The Future of AI in Distribution

Margaret (00:09)

Welcome to the Kodaris Community Show with your hosts, Tony and Margaret and the occasional friends stopping by. This is the podcast where we explore how innovation and technology is reshaping distribution and the supply chain as a whole. Discover how technology is making companies more efficient and profitable, making their customers happier, and is paving the way for our future. Join us for insights from industry experts, interviews with innovators, and actionable ideas to stay ahead in our rapidly evolving world.

Margaret (00:42)

For this episode, we're joined by our very own Mike Mazzoni to talk about the future of AI and distribution.

Margaret (00:51)

Welcome back, Mike. So nice we're having you twice.

Mike Mazzoni (00:55)

Thank you, it's great to be here. I'm excited about today's conversation.

Margaret (00:59)

Yeah, me too.

Tony Zakula (01:02)

He's still at Kodaris since the last podcast. But the good news is she doesn't know you well enough, Mike, because she called me old last episode. So we'll see how comfortable she is with you.

Margaret (01:15)

I did call him old last episode. At the very beginning of the episode too, we kicked it off with me calling Tony old. I wouldn't dare call Michael old though.

Mike Mazzoni (01:25)

Even though we're the same age. Sorry.

Margaret (01:29)

Well, segue–what's new is AI, but is it really? And so today we're going to talk about the future of AI and distribution. And I know both of you have really interesting takes on this. I guess I'll start off. Tony, TUG 2024, you gave a really interesting presentation about AI and broke it down in a way that even me, when I was dabbling with it, didn't quite understand, like what it was or like the technology behind it.

So I'm imagining some of our listeners might also be like, yeah, I've heard about it. kind of know what it is, what it does, but can you give us a little spiel on, what the heck AI even is?

Tony Zakula (02:13)

So 30 minute presentation and 120 seconds, I mean, as for AI, it's all math. It's mathematical algorithms, pattern matching. The difference we have today is what seems like AI knows things. The hardware, the software, the capabilities now to be able to store billions and billions of data points and search them near instant is now available.

So AI has been around for 40 years. But the ability to basically have all the knowledge of the world or a good chunk of it that's available directly stored in a specialized data format and then you could query it, look for it, and then the math to pattern match with variations is now available.

So, you know, AI comes down to a lot of data and, at some point, a lot of process. But it's powerful because now instead of being able to have a hundred million data points, you could have a trillion data points and basically access those. And it's getting better, it's getting faster. We learn and the tech industry iterates.

And one can get into more detail about the iterations of AI. But at its core, it's pattern matching. It is somewhat human-like, because you have to think about us as children. As we grow up, we learn. We store information. We're influenced by our experiences. Our brains learn not everything instantly.

But we can only learn so fast, comprehend so fast, where AI doesn't have those challenges. But it cannot fully reason and understand or be creative. The generative AI is starting to feel like it's creative, but it's really doing a lot of pattern matching.

So there's a term in the tech industry, AI producers versus consumers. We're all for the majority of consumers, but the producers are the ones iterating with it and creating new technologies with it to make it go faster.

Margaret (04:27)

And it's like vector, it's like math, but vector math, right?

Tony Zakula (04:31)

Yeah, vectors, when they use that term, let's say you have a 100 page document, AI really chunks that up into sentences, even words, stores them in vectors, which is just a series of numbers, mathematical numbers.

But that's how they can query it so fast and pattern match, because instead of looking at words or documents, it assigns pieces of that document but into mathematical vectors and that's how it stores and queries and pattern matches so quickly because it's all mathematics.

Margaret (05:01)

Mic drop, we're done.

But I guess specifically in terms of technology for distributors, what has changed over the last two years, but even last 12 months?

Tony Zakula (05:13)

Yeah, when I spoke at TUG, you know, we were working on e-commerce search, right? And, of course, a year ago at TUG, AI was a buzz. I think nearly every session had some AI in it. It wasn't so this year. There was some, but it wasn't a buzz. ChatGPT, you know, we're all getting used to it. It has its limitations. Whatever AI you want to use.

We actually didn't launch our commerce search, AI assisted hybrid search, until about two months ago, I think, or three months ago. It took us nearly a year to get it where it was better than the traditional search. And that's because any new technology has challenges. AI is good at certain things, not other things.

