AI & Industrial Innovation: Evan J. Schwartz on Transforming Waste Management

So, we put cameras on there and we taught AI how to look at stuff and see the container, identify the container, and you see the lids are up. Well, there's more than 90 gallons worth of stuff in there. So, now my algorithm is messed up. Well, how do you deal with that? Well, if I'm if I'm charging you for a 90gallon container and you're putting 120 gallons of stuff in there, then I'm not charging you the right amount of money. There's revenue leakage there. So, I need to have a conversation. Let me try to educate you on keeping the lid down. Don't over stuff your containers. or do you need to buy a second container, Mr. Customer, because you guys produce a lot of waste, right? Either way, that helps me get closer to my idealized route to save my 17 gallons of diesel and

make sure that I'm actually charging for the service I'm doing. Right? So, that's one. >> Hi, my name is Demetri Bonichi and I'm a content creator, agency owner, and AI enthusiast. You're listening to the AI Agents podcast brought to you by Jot Form and featuring our very own CEO and founder, Idkin Tank. This is the show where artificial intelligence meets innovation, productivity, and the tools shaping the future of work. Enjoy the show. Hello everyone, my name is Dmitri and welcome back to another episode of the AI Agents Podcast. In this episode, we have Evan Schwarz, the chief innovation officer from AFCS Group. How you doing, Evan? >> Living the dream, brother. How you doing? >> I'm living the dream, too. I I love to hear that. you know, uh, the chief innovation officer. This is a, um, if I'm not wrong, I haven't heard this

term or not term, this role used, uh, in a while. I believe it was a Simon cynicism in a book uh, from a while ago. Is it not? Is the chief innovation officer uh, or how did you how did you first of all, how did this role name come? Because I don't think that's a common seauite name. Maybe I'm maybe I'm crazy. So you you'll see it across a lot of larger companies, right? If you think about your small and your mid-size customers, innovation is really something that's shared across all of the leadership roles generally. And smaller companies can adopt it. They can run at it really well. You you grow to a certain size and then those individual departments start to take on commitments to a board. They have commitments in delivery of what's in their department. Innovation then tends to start to inch

out. you start seeing slowdowns in delivery of innovation. This has historically been the it's been modeled in a way of saying smaller companies more agile, bigger companies turn like a battleship in the in the ocean, right? Well, the the reality is the answer to that is that you can start to compartmentalize components of acceleration. Chief innovation office does that where I I' haven't kicked out all the people that have been innovative in AMCS for like 20 years. They're all still there, but they do have day jobs. Essentially, they have core responsibilities. So, my job is now to organize those ideas for innovation, get it funded, get it justified, and then be able to accelerate delivery faster, but I don't own the product, right? So, I might just take on an initiative, drive it to the delivery, and then lift it back up and hand it

back into the product organization or where the innovation needs to go to live long term, and then I run it the next thing. This is just u it's maybe an answer to symptomatic slowness that large companies inherently get just due to their size, right? So it allows particularly if you're in a technology company, if you're not driving innovation all the time, you're not really in technology, right? You have to have a focus on that. So this is just a way for a larger company to drive focus and innovation and accelerate delivery so we don't lose that edge. And that that's been at the core of AMCS Group since the beginning. And once we started to see that the concept to market was starting to slow down, uh we put this into place pretty quick. >> You know what's interesting? I um just to call it

out, uh my inkling there was actually towards the chief vision officer. That was the Simon synchism. I got it wrong. That's it's another term that I hadn't heard. But very cool. I like that. I know that a lot of big companies Yeah, that is the critique mainly is they're a little bit more slowmoving. They're not as agile. they can't just have somebody pop in and you know do everything um for them on a quick uh basis. So my question then for you is tell me a little bit more about what you guys are doing at AMCS Group and you know I I think in regards to the start of the conversation focusing on obviously what you guys do in general and then honing in on the AI components we discussed pre-call. >> Sure. So if if I had to really kind of describe in a

a handful of short words what AMCS is, AMCS services the resource inensive industries, the wealth industries of the world where you're doing something with raw materials, sourcing it, manufacturing it, putting it into a product and getting it out to market. But we have a eye on sustainability. We believe in our heart that technology can solve the sustainability problem. I don't always have to strip mine every single thing once I've put it into the market. And we've been trying to recycle for a few decades. But what people don't really understand is we've been we've been harvesting materials from the land for thousands of years. We're really good at that. The logistics from taking something out of the dirt or growing something out of the dirt or manufacturing it, getting it into a logistic supply chain, getting it all the way out to the market. We are

