Data + AI is a powerful combination resulting in real-world applications. Thomas Helfrich, CEO of instarel.ai, joins the podcast to chat on this & explore where AI is headed.
In our rush to create volumes of data, we have become awash in it to the point of madness. But, when you have the mindset and goals to leverage “Data + AI”, then you have a powerful combination.
Additionally, data + AI can accelerate insights and decision-making that would have been impossible in the past. The efforts in the realm of AI are now focused on accelerating, augmenting, and collaborating with humans. So, what’s on the horizon?
Joining me on this episode to discuss this is Thomas “AI Nerd” Helfrich. He is the CEO of instarel.ai, an advisor to the SwissCognitive, content contributor to The AI Journal and Entrepreneur Media, and an official member of the Forbes Technology Council.
01:31 – AI has significantly grown, as there has been rapid adoption of AI across companies in several industries. However, there are still some who don’t see its full potential. So, has an AI overload caused people to be overwhelmed and not fully understand this technology?
04:44 – There’s a misconception that AI is going to replace people. But the human element is still critical to AI success. Rather than taking jobs, AI will create exponentially more opportunities and high-paying jobs.
06:45 – There’s still inherent bias, as humans are the ones writing the algorithms. Often, it’s just a bias that appears because something wasn’t accounted for. It’s not intentional, but it requires humans to correct it.
08:45 – In order to build trust in the system, you must be able to explain why decisions are being made based on the outcomes of that system or algorithm.
09:28 – When patterns of bias arise on both sides of your model, you can refigure the model to eliminate areas of unfairness if you have the transparency to make revisions. However, there’s a greater issue if the unfair model is hyper-profitable or accomplishes the business’ goal. While there needs to be transparency, there will also still be a gray area in how to address perceived bias.
11:52 – Although machine learning is intended to improve the customer experience, it can also hurt the customer experience. How can there be a counterbalance to using machine learning?
Should we be doing AI? That’s not our core business? You need to have it as a core business strategy to have a chance, and not become the next Kmart or Blockbuster.
16:48 – Modern-day AI systems need to be built for purpose. If you’re going to use these technologies to enhance the customer experience, they need to be equipped to process information that will enable them to help customers.
18:11 – AI helps with initial engagement, but does it help improve the business process? Does it provide data that suggests ways to improve the system?
20:09 – Small and mid-market companies may have to navigate AI tools and leverage them in different ways, as a higher level of AI may seem out of reach for them. Fortunately, as AI continues to develop, it will become more accessible and easier to use.
22:05 – Along with the rise of cloud technology, there has also been a recent rise with AI clouds being purpose-built for certain industries. It has shifted more to customization versus just building a system.
25:30 – Intelligent automation, AI, and other advanced technologies, nowadays, have to be part of your strategy, as it enables businesses to make informed decisions faster.
27:26 – When it comes to expectation versus reality with AI, realistic components are that it’s built for purpose, looks at big picture items, and integrates well within your system. Additionally, AI is a tool to be intentional with and should be a part of your core foundational strategies.
Aaron Back 0:03
Well, hey, Thomas, thank you so much for joining the show today. It's pleasure to be chatting with you and unpacking a bit more about AI and as your kind of, I guess, your moniker of yourself as the AI nerd, but I was wondering if you could just give us a brief introduction of yourself. And before we dive into discussion today. Hi,
Thomas Helfrich 0:23
I'm Tom. know, hey, Thomas. Thank you, by the way, a Ron B.
Aaron Back 0:30
Yeah, I'm not gonna let you down. I ain't gonna don't mess up anything.
Thomas Helfrich 0:35
No, you don't? Um, no, you know, it's nice to meet you. I'm a little Atlanta, Georgia. You know, I wear a few hats as a YouTube host on AI nerd. And, you know, CEO and founder of install.ai and Intelligent Automation expert and I played racquetball once my life. I'm a Scorpio, and I smell nice.
Aaron Back 0:56
Which is great. Well, it's good. I mean, that helps helps shower this week for you connect. Yeah, connecting people there. And I I did play racquetball a long, long time ago. I did that for decades. But yeah, I had to.