So I think what's changed in the last year is one, there's a lot of AI but you have to have business value and it has to work. There's a lot of rush to market, a lot of churn, I think there are business value use cases starting to appear where you can point to and say it's providing real value or improved experience. But it's difficult.

We spent 25 years honing the current technologies, and now it's a new technology. And it's going to take some time to really hone it and leverage its power, but do it in a way that provides business value.

So I think that's the biggest, and we can talk about some of those changes, but that's the biggest change over the last year is one, distributors saw, this is cool, but I'm not seeing how I'm driving value. Or two, there's security implications. Or three, I've tried it, it's not giving me value. I'm gonna go back to the old way until I can actually see value with it.

Margaret (06:52)

Yeah, Mike, what are your thoughts? Are you seeing the same things, slightly different changes?

Mike Mazzoni (06:57)

Yeah, so first of all, I'm glad Tony's here to make it sound smart in terms of vector math and words into numbers. And obviously, it's very important that people understand that. And I think that understanding is going to be critical in our success as we continue to build out new capabilities.

But in my career, coming from the product side, the business application side, I think the challenge has been as it becomes more accessible or as the technology improves, to Tony's point, what are the scenarios or the processes within my organization where we're gonna be able to derive value from implementing artificial intelligence? So that's been the challenge. And I can tell you over the last, three, four years that it's been the big buzz.

That's the message that I've been reiterating to both the distribution community as well as the technology community. The ones that are building applications is, this is really cool. What's the use case? What are we going to be able to do with it that the distributor is going to be able to recognize value or the company in general is going to be able to recognize value.

And then as Tony said, it doesn't happen overnight. We define where we think their potential value is, and then we work to incorporate AI tooling, often driven towards automation, to find the value that we can extract on the other side.

Margaret (08:24)

I think, a couple years ago up to last year, it was like so buzzy that people were like, what's your AI strategy for your business rather than how might AI be a tool to help you actually in your current and existing business strategy, right? Like I think a lot of people were asking the question bottoms up rather than top down.

Tony Zakula (08:47)

Yeah, I think some important things about AI. So now that we have the definition of AI, there's a couple of things that distributors can be aware of that become, you have to look out five, 10 years. Where are we going to be at with AI? What am I going to be doing with it? So I think there is a strategy to be had, but it's not from today. It's more if you look at an executive plan, five years.

It's not even the, what will we be using AI for, right? It's one, how can AI disrupt us? Or two, how can we disrupt with AI or grow our business? And how does that interact? What do we have to do today to start preparing for that? Because the technology is going to take a journey that's going to change the world, change how we do things.

But predicting where the AI is going to go, nobody can predict right now. We see patterns and trends emerging. But if I think about, one thing AI requires is data. We all know data is valuable. We have all kinds of laws and audits and everything about stealing data.

Your data is your data if you keep it private. The thing that people don't realize with a lot of AI tooling is they're sending their private data out to AI tools when you're using third parties who are not designed to protect it. I mean, that was one of the early things I saw, that personalized data becomes your intellectual property for your business that can be a differentiator.

Because when you think about hiring employees, talk about artificial intelligence humans, you hire employees with great knowledge and skill sets. And those employees come at a premium usually, right? So when you think about AI, you don't want everybody stealing your best employees or your best knowledge that's inside your four walls at your business. So you think about protecting that, you want to partner with someone or you want to use AI tooling that's going to protect your data, but you also want to have control of your data.

So the first thing people want to do is, I signed up with this AI company, they said, okay we just need to suck all the data out of your system so we can produce AI, which they need the data, right? But you should be asking the questions, how do I protect that? How do I know you're not going to share with others? How can I get that back?

You're gonna give me benefits but, like, if I don’t finish my contract with you. You know, how does all that work? Because that is your intellectual property that you're sending to them. When I think about, your best employees, it's like training your employees for, you know, 10 years and then your best employee leaves.

Essentially as we build our AI data, you're building a great employee. And so you have to have strategies. How am I going to build great AI employees? How am I going to house that data? How are we going to leverage it? But also, how am I going to protect it? And how am I going to use it as a differentiator for me?