super good at that. In fact, we're so good at it, it's it's been cost prohibitive for many decades to go back in the market, get that thing back and run it back through the supply chain. And technology is the solution for that, right? And if if you if you have any age on you at all, you've lived through the days of the first attempt was putting a whole bunch of cans out in front of someone's house and they had to sort plastic here and aluminum here and glass here and that was too hard, right? Nobody did it, right? It's hard. You're not going to get a consumer to do that job for you. And so then we got better sorting. So you could just now put it on one bin, it goes out to a murf and a Murf is sorted. But then that still

has a certain amount of risks and inefficiencies to it to where yeah, we're doing it, but we're not making any money with it. But you know, at that point, sustainability comes and puts their boot on your neck and says, well, you need to make a better world. We want to make a better world, but it still has to make sense to the board and the people invested in the company. You can't do it at the sacrifice of the company. We'll be out of business, you know. So AMCS is looking at these techs from a performance stain sustainability perspective. Let's do the right thing, but it makes money. If I told you that I could optimize the route of your fleets, whatever the fleets are service, you're going out, you're servicing customers, and you're coming back and I could use an AI algorithm to make that

efficient and shave 17 gallons per vehicle a month off of your your expense report. You're doing two things. You're not pumping that into the air. You're not burning that diesel. And that's 17 gallons of savings and you're doing the same a lot of work. Why wouldn't you take it? That's just free money. All right. So that that's where we kind of started was optimizing the routes, looking at inefficiencies, trying to figure out how to do processing. And when we put it together, we realized what we were building was an idealistic route. Then the real world came in and we noticed, well, we we really put a strong algorithm here, but we're not seeing the optimal returns. It's idealistic. Well, what happens if I'm going to that service and let's just say it's trash pickup and I've got a 90-gallon container there and part of my

optimized route is I can put so many now 90 90galon bins in this before I have to go to the landfill or the murf to sort it and dump it. So, I'm expecting to be able to do so many houses before I have to do that. So, I bake that into my algorithm. I get a nice little optimized route and when I find out I'm having to dump sooner than I expected. What is going on? We don't So, inner vision AI. stuff in there. So, now my algorithms for a 90 gallon container and you're I'm doing. Right? So, that's one. And then we looked even a little bit further at all that in place, we still weren't hitting the idealized route. What was going on? Well, some trucks, depending on age, maintenance ability, stuff as you know, you know, you drive out of your car

just after a fresh oil change, it just seems to drive better. You can't really put your finger on it. It just feels better. Yeah. But it does. And that it burns less fuel. It's all kinds of stuff. So, we're like, "Ah, let's look at the maintenance side of this. Let's get into where we can take a sample." It's like a blood test for your machine, whe you know, and it's any machine, right? So, back into the Murf where I'm grinding up metal. Yeah. Or working on grinding up pulp. All of those things have moving machine parts that rub metal against metal that have oil in it. And every time I have to take it down to maintenance it, that machine's not producing money for me. That's number one. Number two, it's introducing waste into the environment. Do I have to? Am I on just some

draconian with maintenance the truck or the vehicle or insert equipment here four times a year? If I don't have to, why would I? So, if I could take a blood test sample out of that thing, run it through algorithms, and go, "No, you're fine. You can skip this one. Go on to next. That's bottom line impact. That's per performance sustainability. We're making sure you're utilizing the assets that you have as efficiently as possible with as minimal impa impact into the environment. And the key so we have sustainability ESG reporting. Every time we brought that data in and we've run your business against it, we're not really just trying to ratch you out and show your carbon footprint. But anywhere there's a process with a high carbon footprint, 100% of the time, that's where inefficiencies are. It's like an indicator where this is where you need

to focus on your business for more efficiencies or leakage. It's just it's a strong indicator. So now it makes more sense for you to do the ESG reporting so you can reflect on your own business. So we have built that into place. We have our maintenance. So we do all of that across the line for our customers end to end to be able to hit that vision of a circular economy and we're we're confident we can get there through digitalization and AI and what what is unique about AMCS is we're looking at the connective tissue between stops right everyone could look at optimization go vertically down and see the obvious use case everyone could see the obvious use case for vision AI everyone could see the obvious use case for you you know, MLA predictive analytics on your vehicle. We're we're saying there's a multiplicative

