and then I want to step up and step up. And and it is start small and move up very slowly. Do not try to throw everything to the walls.

Cabo Jim:

That's an that's great advice because, like you said, you know, with anything, you want to make sure it's working, you want to make sure it's doing what you want it to be doing, and then you know, you want to make sure it's helping your business at the end of the day.

Ken Johnson:

Absolutely, absolutely. So that that's the parts and pieces to to implementing it. So now, what questions have you got for me?

Cabo Jim:

Well, I I mean, I guess you know, there's a lot of different AI out there, you know. It seems like every program I have right now has an AI element attached to it, you know. Absolutely. Do I need to be using all that, or do I again do I'm picking and choosing where I'm using it?

Ken Johnson:

I I'm gonna tell you to pick and choose because yeah, everybody, every search engine has got some AI. Well, what is that AI doing? It's predictive AI. So that predictive AI is is directing you to where they want you to go, not necessarily where you want to go. So as you go through and you you engage that AI, understand that that engagement is predicting based on past experience. Now, I know you, and I know you very well. You and I have been interacting for what 15, 20 years now, and and so I know that you do not keep doing the same thing over and over, expecting different results. You constantly are changing, you're constantly looking at new ways to implement things, you're constantly looking to grow as a person and in technology. You've never been stagnant, and and that's that's really key. And a lot of this AI and predictive AI is pushing people to be stagnant and doing the same thing over and over again, expecting different results, which is the definition of insanity. So I don't like all the predictive AI. Is it good for my thermostat? Yeah, okay. I kind of like it predicting my thermostat. Is it good for me doing research on particular information and technologies? Not necessarily. Uh the AI model may know that I've looked at a particular phone system in the past. So when I ask the next question of uh which phone system uh that I sell uh does uh call center implementation, it's going to come back and recommend the specific one that I've looked at in the past instead of giving me all of my options. I don't like having options restricted. Uh the other piece that that uh is out there with the AI is is healthcare prediction. And I'm I I really, you know, when I'm looking for a doctor, um, I don't want to have the doctor recommended as only the doctor that I've seen in the past. I may be looking for a different situation, a different product, uh a different specialty. Uh, same with a dentist. I may be at my age, I'm looking for a dentist that specializes in implants instead of just doing standard fillings. But the AI knows that I talked to a doctor. Um I'm trying to think of the doctor's name from Little Shop of Horse. But uh I'm um I'm looked at that doctor in the past, it's going to recommend that doctor only. And that's those are the things that that bother me about about AI and and what you're implementing. So as you engage it, it can be helpful and useful, but it can also be restrictive.

Cabo Jim:

And it's only as smart as, you know, it's it's a learning tool. It's only as smart as the information somebody is putting into it and interacting with, you know, can it help uh you know come up with different ideas in certain situations, but it's only gonna it's limited in the fact that it's only the information that's uh been out there already or that they've found or limited memory AI learns from historical data and past experiences to inform decisions, and that's that's the key.

Ken Johnson:

That's what we've got right now. Theory of mind, it's still experimental and not there yet. Um, you know, I'm I'm I'm really glad that uh if I go in for surgery, that there's a human making the decision of whether that part needs to be removed or not, right?

Cabo Jim:

Yeah, yeah, I'll leave it up to a computer. Yeah, I don't want to leave it up to the computer. So it's and there's there's a lot of different AI in the fact that you know a lot of people offer it, but on the other hand, like you mentioned, it's uh there's competition, people want you to use their platform and their version of AI. They don't talk to each other, right?

Ken Johnson:

Uh they really don't. Um, it's it's called market differentiation, and uh, we've all experienced it for for many, many years. So um, yeah, it's not it, it's it's not uh universal where everybody's talking to everybody. You got perplexity, you got chat GBT, you got Groke, uh, that there's all these different pieces out there, and each of these companies want to sell uh their backbone, their program, their tools, their their machine learning. They want to sell that to um XYZ phone company, uh ABC, uh, data analytical company, um, all the rest of it. Now, I'm gonna I'm gonna tell you everybody knows Salesforce. Salesforce is the big monster in the room. There's a company uh in Europe that used um an AI set of tools. I'm not gonna promote one or another, to write a competitive program specific for their company and replaced their Salesforce program with this very targeted uh uh management system for managing their leads and and all of their technology. So that that is you know a huge, huge disruption to the industry being able to do something like that. Uh in Orlando, there's a gentleman I know who has written a marketing program in order to take and in used AI to write this program in order to take and do direct mail marketing, exceptionally targeted for some very specific industries, so that that industry can turn around and look at their database of past customers, predict what customers they should be selling to, go out on the internet, find those customers, find those contacts, find those email addresses, and turn around and send a direct mail campaign out to those existing customers, and another direct mail campaign to the prospective customers, and a third direct mail campaign to customers that haven't bought from them for 90 days, 180 days, 365 days. And so this was all done and written with AI with a very, very minimal uh financial cost to get it implemented. And and they just rolled it up oh about a month, six weeks ago.

unknown:

Yeah.