Mike Mazzoni (12:03)

Tony, you mentioned data multiple times and you talked about bringing on an employee and training that employee and making them the best employee they can be, and then they leave and you lose that data. But I think if we just hone in on artificial intelligence specifically, you know, the term is machine learning.

We're training that AI tool or that AI automation agent to do something, to do a job based on the processes that are defined. And in order for AI to actually be effective, the data that's leveraged is critical. And that's what I've seen as we start to roll out more AI tooling.

And often the question that I ask our partners is, okay, as a human, what would you do? Or how would you know what to do? Or how would you know what decision to make? And if the answer is, well, I just kind of know, because I've been doing this for a while, then we know the data doesn't exist. The data is not there to take advantage of AI.

So, yes, it’s important to keep that data secure. That's, you know, your property. But also to make sure we're continuing to evolve in terms of the content and the data that's being captured, that's being used in the decision-making process by humans, because eventually we know if we want to take advantage of artificial intelligence and see that value we talked about earlier, the data becomes critical.

Tony Zakula (13:31)

Yeah, those are great points. Again, we talked earlier, data comes down to being the heart of AI.

And then actually, process is starting to emerge. So when we look at where AI is trending, there's the whole initial one-shot ask a question, ask it to summarize a call, it does things. But then we've got to these prompt type engineering where you have to ask AI the right questions, just like you have to search the right terms in Google to find what you're looking for.

You have to ask AI the right questions to get some answers or say, give me best practices on this and I'll use those best practices to summarize my call. And it uses context and all of that, right? And now what we're seeing, the trend and again we're iterating where you can't just have one AI person. What we're seeing is the emerge of what everybody's calling AI agents.

I'm just gonna call those basically AI people. So when you think about AP automation, you have an agent that's analyzing the documents and summarizing and matching your fields and say, this is ready to go into your invoice center or those things. And then you have order automation and that AI is operating a little differently.

So let's step back just parallel. I like to try to boil things down for us. You do have an AP clerk, right? And then you have a salesperson or somebody who's entering orders. Then you have somebody who's, and even as we interact with other computer computer systems, well I know how to go to this other software, let's call it, I don't know, another ticketing system and enter a ticket, right? Because I have a service request. Not everybody in the company knows how to do everything.

Everybody thinks we're creating something new with AI. We're not. We're starting to emulate human behavior. And now we even are inventing, an AI boss who's interacting with these other agents and saying, you go do this. Then once I get that answer, I'm going to give it to this next agent.

And it's actually developing very quickly. And so as you think about one, each of those clerks, your AP clerk, your AR clerk, as you think about those being AI driven. Good data and good process to Mike's point–If you don't have good data and good process, it's like training a new employee and saying, just follow this person around that's been doing this for 25 years and whatever you feel like that day you put in.

I mean, it's not going to work. An automated agent needs very specific guidelines, very specific data points, very good data. So you almost have to clean up your operations to make this work effectively to get the automation.

I know we preach to our customers sometimes I call it–good data, good process makes for great automation. Without that, you can't do it. It's just not going to be possible. So that is a, you know, when we talk about AI strategy, that's a long-term strategy. We have to change how we might do things operationally to then drive automation to be able to do more with less people or take advantage of this.

And unless you do that, you'll never get the return on investment. So as you think about that and we think about the future, in five years, do we have great processes, great data, is that building? And we might have AI agents that do an end-to-end business process across multiple departments, pick up, trigger, do things. It's not going to be today, and it's not going to be

perfect, but then as humans we're going to manage exceptions, which we talk a lot about, was even before AI, like running your business, managing by exception. So it's exciting, but it's challenging. And nobody's going to get there in three months, six months. The business value will be there.

But probably you'll pick up value just with your data and processes along the way, standardizing them, making them better, becoming a more efficient all-human business, but then it's much easier to leverage AI.

Mike, I'll let you chime in on that. I know, when you think about operations 10 years ago when there was no AI or five years ago, probably very similar, why do still need good data? Or why do you do it that way? The system's not designed to do it that way, right?

Mike Mazzoni (18:13)

For sure. Those conversations have been going on forever. The way the conversations happen have changed, but in the end, look at your process, understand the process, try to remove non-value added steps in the process. And when I say non-value added, does it add value to your customer? And that could be internal, the employee doing the work, or external, the customer partner that you're working with.