Thomas Helfrich 1:08
It was not fun back then. Yeah. But actually a lot of people who played racquetball or are technologists just to be fair, I mean, you can't get much nerdier than racquetball. It's great.
Aaron Back 1:16
Yeah. Ah, yeah. But that was Yeah. Again, decades ago. Wow. I think it was my gains early. No, no. Yeah. So yeah, late teens, early 20s. I'm talking so been quite a while, but maybe I'll pick it up again. All right. Well, hey, AI. I know, it's a big, big topic, something you're passionate about. And but I think for a lot of folks, it's, it harkens back to I think, when I started remember the funny news interviews people have when they still couldn't figure out the Internet back in the 90s. And what is this internet thing? And then the cloud comes along, people are like, Okay, what's, what's cloud? What's that mean? I think for some still, it's, unfortunately, some are still thinking AI, that semi, you know, AI came along. So a great growth, a lot of companies are really taking advantage of it in great ways, seeing a lot of great potentials coming out of what they're doing with AI, but others still think of it's this kind of this nebulous thing. Misunderstood. Myths versus reality, if you will. But I want to circle start off on this topic here of, obviously, with this huge rise of AI. Are people seeing a getting sort of an AI overload? Like, is it too much information? And then it's causing people not to know what to do with it? Really? It's like, analysis paralysis, if you will? That I can't understand it. So therefore, I don't know what to do with it? Or is it? They just don't know where to start? I mean, so they are overloaded? I mean, what's your what's your thoughts there?
Thomas Helfrich 2:55
There is a I overload, and you touched on a number of a number of areas, and I know, we only have six hours to do the podcast. So Ivan people got popcorn. The first of all, what is AI, it's artificial intelligence. That's piece if you could think of it maybe more in the terms of augmented intelligence or Accelerated intelligence, only for the fact that it's, it's, it's a brand marketing thing more than anything, now that we have AI, we this and that, and, and as a, as a founder of a company that uses AI, we use it not just a little bit enough to really, but the part that we really use it on saves 90% of the time. And so the, how you leverage a technology like this is important, because it's leveraged in so many ways, and has so many different meanings, it creates, in me another turn of cloud of uncertainty. And that then creates the the fact that people don't know what to do. So when you don't have to do you freeze, and then you don't know how to apply it, you can get tricked, and when you feel like you're getting hoodwinked or tricked, then you just don't do anything. And so you'll see this over and over with companies that are just trying to figure out how to go leverage this technology, or these sets of technologies. And they they're kind of paralyzed, and they're also mesmerized by what it could do with being told to do I think just high level that's that's where I would start is that it could be a lot of things. It's all those things you said is actually cloud internet and others and compute. A lot of this is resting on machine learning algorithms that were written in the, you know, 60s, so it's been around a long time. And and I think we can take the conversation wherever you'd like to from there, but like there's it means a lot of things to a lot of people and for a reason.
Aaron Back 4:44
What some of the uncertainty is with that misconception that it's going to sort of take the place of people in a lot of ways. Oh, the
Thomas Helfrich 4:55
robots are coming right. Yeah, robots are
Aaron Back 4:57
coming. It's it's it's like why Do I want to trust it? How much do I want to invest in it? And is it gonna cause a massive shortage of people wanting to work for us, you know, type of thing or take my job and but I'd love to highlight the fact that there's still an element that needs is still critical to AI success as a human in the loop element?