impact. If my vehicle is running better, my ideal route is better. If I'm doing what I was serviced to do, my ideal routes better. Right? So, we're seeing multiplicative effects by looking at the connective tissue between processes and looking for the handoff from one step to the next. So that's AMCS services all of these industries across scrap metal, waste, recycling, pulp and paper. Anywhere that you're sourcing material, extracting material, one guy's waste could be another guy's feed stock. We're looking at ways to interlink all of these businesses. You know, the whole point of the reason why I wrote the book, people, places, and things. You get up high enough, it's always people buying something from somebody, shipping it somewhere, and getting it to somebody else. It is people, places, and things. >> Yeah. Yeah, I find that really interesting cuz um my uh current understanding

of uh logistics and stuff like that in regards to the AI um adjustments is it's just maybe that they're going to automate the uh truck industry, but I I hadn't quite heard that I I've always I've always heard from some people that, you know, like Amazon, etc., those types of companies make their uh money on logistics and whatnot, but the advent of AI inside of that those types of industries had never really uh I guess come to me, you know, like I never I I had never heard somebody break down them using it. So, I find it I find it pretty pretty cool and and it kind of shows you where everybody whether you're a what could be traditionally seen more as a physical goods company, right, or uh physical work in general company or mechanical company has room to utilize AI. uh where most

of us just think about it like knowledge work stuff you know like sitting at a desk uh click clack stuff I most people I don't think the the complete uh understanding of these things is like okay I don't know maybe AWS powers something I don't know like you know you see the football you see the football ad where like Russell Wilson throws the the 80 yard the the 60 yard touchdowns like catch probability I'm like I don't know what physical uh instances that AI is actually helping with these companies, but this is this is very tangible. And and just out of curiosity, cuz most people have only been aware of AI for a couple years at this point. When did you guys start getting into uh these enhancements of what you were doing with these AI tools? I know you were talking to me beforehand

about how there's an agentic component that'll be coming as well soon. >> Yeah. So look AI we we use the term AI in its and this is be inflammatory but I'm going to say in its proper use case you know AI algorithms where it's really just predictive high-end statistic work right I can the AAR algorithm someone at at a base level could say is AI just finding the best route through a maze right so those early algorithms were very powerful we've been using them since then right building on top of it when we start talking about the introduction of transformers and generative AI and things of that nature. Now we're starting to expose opportunities where it's not there isn't no right and wrong answer, right? You get out of the binary condition, you go, okay, there's some there's some fuzzy room depending every time, you

know, if you're ever doing software, you're trying to get requirements from someone, they go, "Yeah, it's like this, but then there's this case and there's this case and there's this case." And then sometimes you just have to make a call and software people start to recoil from those, right? I either I need the rules or or there are no rules, right? Um but where aentic AI and the transformers gives us enough to have bas basic reason to give you a stochcastic range for that to go and flow through. That's where aentic AI is shining. Now right away the hype comes in and you've heard all these failed implementations where people try to boil the ocean with it. The the secret is to develop an an agent extremely narrow use case. If it's your first time, strongly recommend you're modeling an existing human process that you've

measured. You have metrics against and you're now looking to replace either some component of it or if it's narrow enough the whole thing, but you have a human analog backup of it and you know what you're running at. And agents are really good at that where you can now say I'm sitting the agent even if it is fuzzy you're giving it guard rails especially if it's like a decolence any fraction in between these two extremes is okay and I need you to navigate it that's something that traditional programming with if then in case statements can't can't do right because it's it's an infinite possibility but it still sits within a guardrail you're not going to go outside this range and run and if your confidence level drops below a point or you don't think you can get there from here based on your guardrails, you

onramp to a human into premium services. So, we're we've added an innovation to make that work was the MCP layer, right? That it sounds simple. You could almost >> explain that for everybody because uh we've talked about MCP before on this podcast, but explain what the MCP layer is a little bit more if you if you wouldn't mind. >> Sure. So, model context protocol is a descriptive language. So it has predictable components that teaches an AI how to interact with the open APIs of your system. How how if I were to integrate whatever you had, let's just say Salesforce or one of our AMCS into SAP or JD Edwards or Microsoft Dynamics, all of these systems have APIs that allow these systems to share data and talk to each other and perform activities. Historically, you've had to insert a programmer that knows one side or