And so actually, as we look at these processes, historically, it was, can we actually just remove those non-value added steps because there are extra keystrokes or clicks? What's interesting, and this was a conversation that Tony and I were having prior to TUG, it's, well, as soon as we can introduce artificial intelligence to try to drive that process, and we can introduce more automation.

The number of clicks doesn't really matter. What we're looking for is the value of the result. So it actually opens up some doors from a process perspective to say, well, what is the right thing to do? We're not just trying to get it down so we can enter an order in under 30 seconds. We're trying to get it so we can automate the creation of that sales order, but drive the most value at the other side.

So there's a lot of neat opportunities when we think about data and process and then that human teaching the AI agent what the process is.

Margaret (19:39)

I'm sure hearts everywhere just sank when we heard that it's actually still a data issue. You know what I mean? It's like, wait, didn't we come up with a magical solution to not need to have perfect or clean or clear data?

Mike Mazzoni (19:56)

So what's interesting, actually, is we're exploring–because it's kind of like you have some ideas, we take advantage of the technology here–the potential to for, if somebody is correcting something over and over again, but not cleaning up the source, can we identify that leveraging artificial intelligence and say, we can fix this for you if you want?

Margaret (20:24)

Make the robots clean the data.

Mike Mazzoni (20:27)

Now, obviously, this is going back to, well, what are the use cases that are out there that we should be thinking about that have value that'll benefit the distributor?

Tony Zakula (20:41)

I think the key thing, one key thing there is repetitive tasks no longer become the blocker, right? Because the robots can do it again and again in a millisecond. So when we think about keystrokes and time consuming things, once you automate a process, the keystrokes no longer matter, right?

And it's not just us talking about this. Amazon Web Services has been talking about it for a year–how SaaS software will become a commodity because you'll have agents that know every piece of SaaS software, be able to interact automatically. They will, I mean the whole industry is trending that way.

So, you know I think what becomes important is that process and understanding interactions and then automating that, whether it's even cleaning up data or anything we as humans are doing repetitively starts to become an opportunity, right?

I think an interesting thing though, is when we think about the future and we think about these agents, right? You have all these employees. The challenge becomes, especially for executives or leaders of companies, well, I'm going to go buy this AP automation software and then I'm going to go buy this order automation software essentially outsourcing every department of your business.

But in the future, we just talked about those agents can now start to talk to each other. And maybe it takes us another year, two years to get really good agents in each department. But if you're with third parties, they'll never be able to talk to each other. And those companies will control that part of your business. So the future is platforms that you can house all your AI on one platform and all your agents together.

And it's not just me saying it. AWS is preaching it. We're doing it.

And in fact, ServiceNow, I think they're one of the largest SaaS companies in world, they are seeing their customers consolidate down onto their platform from other SaaS software providers already. And their AI is going up. Why? Because now the AI can start to interact. Having everything under one umbrella, starting to be under one provider, controlling their destiny more and more.

By limiting who has their data, how it's being used, they’re saving cost and AI can start to interact across the business, not just siloed. So there will be a, you know, there's a lot of froth in the market, a lot of ideas right now, a lot of startups, a lot of people doing a lot of things. At some point when it comes down to raw business value, you're going to need your AI under one house because it's just not going to function well spread all over the place.

And so I think people have to start thinking about that when we talk about, as a good point, AI strategy. It's going to change. But when we look at where we will be at in five years, you want us to be thinking, what am I buying today? Where am I putting my data? How am I preparing for a 50% automated world or 60% automated world where my employees are experts at the human relationship, at the exception management, decision making. But we're no longer doing repetitive work.

It's interesting because people say, I'm sure clerks and people start to worry. But honestly, one of the most cutting edge pieces of it is engineers, software engineers. The AI is probably fastest moving at eliminating the job of software engineers in any other business right now because the world's filled with trillions of lines of code that AI has access to.

So it's going to just change everything. It doesn't mean the engineer who can solve problems and tell AI how to solve them. The team lead is now the team lead of the AI agent to go write the code. So it's interesting times.