Thomas Helfrich 5:17
Well, no, it's, yeah, and I'm gonna wrap it sorry, but he did pass, it absolutely is going to be for a very long time, I would have no fear of AI taking jobs, you know, there will be, you know, jobs that will come and go. But AI and these technologies will add an exponential level, increase the number of opportunities and high paying jobs out there in the world. You know, you've seen this and other things, right, when you go from people who bred horses, and trained them and fed them and cleaned them to kill those jobs in a way, but then you have people making cars, servicing cars, and designing cars and all the things. So in that way, there's way more cars and horses somehow. Well, and that's like a very fundamental example, like, you know, there's people who make candles a lot, but then we got electricity and wireless. Yeah, we got you. And so maybe an older example, and in a bit, but a very clear one that detaches you from modern day to so you can understand that that impacts always going to be the way that it is. And even if technology goes completely, you know, snafu on us. There'll be a whole bunch of stuff, we got to clean up, it's like crazy to
Aaron Back 6:22
say I had a recent chat with Peter Lee. He's the CEO over over at rapid miner and he had it you mentioned the candle there. And the way I think he phrases that continued innovation of the candle didn't result in the electric bulb, you know, so it had to have somebody thinking totally different. Instead of let's reinvent the candle, let's think totally different. So in this aspect of AI, you know, you know, everybody's got uses all other technology and having sort of a control over what happens. It's like, well, wait a minute AI is, am I going to lose control? What's going on? Is it kind of overrun, my work isn't gonna, you know, so that leads to that sort of building of trust with AI, you know, it? Yes, I still think it needs some work. Because there's still that inherent bias that sometimes falls in there, because humans are still writing the algorithms and things like that, or it's, it's just a bias that appears that something this wasn't accounted for. So it spits out some sort of false bias. In other words, it wasn't intentional, I should say, but it has, by bias. All the same. I had an example. I was chatting with someone previously, where they were using AI to help somebody, oh, I know what it was maximize their, their hiring pool for potential candidates for jobs that are trying to fill. And they were using AI to source that and sift through a lot of the resumes and check a lot of data. And it was unfortunately filtering out single mothers, because of some of the criteria, because they have more. They don't have standardized sort of hours that they could work sometimes, you know, they have to leave to take care of their children or what have you. So they adjusted the AI model to account for some of that, and it changed the output. And allowed for that now. And it wasn't an intentional bias towards single mothers that they were trying to build it to that. But it was an uncertain. It was a an unfortunate, I should say output of it. But it took a human to recognize something was not quite right here, to go back and say, oh, we need to adjust something. So that that's where I'm I mean, it's tying that human in the loop back to that continued building of trust inside the AI models and the outputs of it.
Thomas Helfrich 8:45
Well, absolutely, you're describing, you know, explainability traceability and you know, without adding noise or bias to a decision making process in those types of systems. And I'm not an expert in delivering that. But I do know that it to build trust anything, you have to explain the answers. And you have to you have to be able to mathematically maybe the wrong word, but like algorithmically show why a decision was made. And based on the data you have. Now in that example, it would have been interesting, and I think the models of the future will show here is in a moment many of them do this, while they're here, you're based on the logic you gave me here are the results, right? This is fundamentally enough. And all these other ones were excluded. But here are some patterns with these other ones that may show bias. And so the the the side that wasn't selected to come forth, let's say in the side that was you be able to show biases on both sides and be able to refigure or recalculate or refactor. Let's say your model two, oh, well, that wouldn't really be you have unfairness now and into this. Now. However, that happens, but if you have the transparency to do that, and then revise models, that the real problem comes though is if if the unfair models, the one that's hyper profitable, or the one that really accomplishes the business goal And then that's where the really big gray area gets of the A, most companies aren't going to disclose that unfairness unless they're caught. And be, you know, who would know, right? If you never see it so so there's no and so that that's a problem inherent that we want to make money and we want to serve the customer base and our shareholders and whomever else. But at the same time, you can't leave proliferates a problem, you know, out there in the in the world of bias. As a mother, I'm already challenged non challenge even further technology.
Aaron Back 10:32
And then sometimes too, it's there's a perceived unfairness, I think, cuz sometimes you have to build in certain parameters, because there's a specific set of, say tasks or things that are necessary to accomplish a this job that you're trying to fill on. If they don't meet certain criterias, then well, you know, you know, they're, they're not going to be hired. But somebody say, Well, hey, that's your That's unfair. I know. So so that could fit that role? Well, it depends, you know, so it does harken back a little bit to that gray area. But I think there still has to be some transparency in some of this. You know, that's that address? Yeah,
Thomas Helfrich 11:10
yeah. It's intentional or harmful, and sometimes that it will. And sometimes it doesn't matter, though, if it is, if it is bias or not, it just, I think the gray areas comes down to greed and, and function. I mean, there's a lot of routes you can go with this of, you know, there's, there's a million ways we could go talk about these but the truth? Yeah, in short, I think you need transparency. And I think the build trust, the step forward is, is build that from the beginning, it revalidate the responsibility of the trust in the factory in the explainability, at least, then you're in the right step to be aligned to be better than what others are doing or being or not showing the those elements to the the AI systems that are built.