the other or two teams that collaborate to build that integration. based on some event structure, timing, heartbeat, pumping of data, and it's it's prone to error. It's high cost, one side changes. Maybe they didn't follow data contracts. It's all proprietary and it it's a it's just a difficult component. The best of breed solution, the bane of bre best of breed is that interface layer because it's all having to be customuilt for every single customer over and over and over and over again. MCP says, I'm going to describe my my API layer, right, in a way that AI can interrogate it, understand it, and then interact with it just in time. I I don't have to have any hard-coded pre-built things. you ask a question of me or you give me a prompt, I can inspect your MCP layer. I can work out for myself how

to maybe create a new customer or query for a customer or get the data you need and then execute on the task you're looking for. That's really the value of MCP. If you go back to the late 80s, early 90s, there was a similar tech for humans and in the API space back when it was SOAP, simple object access protocol. I'm not sure how old your users are, but with the SOAP protocol, there was a thing called WISL, which was the Windows system definition language that I, as a developer, could point to the WISL for your SOAP and it built out all my proxy classes for me and abstracted me from having to call your APIs. I could just interface with the object in my code, and then it magically went and talked to your APIs and did what I needed to do. So, for

me, the developer, it was no different. It's kind of that same thing. It's got security, authentication, contracts to make sure the AI is doing the right thing. But because that's there and I can now expose an MCP that has a publish standard, I can put it into a space and as long as the other side has it, AI already knows how to talk to these two systems, share data, execute on tasks, no developer required, right? So that's the real power of it is that I could take just our standard chat bots, Anthropic, OpenAI, they now allow you to connect to agents. The requirement is that whatever you're connecting to has published an MCP layer. So as long as that's there now, it's there available to my chatbot. I can ask it questions. It'll attempt to execute based on what it understands from the MCP layer

and come back and get me the the answer or do what I need it to do. That just opens up all of our digital architecture to an agent to work through the solution without any heavy function requirement document, technical spec document, you know, success criteria, test cases, the build, the developer, and hope nobody changes anything on either side that would be breaking changes. So, I have to go through it again. It's just as long as that MCP layer is current up to the date, the agent's able to go through and work out with a high level of confidence, right, on how to do what you're asking it to do. That confidence level becomes much much better if your use case becomes very narrow. Where you've seen people blow this up is, and I'm not going to give a name, but there was a car dealership

that had thrown an agent out onto their website to help >> or mechanical company had. >> Yeah. and the guy tricked it into giving it a car for a dollar, right? That's the problem with just open-ended prompting out to anybody, right? So those prompts need to be tightly controlled within your agent. If you're just open-ended prompt someone, you have infinite test cases. So that's the challenge. In old software development, I had some kind of guard rails that go, here's what I intended to do. Here's the maximum range of edge cases I need to test against. It's definable and quantifiable. So I could test and build a good piece of software. But if that range is infinite, I now need to think of this same problem in every possible language. I need to think of this problem. Someone could create code words and say when I

say this, I mean this and get past your hidden layer that would protect you from that. That's too much, right? So you agents are going to have to be constrained by specific use cases, control prompting and activities, but the value is still very much there and being able to interact with them. So hopefully that kind of frames the value of MCPs. It just stops me from having to train AI on everything future and past and current. >> And this got rolled out. I think originally like anthropic was the the >> they were the originers of the first revision. The big boys have all gotten together in a little consortium. They're all agreeing to adopt this as an ISO. They're all contributing to it. Anthropic still has the pull request right to say, "Yeah, you're a you know, I mean, if you were to talk to

Microsoft, they're like, I don't like the MCP first standard because it doesn't allow agent to agent, right?" So that's So now we're getting very interesting there. It's just agent to thing and back and now you have to build this complex coordinator. So they're evolving it, but as long as they don't evolve it in a in a vacuum and they end up coming up with five standards, we're good, right? give one standard to the world, let us adopt that standard. And now you've opened up AI to interact with all of our systems because that's going to manifest where I think AI is driving it. We're going to get rid of UIs. They're gone. No one's going to be hand jamming in data into a screen anymore. That's ridiculous. I mean, I'm If you get up and go to work, I guarantee this has happened to you