But, I see that's where the future goes, and if we're engineering now for our customers, it's perfecting different pieces of the business right now, different pieces of where we're at. And we're already doing R&D on how those will talk to each other, how they'll trigger each other, because that is the real future of where we're going.

Margaret (25:20)

Yeah. And I know you're saying it's, it's going to be hard to predict where exactly we are, or where distributors will be in five years with AI. But do you have any outlandish ideas?

Tony Zakula (25:33)

I think the whole transactional system, think about a price increase from your vendors, you get an automated feed. It's going to trigger AI to run your margins, figure out your margins, stage your pricing updates, then ask a human, here's a price increase, here's how it's going to affect margins, here's how I'm going to adjust pricing. Can you please approve?

I mean, I think that's two years out max. That's easy. Hey, I got an AP invoice. I put it in automatically for you. There was no exceptions. Hey, it's due in one day. Here's all the invoices I'm going to pay. Click approve. Like that's totally possible. That's not outlandish. That's just perfecting some flows.

We've got this order volume increase. We need to adjust our inventory levels. Based on historical and trending, I'm recommending we buy X. You'll still need a human that understands the business. But anything you're having junior people do, or data entry, or any of those things, that all can become very, very automated.

And I don't think that's a long ways out there. That would be core business functionality,

Mike Mazzoni (27:03)

Just to elaborate on that, Tony, so you talk about how the data sets and the transactions really are correlated, they're tied together. So when you think about specific solutions or different third party applications focused on different processes, that becomes, you have some scenarios we're working through right now, but something as simple as an acknowledgement on a vendor purchase order, right?

Okay, the dates change, the cost might need to be reviewed. That seems like a simple update. Yeah, it's a great process to automate, but then you think about the impact that potentially that has on the rest of the business. First of all, if I am correcting the data initially, because the POs told me what the data should be or given me an opportunity to make adjustments,

From a downstream perspective, now my accounts payable process is streamlined. And that automation flows through without exception at a higher rate. Along the same lines, we talk about inventory levels. If a date changes, if one date changes, should AI be able to recognize that that's going to potentially result in a stock out and make a recommendation on potentially alternative sources?

Should it also be looking at customer service and fill rate targets? Should it be looking at open transactions on the sales order side to determine if we need to, should AI be saying, hey, you've got an issue here, you might want to connect with the customer or look for an alternative.

So when you talk about that, you know, having all your AI data, all your data and your AI tooling, your different agents in one place to have the ability to easily communicate understand data across the organization to make those suggestions so we can truly effectively manage by exception without having to have multiple systems and multiple data sets. That stuff where having those conversations today, the user experience may continually improve, but, you know, to have the data in a common platform enables that visibility today.

Tony Zakula (29:18)

Yeah, just like with any new employees you hire on day one, they're not going to be the most productive. It takes time. And we know with technology, we iterate constantly with our community of customers and perfect and polish. And it's the same. The AI tooling today versus a year ago is unbelievably better. It's stable. It's getting better.

Margaret (29:39)

It knows how many Rs are in strawberry now.

Tony Zakula (29:42)

Yeah.

Part of the challenge with doing all this interconnected is the data, the agents, the interactivity between all of them. But it's moving very rapidly and we're having those conversations about how do you do that? How do you trigger that?

And really when, even internally when we're studying how to engineer some of these things, it's interesting how many parallels there are with our human behavior because essentially you're trying to parallel that decision making triggers criteria and sometimes you just have to lay out well how would we do this and then you back that into data and AI that then can perform it know thousand X faster, ten thousand X faster how do we give it enough knowledge so that it knows but once you have the pattern to build off of it becomes faster and faster, right?

That's how technology is. Every breakthrough leads to a faster iteration on the next breakthrough. It's the same with AI right now. It's moving, but going back to distributing real business value, now we're talking a reduction in overhead, speed in decision making, understanding your business, your customers better at scale. Being able to more efficiently serve your customers. Potentially sell more with less overhead, better margins so you're back to your business.

So I've always said for the last couple years, wait and see. Take a slow approach to AI. Take a slow approach to adopting things. Because if you invest too fast, too heavily, and have a wrong solution, then you're going to have to back up and retread. So let things settle out a little bit.