Aaron Back 11:52
So I know we've hit on some sort of higher level concepts here on this, but I'm going to take a step down, but lower. And now I've seen many uses of machine learning out there, and many examples and, and so forth of and people packaged it up nicely and sold it as sort of componentized, or democratized usage of AI and ML. But a lot of goals I've seen with machine learning is trying to improve that customer experience. Not just retail, but any customer experience. And I've seen it with some success. But there are times though, when people really want to talk to a live person, they don't want the chat bot, they don't want this set or any other and it's just like, how do I circumvent this or get around it? Or what do I do? 0000? Exactly like the old day. Yeah, pound Hey, Ryan here, oh, yeah, somebody's automated, you know, call centers now and routings, I get it, I get why they do that. But at the same time, and the the rush to fill a gap, or say, Hey, we got this nice, shiny new toy. Let's use it to the enth degree. And the same token, though, you're hurting your experience with your customers. And that translates in the it could hurt your brand, it could hurt your sales and your bottom line. So what are you seeing that is kind of that counterbalance to that? I mean, there's Is there such a thing as a counterbalance to the machine learning? And Oh, absolutely.
Thomas Helfrich 13:26
Yeah, absolutely. So you're you're spot on on something with with that, as with any new mention earlier was when the bring the human loop. And so this technology, if you take the point of view that I have is you accelerate humans, right, or you augment them, because as long as humans are involved there, they're going to be involved up. So those I don't like IVRS, but I'll say more dynamic chatbots, that can actually solve problems that don't need to interact with awesome, but when I really have a question or a detail, or what's the difference between these two things, I can't figure it out based on data. Knowing that get a human that has knowledge is a big key, not one that's gonna need to then therefore transmit yet to somebody else. That is a really powerful use of technology, because then you really got the customer experience you want, which is, I couldn't figure it out. But I thought I could there's some stuff I certainly can online or on a chat. But then when I had needed a person they were so they were all over it. They knew exactly what I needed. And then they were empowered to give me a deal do this do that, you know, and are they already knew what my problem was when I came in fully. And I have to do all this that is so awesome of an experience. That's where technology like this applies, and this is how it should be applied as a human loop of the right moment. And then no more and then that way they the human gets a better life because they actually don't get people. Let me tell you, I'm coming from IVR already have cussed out the IVR because you just know someone's on the other in the line that heard that like, Oh, hey, sir. How are you doing today? Mike? Thank you for asking. In eight minutes.
Aaron Back 14:52
I actually had an experience with this just a few days ago actually. I was on a doing a chat. Got through the chat bot, and they get transferred me to actual person through the chat though. It was actually an Amazon chat issue. And they, I was just trying to return something and it wouldn't spit out the return QR code that's typically could do because there's some local store says that it handle Amazon returns. And so I just want to send something back. Well, they were asking these bizarre questions like, Look, all I needs this, kick off this process for me, send me the code, so I can be on my way. But they were going around and around. It's like, wait, I already told you this stuff. Or, you know, so I was just getting very frustrated. So finally, while they were taking forever to answer my question, I, I did some searching, digging further, finally found the right link, generated my own QR code, and got what I needed. And then I went back to them, like, Hey, thanks, I already got what I needed by so by Yeah, yeah, wasting my time. And then. But at the same token, the customer experience can only happen like you outlined, if the person on the other end, say the customer service, customer support, whatever you call it side has access to that information like that. They know the customer they're talking to they have their order history, or can even see their sentiment from past purchases, things like that, and can engage better. And so like you said, they can jump on it like that offer something a discount, refund, whatever. So it takes it's on both sides of the coin. It's that customer experience of engaging through whatever channel and then on the other end of how that support happens. So it's they're equally need to be balanced on that.