because it happens to me. I get up, I get dressed, I go to my car, sit in my car, my phone goes, "Oh, you're 23 minutes from work." I didn't tell her I was going to work. How you know I'm going to work? But it does. That's context. So we're feeding just streaming I we are a device as IoT streaming data into these things. And now it's going to have contacts. It knows I'm in my work truck. It knows I'm delivering something to my customer. I'm at the gate. Why do I have to have a guy that vets me? It knows it's me. It's got my phone. It knows what I came there for. I'm on time. Open the damn gate. Let me in. Drop my stuff up and get off and give me my money. Right. As you see the the ISO 2022 standard

and back-end finances coming to place, all of these technologies are converging that are going to allow microtransactions across a digital space when you're working and doing you're just going to see a money ticker going up of all the value you're bringing to getting things done. And you're going to be orchestrated where you're owning or command and control hundreds of agents to get things done by just making play calls and pushing things to get them done rather than the complexity of death by meeting to get 10 people to agree to do one thing in a way that we can all agree on. Oh my god. So look, I'm look, you know, there's a part of me that's looking forward to it, but there's there is an optimistic future to this, but we have to choose it. >> Sure. Yeah, I agree. No, that's that's a good

point. And my I have a question for you then in in this regard, you know, it seem I from what you told me beforehand, you're working with uh those of us uh the youth so to say. No, those uh college age uh people uh who are kind of navigating this transition from more of a develop development's a good route to go into to now AI is a good route to go into when it comes to a future career. though. Could you talk a little bit more about that and how um all of this kind of plays into what you just discussed? Cuz I think there's so many different new layers of AI, right? And new components of AI. Like I I I host a podcast about it, so I'm I'm at the forefront of it. It's not a problem to me. But the average person,

you know, it's kind of it's kind of hard. So what what are you telling the tell tell us a little bit about what you're doing and then how you're kind of talking to these uh >> Yeah, look, they that you're you're spot on. So the the pressures into this is I I feel very bad about the guys that just got their masters in computer programming, but there's a whole host of students that are along that journey maybe a year, two years in and they're seeing what feels like being rugpulled that the universities aren't going to offer a degree for that anymore because there's no work to be had under that degree anymore. And to me, those are the lucky ones, right? So now we're at a point where we can build a better curriculum. And at Jackson University where I'm a professor, I I teach

on the data analytics and AI side and technical project management. So those are the two areas that I think humanity is going to have a a strong uh long-term viability is risk management and understanding how to use the tech. And as part of that is to as as a member of the advisory board is to help build out the curriculum for the computer science. And in that there's still going to be a little bit of basic programming because it teaches you logic, it teaches you form, you know, Python's still a big deal. So of programming, but no one's going to be master's senior highle degree in it for very long. Then the next piece is no one's building the next generation LLM or LCM, right? They're not they're not going to come in and compete with OpenAI or compete with anthropic. But understanding the fundamentals,

what's under the hood, what is neural deep learning, what is the multi-layers, what's a Boltzman model, how does data get in, how does dimensionality work? Those fundamentals are important. The applicable skills is how to leverage tech as it is like the product innate in.io. I'm sure if if you're in the AI space, you've seen that. >> Yeah. So, I'm I'm like pretty um nice thing about this podcast is I'm I'm pretty deep in uh AI uh building and um automation building for my own company. So, usually I'm I'm I'm at least like sticking with the the guests to to a reasonable degree. So, absolutely. NAN is a is an amazing platform. >> Absolutely. So, that that would be we're we're I'm not going to say it's across the line, but we had pitched that is maybe that gets taught in the class because it's really

good on flow and how to connect in models. I'm an advocate, but I got there's a whole board of people I got to convince. Yeah. And someone could come out with a certificate or a degree in those because the reality it's the fundamentals of that. Even if I didn't have NA in IO, I would have someone says, "Here's your agent. Here's the LLM brain for the agent. Here's the MCP enabled task for that agent. Here's the flow of data." So, those are still fundamental skills regardless of what tool is available, whether it's Power Automate or some other tooling. you know, how to do prompt prompt engineering. Um, a great example of this was an article I read where u a beautiful artistic photographer was wondering whether or not he had a future and he was showing his pictures and artwork and they they evoke emotion,