Margaret (31:41)

Well, that's a core Kodaris value, smaller bets mean quicker feedback loops. So you've got to go all in all at once. As you both were talking, what's your prediction in five years? What will be the least amount of headcount at a $1 billion distributor?

Tony Zakula (31:58)

I'm going to say that like when cars were invented, the horse industry became obsolete. That's very hard to predict because you still have your service levels, you're going to have your relationship salespeople, I don't think it's about the non-humans. I think it's about service levels.

Look, we know technology has raised the standard of living globally in the last 25 years. I think it gives everyone a better standard of living. Will people's jobs change? Will there be less software engineers and more creative workers? Will people transition into other careers? Absolutely. And that's part of our economic cycles, right?

So it's hard to say. I think you'll have a lot less people doing repetitive tasks, a lot less. But better buying decisions don't mean less deliveries. Better salespeople having more customer relationships and having fun with the customers maybe rather than typing up orders. I think that's a hard question to answer. It will look radically different than it does today.

Mike Mazzoni (33:10)

My perspective is having spent 25 years working with wholesale distributors in North America and around the world, we've always talked about how do I do more, get more accomplished with the same, right?

And that's core to the fact that these businesses, their families, their teams, a lot of them have long standing employees and long standing relationships with their customer community. And so, it's always been about how do I grow? How do I scale? How do I do more with what I've got today?

And so I agree with Tony, we don’t know. Things are going to change, responsibilities will change, the roles, the redundant tasks will be reduced or eliminated. What it does to the actual makeup of the organization in terms of humans, it's hard to predict, but I think it's exciting to think about and I think it actually provides opportunity.

Ae talk about moving those people, those humans from non-value added activities to value added. And distributors in general should be continuously thinking about how do I add a value to the relationship I have with my customers? So we start thinking in those terms. We want to sell solutions, not products, right? You know, what is the additional value we can create?

Then it becomes not a matter of how many people will it take, but what do those people do and how are they providing value to the supply chain and specifically to their customers?

Tony Zakula (34:50)

Yeah, it's an interesting question because even for Kodaris, we are leveraging obviously some internal AI tools around development and just the whole where we spend time working that's maybe not value added. And then we've talked internally also about having internal classes to help others understand how to more effectively in our business.

So if you cut development time in half, and we're training engineers to make requests from AI correctly so they get better results. As we deliver faster, I don't know if customers ever said, OK, you did that faster. I paid less, so I'm done. They said, OK, so now I want this thing, and can you give it to me even faster?

So we haven't seen a slowdown just because we're getting faster. We've seen, if nothing else, more demand because we're getting faster. But it is giving the customer more value at a faster rate, which is exciting. It's hard to know where it's going. And so maybe software engineers who learn code, that might be gone in five years. But software engineers who can talk to AI and have it write the code for them effectively is something that was unheard of last year. So it becomes interesting how the markets change.

I read a lot of what Jeff Bezos talked about over the years, is two things our customers always want it faster and they want it cheaper. Those two things never change, right? So if AI is helping you give them more of that, you're going to be successful. Doesn't necessarily mean you have less demand and less people per se that are doing other things.

Margaret (36:46)

Awesome. I have a fun personal question for each of you to respond to. What is one way that you're using AI either in your work or personal life that you were not using it six months ago?

Mike Mazzoni (37:01)

Thinking about my personal life because that's the direction I wanna go.

Margaret (37:06)

You asking chat GPT things instead of Googling?

Mike Mazzoni (37:08)

I'm not sure. No, should I be? You know, I think, I mean, honestly, where it's infiltrated. And so I don't know that I necessarily looked to take advantage of AI is, you know, when you're watching TV shows and movies and you're watching Netflix and whatever, I do let the suggestions influence my watching habits in terms of what should we think about watching next?

So, I mean, from a personal perspective, that's one where I haven't actively sought out AI and said, how can I use it? But I think we see that influencing our lives. My wife would also say I'm very susceptible to, at the grocery store, things at the end of the aisle. When you're in the consumer world, when you're shopping online, well, I need that.

I didn't know I needed it, but somebody thinks I do and I'm pretty sure that's AI in the background telling me I need it.