Thomas Helfrich 16:48
On exactly in to build that AI system to do that you'll never judge the machine learning the learning piece, right? And somewhat how human interacts and the you know, is that good or bad thumbs up thumbs down the ability to go across disparate systems with say, robotic process automation, or the ability to look at forms data with intelligent document processing, or complex business decisioning through an AI inference engine, if you will. And if you combine all those and put them on a kind of a platform approach, that the user interface is chat, or it's an email, or it's a call, but it's all centralized off have the ability to answer questions. That's what I think modern AI system is. From the work though, like you described in chat, it's very important that they're narrow and deep. They're built for purpose. There is no Skynet right now that we eat, if it was, you wouldn't be aware of it. But the point is, you know, there was it wouldn't be use for you helping you with an Amazon order. Amazon orders are built for purpose. They're built for all those pieces. And that's where the technology gets really, really, really well as when it's industry specific. And it's really built for a specific specific business function or set of functions, then it can, it's, they're fantastic. And they've come a long way. And and that's where your data is important. And what your business does, and how the outcomes are in the processes that support it are all really important. But enabling all that together is a modern day, what I call AI,
Aaron Back 18:11
art. Let me take this a step further is so obviously AI is helping for that initial engagement, and then that I'll call it sort of that transactional triage, almost Yes, yeah. But after that ends, are you seeing AI help with improving of the business process? So you learn from what happened during that? Transactional triage? What have you? And is it going back and suggesting ways to improve processes or improve systems? To say, suggest things that could be updated or changed? Or, you know, ask the right questions, the next time? Are you seeing that happen? Or is it just people aren't Fast, fast enough adopting, or taking the suggestions for real or not?
Thomas Helfrich 19:02
Yes, that's definitely happening, that the most advanced companies in the world are doing that that's, they take the interactions and you learn from it. Therefore, the interaction next time is not only more consistent, it's likely more accurate or to meet the need. Not all of them do that, though. Because you have to get to build for that you have to account for that in your data. And you have to account for that in the models. But there's some that are least collecting the data, maybe they're just not acting on it right now. So you know, the train these models takes a lot of data and a lot of this right now is just being collected of how do you learn from it? Where does a human you know, you need the feedback loop for learning piece, it can do it, you know, on the machine level as well. But the human really gives the best responses to some degree when it's human human on the machine level like you think trading platforms, right that like just the outcome, did you make money or not? Right? That helps determine if the the algorithm is running correctly. So it depends on what it is and in that thinking, you can learn but but that also means that can go haywire if it learns and then it doesn't know what to do. But then that's where a human comes in Lupica the answer question in very long winded form. Yes, they are being done that way. But not everybody.
Aaron Back 20:09
Yeah. Well, let me you hinted on something there. And your response about the large companies are doing that is, is I still kind of out of reach for the small and mid mid market? Or is there sort of easy ways that they can jump in and start leveraging these tools? I know some come with some solutions. Cloud solutions come with pre built AI components that they can layer in drag and drop components, even, like in low code, no code tool situations, there's AI components that can be layered into some of the functionality, which I think are easy steps for them to adopt. But for some I'm hearing that it's they feel like aI still kind of out of reach to the level what we just discussed. Is there ways that they can continue on their AI journey without feeling like, I can only go this far. And I can't go any further?