you know, the sunset to people and he went into AI and he just took a selfie of himself like you would be right now, put it into Open AI, described it and he came out with this shocking, jarring, emotional black and white face, waters one side, shadows on the other exactly the way he'd envisioned and he's like I think I'm out of a job. And then he slept on it and came back and he goes, "Yeah." He goes, "You know what? I'm not out of a job." Because as I was reading it, I knew it right away. I said, "There's no way I'm making that thing produce that thing you just did. The only difference is one, you had a 35mm camera. The other you used chat GPT to evoke it." That's a prompt. It's just a different lens. It still needed you, Miser Photographer,

to have that vision. I wouldn't have had it. I I don't I wouldn't I couldn't even gotten close to understanding the lens flaring discretion and the lensing. So, it's just a different camera that didn't need a physical camera to evoke it. It >> Yeah. To to put it in the words of Steve Jobs, like the phraseiology he used for computers back in the day was it was a bic bicycle for the mind. Um we're just now seeing more of an unlocking of what that means. And just because we had other uh physical mediums in order to make said thing happen written uh drawing camera does not mean necessarily that just because it's removing the need for the camera being used does the eye of photographer not need to exist like I yeah you can't say like no like you or I mean I work in

media so maybe not the same thing but I and you as well but let's say somebody who's your average everyday person says, "I want a nicel looking picture for this." Okay, the AI is not going to know what to do with that. It's going to take somebody who says, "Have it look like this at this aperture, um, this angle, sun coming from this direction cuz that's they're just doing the same thing just with a camera." >> Yeah, that's right. And it'll get you closer. Then the ability to describe the emotion you're looking for, the blending of colors, understanding color science, the color wheel, how to talk to it, it can produce it, but it still needs someone just like you would a camera that sets up your lighting, finds a good frame, decides what this subject matter is going to be, and to click that.

That clicking of the button is like the least important part of taking the picture, right? It's understanding all the settings. It's just a different. So then he came back very excited going I just I can now take pictures anywhere in the world without leaving my chair and I can still give the world my art. That's the optimism of AI. Right? If you think about what the industrial age gave to us and it it sort of while there was some attrition, it removed the burden of labor from humanity and allowed machines and automation to do that. As we move forward into AI, it's removing repeatable excellence burden that I have to do the same thing. I have to come up with the right answer and I have to repeat it over and over and over again. Because that's what our schools do today. They beat creativity

out. If you've got any young child in your life, anything's possible. Not afraid of failure. And they will run at trying to do anything. And every And that's right. And then they go to school and they learn that 2 plus two has to be four all the time. you have to repeat it and memorize it and then you reach a level in even grade school of honors and status and straight A's and then you stop taking risks because you don't want to lose the status because if you take a risk and you don't get that A you lose your honor status what a terrible way to teach kids my hope and dream is that AI takes that burden from us and we allow humanity to learn and evolve all the way through the grade schools with fearlessness on making mistakes. Make them. Learn from them.

I don't know about you, I've not learned a single thing from a success I've ever done in my life. If I walked in and it just was easy and I did it, I learned nothing. I I still carry stories with me of things that I messed up for 30 years. That's where you learn stuff. And if Yeah. And if you're too afraid to try, you're too afraid to fail, you're not going to learn anything. So the brightest future for me is that AI takes that burden, enables us to try to be fearless in creativity and we usher in a different age just like manufacturing. No one's no one is complaining or saying that an assembly line was the wrong way to go. Nobody is. But we had to choose the next thing that everyone's going to work on. So, we can go down the apocalyptic

view where, you know, almost everyone's out of work and nobody does anything. Or we can take steps right now to build a curriculum that trains human beings on how to use a camera that doesn't look like it used to. It's a prompt that generates the art we've always come to love and use, that can generate the outcome, command 200 agents to drive your business, and amplify. That's an infinite game. I can only scale to zero of people. That's that's a finite game. At some point I lose, right? But if I could amplify your power and enable you to do more and more and then I can go get another one of you and give that you 200 agents and they can do more. That's infinite. Now my scalability has no limits. That needs to be the way we're thinking about this. >> Yeah. Absolutely. No,

it's a good point. I the um Speaking of Infinite Game once again, Simon Synynic reference. I'm pretty sure that that was a book by him. Yeah. So accidental. Um yeah, he's a good Yeah, he's very good with what he does. So um interesting. By the way, how did you get into uh teaching? I was just curious. I don't >> by accident. I didn't think I'd love it to be honest. This is pretty much just about my whole life is I felt best at so good friend of mine, Matt Burcth. He's on NLP Logix. He's he's brilliant with AI and he was covering this course and he got two courses at the same time and he couldn't cover one and he says, "Hey, would you mind covering this?" I said like, "Yeah, I'll give it a shot." Fell in love with it. Absolutely fell in love