Margaret (38:16)

Impulse buys. Good to know that an end cap will get you every time.

Mike Mazzoni (38:20)

Yes.

Tony Zakula (38:23)

Yeah, I think for me, any good software engineer is traditionally lazy. If you're going to build something, he doesn't have to do the same thing 10 times. We have this old saying from over the years, the fastest way to get something built is to make an engineer do the repetitive process five times, and he'll be like, I'm done with this.

So for me now, you know, we're dabbling in AI, whether it's coding, whether it's entering some data and some piece of software, my brain immediately goes to there's got to be a way AI can do this. If I have to do this five times there's got to be a way to develop something faster because I don't want to do this repetitive task. We're taking advantage of AI whether it's entering tickets, whether it's call summaries, whether it's coding.

And I think that's the fun part about it too is now we can have these potentially superhuman kind of people maybe helping us all the time. And so now it's like, where can I find somebody to help me so I don't have to do this, right? You start looking at the world that way, right. And, and you have to be careful because you want to do it right.

But, it's interesting because that's immediately where my brain goes sometimes three or four times a day. Can we get AI to do this? Maybe not today, maybe in a year, maybe in two years, but it's interesting to contemplate the world.

Mike Mazzoni (39:48)

Actually, Tony, going back to what we talked about earlier, which was, all right, the technology has come so far, but what are the use cases? Where can it be applied? I think asking that question, and sometimes the answer might be no or not yet. But if you're constantly asking that question, whether it's in your business life, your home life, that's what's going to drive those use cases.

Tony Zakula (40:09)

Every time when I go on, I rarely do it, maybe once a month. I buy something off Amazon, mostly the same thing I bought last time. I go, why do I have to log in? Why do I have to go to the cart? Why do I have to go look at what I bought last time, put it in the cart? Seems like somehow I should just be able to order me this again. And it's probably coming, right?

Margaret (40:33)

We need to get you an Amazon Alexa. You absolutely can do that.

Tony Zakula (40:40)

Then she's gonna listen to me all the time and that's bad too.

Margaret (40:43)

Yeah. I mean, I think the consumer applications with ChatGPT and stuff and personal lives, I've seen people do some really creative things of like put in their, you know, the diet and their goals and what kind of food they like, and it'll just shoot you out a grocery list and a meal plan.

And then you can probably even sync that, to your point, Tony, to your Amazon, Whole Foods cart and just get it. It's amazing to think of what we can do. I had it analyze a message and give me, like, personality traits, pull out personality traits, and how best to communicate with this person. So the options are endless.

Tony Zakula (41:23)

She was studying, Mike.

Margaret (41:26)

I'll give you some ideas, Mike, to play around with.

Mike Mazzoni (41:30)

There you go.

Tony Zakula (41:31)

I like a little bit of control in my life, so I do like to go to the grocery store and choose. Because I just like to.

Margaret (41:40)

Because you might get inspired by an end cap item. There's no end caps on a grocery store online checkout.

Mike Mazzoni (41:47)

Do you know what kind of cereals they have today? They’ve got crazy stuff now. And it's always at the end of the aisle.

Tony Zakula (41:55)

And I, you know, as an entrepreneur, I hate people telling me what to do, so I just have to balance my AI with, it's gotta work for me, not me yet. So that's kind of my line there.

Margaret (42:07)

I like it. You're going to boss the AI around. They can't be telling you what to do.

Tony Zakula (42:11)

Exactly.

Margaret (42:13)

Well, thank you, Mike, for joining us and Tony, always a pleasure. Any last words about the future of AI?

Tony Zakula (42:21)

Embrace it and have fun, do it sparingly until you figure out the value.

Mike Mazzoni (42:26)

I would agree and only that personally I'm excited. I think our customer community and the companies that I talked to are excited about it. But as Tony said, think it through. This is a journey. We're gonna get there and it is gonna be fun.

Margaret (42:44)

Mic drop.

Tony Zakula (42:45)

That's a wrap.

Margaret (42:46)

If you've been enjoying this podcast so far, take a moment and like and subscribe or leave us a comment. If you have any feedback, we'd love to hear it.

Margaret (42:55)

Thank you for joining us this week on the Kodaris Community Show.

We'll see you back next time.