Thomas Helfrich 21:00
It's a great question. And one that I post to, I've posed to many companies as well, as you know, there's a lot go back to your original thing. I think you said the first sentence was the internet, the cloud will because it's on the cloud now there and there's been such low code, no code development, and the computing power has gone up so much, gotten cheaper? Yeah, it's readily available to some level on a number of business functions from document processing, to chat to just, you know, predictive models for your business and what have you. So yeah, the midsize business can do it. If they want to get into more of the data science really number crunching product side of it, then you're you know, you're you're, you're in the game of the great right, great resignation, and all the rising salaries for technology. And so yeah, you probably can't afford it, let alone you know, do it, but maybe the consume it's already been built that's becoming more and more available that democratization and adoptability. Will, that will definitely be a trend moving forward as it'll just become easier and easier to use. Right? Therefore, you're gonna hire different people. So the question and the answer is yes. And yah,
Aaron Back 22:05
yah, yah, yah sant. And, and to your point of the rise of the cloud and things, you know, previously with the cloud coming on, into being it was, it was super broad, and people are trying to figure out how to harness it. And we've seen over the last few years, the rise of these industry clouds with more purpose built specific built with sort of language processes workloads built inside, towards those industries. I'm also seeing the rise more and more of these AI clouds that have that are designed for certain industries that I think help that next step for some of these midsize is like, here's a solution already purpose built for the industry. And then I can leverage that
Thomas Helfrich 22:45
you use low code, quitting click and drag components to make it your customized versus build. And anymore. I mean, I said it is dead. And I mean, what I mean is most companies that I've seen are unless you're, they're going to cloud, you're running software. So the idea you know, security's ran, it's to some degree, office 365, even saps and cloud, right, like you, you it's it's getting away, you know, and a big question comes up, that comes up all the time is the core, is this a core? Should you do this as part of your core business? Right? So should we be doing AI? That's not our core business? We're a manufacturer, we're a pharma company. We're, the answer is I think, yeah, it is. You need to have it as a course business strategy to have a chance to you know, not become the next Kmart. Right. So they are blockbuster, which I missed. There was like, a Friday Night Live date you go.
Aaron Back 23:40
I do back with like, they haven't they have I think blockbuster I think Netflix even approached blockbuster to see if they want to purchase them. And they laughed them off. And now what's funny side anecdote here is that I watched the documentary about the very last blockbuster on Netflix, a few weeks back, so I guess Yeah, well,
Thomas Helfrich 24:02
I will take a tangent with you on this. I think Netflix does a great job of in the Spotify, eyes of the world and what they built for blockchain and things. But I will say that I remember being a Netflix mail subscriber and how cool. And then I felt like it died so fast and then went to box for like six months. I'm like, why? Like, you don't stream and I'm like, I know what they meant for spoiling this like three years ago, five years ago. I'm like, yeah, no, I just don't really. And now I don't even have I don't think I have a DVD player anywhere in the house. Right. And except the car on either. Yeah, the 2013 Honda Odyssey. Well, though I love my Honda Odyssey it needs
Aaron Back 24:39
the last thing I have with a DVD player and it was my prior Xbox console. And then when we upgraded the new one, I got the fully digital one. We don't have to slide right here.
Thomas Helfrich 24:50
I'm going Aurora 14 Man 128 gig 3200 megahertz ram with a 59 50x an AMD 16 core with a 3090 Nvidia. Yeah, it's gonna be The 18,000 gaining points, so it's gonna be off the charts. There you
Aaron Back 25:03
go. Yeah. We'll have to you'll have to do a YouTube special on there.
Thomas Helfrich 25:07
That is, I mean, I think it's straight out nerded out. That is.
Aaron Back 25:11
Do it do a full on. Yeah. Overview of your system. Get people jealous get those jealous points out there.
Thomas Helfrich 25:18
That's right, Joe Ellis points. That was jealousy points, man. Yeah, there, we have taken a tangent. This is how we probably should have started the whole thing. Because if anybody's still awake now, please wake back up. Yes, um, let's shift gears a bit, hey, for the 18 people that are gonna listen to this, I would tell you this, if you are not looking at the Intelligent Automation AI because you you don't or or any of these advanced technologies to help automate business processes on fundamental level, you're, you're just there's no way in and maybe just don't want to scale your mom and pop shop, that's fine. But if you're a midsize company trying to go from like 100 million to 500. And that is not part of the core strategy. I don't know how you make it. Honestly, I really don't, in any form, I don't care if you're selling something if your services, if you're, you know, literally transactional in nature, to me, just the margin shaving on operations and the ability to make decisions faster and more knowledgeable. It just has to be part of your strategy. Otherwise, you're just in the dust in 10 years?
Aaron Back 26:14
Well, you are and I've even I've even seen the extension of AI out to the edge to they're they're actually building Well, yeah. Oh, yes. devices to basically get it closer to the end user that's using whatever. Well, it's a nice piece, right?