with it. >> Okay. Interesting. So, and and I guess this kind of leads us into the next thing, which is another form of teaching, which is uh bookw writing. So, you are um and and course uh creation. So, you uh have created a book and it is called people, places, and things. You referenced that earlier in the podcast. And my question for you is what was the inspiration um to obviously I understand the inspiration for what was in the book but what made you want to write a book cuz you went from that to also courses and it seems like you really are passionate about teaching this. >> Yeah. So I I started the book before I realized I was passionate about teaching anything. So maybe those converged at some point and I'm also an avid D and D player even to this day. So

when I went into teaching because I didn't I I didn't know how to teach really. So I gified the course and that turned out to be the best way to teach these topics is through gamification. So one of the course yeah I I wrote an entire game dock around the project management. I got I got the kids to take up and be their teams. They each had their roles. They had to write a vision statement and they had to put scope and they had to put you know assumptions and mitigations and those were blocking my attack roles on them of oh you just lost your your lead developer. What are you going to do? Right? Oh I've got a mitigation for that. I'm able to mitigate. So that it it teaches those principles but in a in a role playinging way. But from from the

book perspective, it all comes out of I've been on the vendor side for 35 years and I keep seeing the same customer mistakes over and over again. They they needed they have a digitalization project. They haven't thought through the vision. What's going to carry you through this pain? Right? A lot of people don't know. If you haven't been through one, you're thinking, I'm just upgrading software. I do it on my Windows all the time. there's a little bit of hesitation, but they don't understand the complexities across an enterprise of having to roll out an ERP. What do I need to do? Right? And it rarely are they prepared. They've not done their asis. Right? If I told you you're taking a trip to New York, but I didn't tell you where you're going to start it. How do you plan the trip? You don't know.

And that's the challenge is we'll come in and find out they don't really know what they have. They think they know at a high level surface, but they don't really know that there's some Excel spreadsheet or some access database rogue system that's holding their entire enterprise up and to find that along the discovery of of roll out is tragic. It's absolutely tragic. So I decided to just say look if I was the customer, how could I make this more successful? And so I I wrote a book that just covered the 30 it's 10 phases called the customer journey. It's 35 years of my experience of here's a story, here's why it's bad, here's what you should do. And it became an Amazon bestseller in a week, which speaks to the pain. Yeah. And it's it's a business book about rolling out ERP and it was

it was mostly 35 years of my painful experiences. What the problem is there's not enough meat in a 95 to 100page book for someone to use that and actually apply things. So I like so I got a lot of questions of okay well what would you have done before you bought what would you have done at this phase what would you so I just decided to sit down and write a course and before I knew it I was at 700 pages of course material what if specific details and and how to get through it and then I wanted to put on top of this the modernization of how can AI help you across all 10 of those phases right the big enterprise projects have hundreds of pages of artifact facts. No human being can know all of that. So things slip through the cracks, you

have mistakes. How can AI make sure that your functional spec matches your requirement? It can do that quite easily. It can do a functional alignment against what I asked for versus what the vendor sent me. If a vendor sent me a 400page document and said, "This meets your needs." Most human beings are going to read through that, but they're they're going to have a hard time checking the list about, "Well, does it really?" And then you find out deep in roll out, yeah, I can see how you thought it met my requirement, but you're missing the detail of what I absolutely needed. And it walks through that entire journey to try to eliminate that friction, get way ahead of it, and make sure because this is in the headlines. You know, there's a company in Germany just wrote $650 million down and never got their

ERP system, right? That's that's almost a billion dollars. People's jobs are gone after that. I guarantee it. Right? That's just try and it's all over the place and you get trapped in these five-year contracts paying for something that you can't use. And you know there there's a little bit of cross. So you know it could be the vendor's problem but a lot of it is you walked into it handed someone the keys to your house and said call me when you're done right and who no one's going to take care of your house better than you. So you have to you have to take ownership of it. But if you don't know how, I can see why it's attractive to just hand the keys over and ask for a call back cuz that's very alluring. So I don't have to do anything. I don't have