Cabo Jim:

So I mean, there's there's some good things that AI is doing, but you know, again, you know, I I guess I warn you, much like the media, they're gonna filter and they're gonna push people in the direction of what benefits them as a company, um, providing that AI, right?

Ken Johnson:

Absolutely, absolutely. So you've you've got to pick and choose, and you've gotta be aware. And uh I don't remember who the quote was, but uh of the many truths, if you pick one and follow it blindly, um, it will become a falsehood and you will become a fanatic. So there's a lot of truths out there. Look at all of them.

Cabo Jim:

Yep, look at all of them. And make your own decision. End of the day. There we go. Well, very good. Any last words for our listeners today?

Ken Johnson:

Um, yeah, if if you see Arnold Schwarzenegger walking down the street, uh, you know, with uh, you know, uh saying, I'll be Bach, uh, be afraid. But so far, we're we're not out there. Uh, we haven't got self-aware uh AI. Um everything out there is is still uh limited memory AI and and predictive AI. So it's uh theory of mind is is coming. Um the the uh Asian countries are way ahead of the US in implementing that, but it is it's coming, and I think that that's uh uh an area that is going to to absolutely mushroom in the future. Um when I I I I think nursing. Uh think nursing. Wouldn't you how would you feel about having a robot be able to deliver the food trays? Um, how would you feel about the robot being able to uh solve minor problems? Like uh I drop the call button from the side of the bed and have a robot come in and pick up the call button and hand it to you. Uh, I think that that's gonna be theory of mind uh implementation. Uh I think medical is huge in that area. Um but it's not there yet. Uh limited memory AI. Uh we've got um uh there's a company out there called Miso that in my experience as a young man, uh running the fryer is is the absolute worst job in any fast food restaurant. And Miso has come up with a AI machine that that cooks the French fries or or whatever else is in the deep fryer and removes the human element from that, which increases safety from burns, which increases efficiency and consistency of the product coming out, and and there's uh that that whole area is absolutely exploding.

Cabo Jim:

Wow, absolutely AI fry guy. There we go. We got a little bit of an AI fry guy that's called the jobs, does the jobs you don't want to do, right?

Ken Johnson:

That's the job you don't want to do, and there's a lot of jobs that that you don't want to do or that are dangerous, and and I I mean I I look at oil fields and and I look at the the the danger involved in that. Um I I I worked uh on uh airbag uh explosives, and and I I saw a guy in the powder room when a flash happened in the powder room and and the guy got burned. I mean putting uh a robot in the powder room that if the if the flash happens and and uh and and the robot gets hurt, you you put another one in there and send that one out for repair. So yeah, I think there's so many areas that are dangerous that AI can be implemented and and it would be great. I I just think that that would be huge.

Cabo Jim:

Very good.

Ken Johnson:

We can only hope, right? It's not even hope anymore, Jim. It's they're coming. Uh the it's out there now. Are they gonna be able to to um you know lay bricks? Maybe, but are they gonna be able to make human decisions? Uh maybe. So the the people out there that are are the ones that have to make decisions to make sure that the structure is complete, to make sure that everything is intact. Uh, those are all things that that those are jobs that are not gonna go to AI.

Cabo Jim:

Very good. Very good. Well, Ken, again, I appreciate your time today and your knowledge. It's been a pleasure getting to know a little bit more about AI as we continue to move forward. But uh, I guess we'll see you next time.

Ken Johnson:

Uh, I don't know if we have a topic pick for next time, but uh I'm sure and uh I'm gonna I'm gonna ask people to comment and and send in there what do you want to hear about? Uh and and we can pull in experts from other fields too, uh, you know in order to address some of this, but uh but I've I've had a couple of requests uh for another topic, but uh well let's let's see what the what the response is when I post up this video.

Cabo Jim:

Very good, very good. Well, Ken, again, uh pleasure. Thank you for your time today, and uh, we will see you out there soon. Always a pleasure, You. Take care now.

unknown:

Yeah.

Intro/Close:

That's it for this episode of Tech Talk, where we've got IT covered so you can keep life and business running smoothly. Remember, whether it's cybersecurity, cloud solutions, or just keeping your tech trouble free, USIT Systems is here to help. Visit us online at usitsystems.com or give us a call at 407-753-4499. Stay safe, stay connected, and we'll catch you on the next tech talk.