Thomas Helfrich 26:30
It's sorry, exactly. Think about so the you know, like these little you know, I'm hold up a phone, but there's so much compute and the fact that it sits at nighttime at rest, I you know, are pushing compute out out to the and they doing it now for device level, then then you're you know, you're closer to the it's absolutely how that's gonna be everything's gonna become one machine. Not everything a lot, they'll give you the ability to do a computer on a one sheet machine edge computing level, where it's more data, small data transfer, and more compute at the level to do what you need. If you need some bigger all the data comes back up to cloud, which is also you know, almost like a, you know, content management, you know, distribution kind of network routes. Er, then it goes, maybe knows when a meeting qubits, it's over well over
Aaron Back 27:16
what needs it only, it only fetches when it needs to knows when it needs to fetch something from a cloud base or update the
Thomas Helfrich 27:22
model. Right. But like I said, when we get to qubits game, oh, well, yeah,
Aaron Back 27:26
there we go, then we're all gonna die, quantum AI and so forth. Ah, yeah. Well, hey, it's been fun chatting with you. I think we've hidden closer up to the time. But I do want to ask, you know, if there's any last bits of takeaways, you can leave the listeners and something I want to maybe qualify that a bit is, I kind of started off with a bit in our conversation, but I want to kind of end it with this. What is the expectations versus realities and AI? A lot of people think AI can solve world hunger or make everybody really, really
Thomas Helfrich 28:02
wait, you can't, you can't say hello.
Aaron Back 28:06
I see some great stuff in the ag space, though, with with AI and IoT, awesome stuff going on there. But water conservation and so forth. But real quick expectations versus reality, on AI, what should be what should be idealistic expectations versus, you know, pie in the sky.
Thomas Helfrich 28:27
They're quite an you know, it's a, the final definition of AI to give us everything we can't do yet. So I think the realistic pieces are is any built for purpose, the realistic is that's, that's here. And that's, you know, still nascent, but incredibly powerful, nascent technology, like it's mind blowing, if you look at it, that that would have been possible five years ago, and you look at an hour, it's like, man, man, right? But but it's, it's there, and the foundations are there to be built upon. Big picture stuff. I think, you know, if you think about state sponsored level AI with data and you know, narrative manipulation, or just geopolitics, like you'd be, that'll be in play be, in my opinion, not to sound like a conspiracy theorist, but it'll be in play be long before already is before you, then we're aware of it. And it's like anything else, it's so it can be so powerful. What can be done with computers and compute, I should say more than computers that you're realistic is that it's assumed that it's already there at any point, but it's just there. Honestly, it's probably not going to affect 99.9% of us, but day to day, your realistic expectations are and if you're buying it, as it has to be built for purpose, have that proved out make sure it's proved out quickly, and then build upon it with with reasonable it's still technology and people so it's not going to just be implemented overnight. And it's not an it shouldn't also take two three years. So your expectations are, you know, simplified business goals, find things that are built for purpose, and if all possible, don't get get another application, get something that actually is integrated and works with others.
Aaron Back 29:58
Right? I like how you touched on a while ago, make sure it's part of your core foundational strategies going forward, not just some sort of afterthought. You've got to be purposeful and intentional with it. And not just from what you want to build, but how it fits within your business strategy, set aside the technology for a moment and focus on the how you want to
Thomas Helfrich 30:17
leverage Exactly, yeah. How does it affect my business outcome? You know, there's two rules I always go with is where do I get my revenue today? And where do I get it tomorrow? And if it doesn't serve one of those needs, don't do it.
Aaron Back 30:29
Well, yeah. Face. Yeah, yeah. Basic stuff. But in modern age, like we still need to hearken back to it. So I like to sound smart. There we go smart on our smart house. Smarty. Smarty Pants.
Thomas Helfrich 30:44
I got educated. Yeah, I'm back way older. Listen, thank you so much. I love your beard. You don't think you got it? Like it's it's the beard in the room.
Aaron Back 30:55
That's what brings it up and creates the great conversations we have. I'd like to get
Thomas Helfrich 31:01
across the middle of it. Yeah, all natural. Oil natural. Yeah. Yeah.
Aaron Back 31:07
All right. Thank you so much for joining the show. I appreciate it. Great anytime nation. Yeah, let me have to have you on again in the future. But till then, thank you so much. Cheers.