to know anything. I'm going to go back to my day job. And that's it's a tough reality. So the goal of this is to teach businesses doesn't matter what it is. Every business needs a piece of software. What are you about to go through? Whether you're losing $650 million or like my friend in the book John Leonard, which was trapped at 500,000 a year, but his business was small enough. He went through the it was the exact same 10 phases exact same mistakes just at a smaller scale but the scale was impactful to him at his business level right the dollar figure doesn't matter that is a heartbeat of your business no business runs without software today so here's how you're going to make an investment to get your ROI this is what you really need to know to make it successful and then here's

how AI can help you do it so you don't it's not you don't have to ingest 700 pages which is of a course material. You can you can take it a piece at a time. Each one of those phases, I'm either here already in a train wreck and I need to figure out how to get out or I'm about to start the journey and I can take it one phase at a time as I go through so I can prepare for the next phase, prepare for the next phase. So, it's written from the customer perspective. It doesn't have a lot of techno jargon. It has things that you could just cut and paste to use to get your answers out of it. And it's meant to allow you to do the core competency or whatever your business is. That's what you're great at. You're not

a software development company, but the knowledge in this course makes sure you get what you need and how to talk to your vendor, right? We were talking about before, don't don't tell them you need a checkbox on a screen because they're going to do it. They're going to put a checkbox on that screen and charge you for it. And you're not going to get what you want. You're not going to realize you just messed up all your reports. You're not going to realize that they can no longer integrate into your accounting system. You need to ask a different way to your vendor so they can be a consultant partner for you. But if you're declarative and you're asking for things, you're going to get what you wish for, brother, whether it's good or bad. If you ask it open-ended in a way that you're trying

to achieve an outcome and you're asking them to design it, that's a better way to go. Now they're partners. Now they're in it to win it with you, right? >> Um, no, absolutely. Absolutely. And I and I find every every kind of layer that you're talking about, whether it be in the business, um whether it be with teaching and whether well in the couple different ways you're teaching, seems like you're you're all in on this type of stuff, you right? And and I guess my my question for you is as somebody who has been in it for what is a couple years now and is teaching about it, how obviously you have your book and this is my attempt to plug you. um how would you recommend somebody to uh get going in this and and what type of people specifically do you think could

benefit from the book and the and the course right cuz obviously it could be anybody but I'm just curious if there's any specific like areas of people that you've noticed more benefit from it >> I I would say that those that had the most to lose the sea suite of most businesses or founders are finding this book very interesting the course material would mostly be valuable for your operational people the people you're expecting to execute ute on your vision for digitalization. So there's a mixed bag. Anyone that's been in the biz of running a business, have gone through a software implementation that you think is what the heck just happened? Um would probably enjoy the book because there's a lot of parallels in there. So at least you would learn it wasn't just me. Maybe there's some comfort there. Um or things you could have

done a little bit differently or oh I like the way he learned this. So it it's it's meant for businesses and the people in those businesses, right? So it's I I want to clarify that it's not written for a business. It's written for the people in the business. Those people are in that company because of a core competency or skill they have. They they unless they're in a software company, then they then they're they're a software vendor. But I've even seen software companies that build one type of software trip over the same things. they do this for a living for other people and then when they take it in themselves they make the same mistakes. Right? So I I think that there's a a large broad brush of anyone that has a vested interest in your company and at an operational level that are either

looking down the barrel of buying a piece of software and how to do it or are in the middle of it and need a lifeline. This book in the course world would save you. >> Awesome. Well, I appreciate that. I think it's um overall like I I find uh the way you're approaching this stuff really really interesting and I I think you have a lot of good solutions for many people. So my my last request to you would just be to plug what you want to plug. I think we know what it'll be. But plug what you want to plug and um we'll close it out. >> Yeah, no worries. Look, if if you're in a resource inensive industry and you're not sure how to tackle sustainability, give me a call. If you're in the middle of an ERP or digitalization, buy the book. It's

not that expensive. Read through it. If it resonates with you and you think the course could help you out, give me a call. Happy to help. It doesn't matter what it is, whether it's an accounting system, CRM, an operational ERP. Um, but what I don't want to hear or read about is a white paper where you wrote down millions of dollars on another failed ERP. I want that out of the headlines. >> That would be awesome. All right. Well, thank you so much, Evan. Appreciate your time. and everyone watching, make sure to leave that uh this episode a like on YouTube. Make sure to subscribe to us on Apple Podcasts and Spotify. Thanks so much for listening and we'll see you in the next one. >> Thank you, Demetri. [Music]