The $3.4 Trillion Problem & Why 70% of Transformations Fail | Michael Schank

Chris Fay: Welcome to Built with Purpose. I am Chris Fay, and this is the show where vision meets execution and leadership leaves a legacy. From design studios to innovation hubs and product lines to skylines, we meet the people building more than great companies. They're creating culture, driving impact, and shaping what's next. These are the stories and strategies behind how the world gets built with purpose. All right, everybody, welcome to Built with Purpose. I am your host. Chris Faye. Today with me is Michael Shank. ⁓ actually we were prepping a little bit beforehand and super excited to ⁓ probably completely ⁓ disrupt Michael's expectation of what we're gonna cover in the podcast today because ⁓ Michael is the founder of Insight Twin, an AI based ⁓ organization, as well as the author of Digital Transformation ⁓ Success. And so lots of things going on. So Michael, welcome to the show, my friend. I'm excited to have you on. I'm excited to be here. Thanks for having me, Chris. As we were ⁓ as we were prepping and kind of thinking about, you know, your background and all the all the disruption that's going on in the AI space. I mean, I just I know that's something we're gonna get into ⁓ pretty pretty deeply today, but just as a as a quick icebreaker, Mike, and I just I'm hitting you with this on the fly. I see a guitar in the background. Are you a guitarist? I I dabble. I know a few songs, not not much. It's more of a prop for when I when I talk to people 'cause a a lot of what I what I build with Insight Twin and I'll describe it more is It's a it's a tool for collecting data about how an organization runs. But I I heard this good quote once that, you know, like this is a this is a tool, but it's only useful if you know how to play it. So a a fool with a tool is still a fool. Love it. So you've you've you've got to have the technology, but you've got to also have the the know-how and the processes and the people that that can make it effective. Awesome. So it's a little bit of a prop. Well, I I would say my guitar is ⁓ is also a prop that I have at the house. I I used to play in ⁓ in School and it's interesting, it cop it probably describes, and you would probably be able to peg how my mind works is that ⁓ I I learn the guitar up to a certain level and then I just hit a complete ceiling. Like I I can if you showed me how to do something, I can do it. But then when you ask me to like freestyle on the guitar, my head explodes and I just and I stop. So about I haven't played in years, and so like six months ago, I was like, I'm gonna start playing guitar again. So I get my guitar out of the attic, I get it all tuned up, change the strings out and everything. And I've had it, you know, my fingers start to hurt, and then I just put it down and it's been down for like four months now. So it's just it's a it's a bucket list item for me. So at some point I'm gonna I'm gonna get proficient enough. Like not like I'm not gonna play any bands, but I don't know, if I could set at a campfire and and you know do eight songs or something like that, that'd be that'd be awesome. Yeah, right, right. I think that's the I remember I was at my wife's, we had a ⁓ like a family get together for Christmas when Time and part of us her uncle and a few of her cousins are like really musically inclined, like just voice, piano, they can pick up any, any, anything. And so it was early in our relationship. And ⁓ so we get to the family holiday, whatever, and ⁓ and they're getting ready to start singing like Christmas carols, and somebody says, Well, Chris plays guitar, and they're like, ⁓ you play guitar. Well, do you want to play with us? And I was I I don't, no, I don't, but they're like, ⁓ come on, you know, you're and I'm the new guy. guys so I can't say no. So I go back and get the guitar and we come out and they're like, all right, do you know so and so? I'm like, nope. ⁓ well how about can you play XYZ? I was like, I I I can't. Do you have like some chords I can play? And they're like, yeah, try this. I tried one song and I was like, nah, I'm done. Sorry. The new guy, the new guy's out. New guy's out. ⁓ Michael, give us give us a I mean I lots of stuff I want to talk to you about, but just in general, like just let's kind of frame out like give us the 30,000 foot view, man, like who you are, what you what you've been doing, and then give us some some ⁓ insights into Insight Twin and what you have to play. So a little bit of background. So I started I've really been in my whole my whole career has been in consulting and specifically in the financial services industry. I started 13 years at Accenture where I was a deep technologist. I was a Java C programmer like solution architect led large system integration projects then i went to bank of america for two years deep technology again i i owned some systems like the whole development team production sport etc but my my purview was just is the code delivered on time is it defect free is it production stable like i knew nothing about the business so then i then i move over to ey where i spent 10 years and at that point they didn't have a deep technology practice. So I s had to start kind of getting more into business side stuff. One of the first projects I did was at a a large, large bank and I think it was one of the end of the year type things. They had $500,000 burning a hole in their pocket. They're like, if we don't spend it this year, then we may not get it in budget for next year. So we they're like, ⁓ just build us a business capability model, which if you're not familiar with is just a model of everything that that business is able to do. ⁓ so we interviewed some people, created this this thing that was like three levels deep. ⁓ but it was like a epiphany moment for me. So a guy who is like deep technology and has no clue what the business does, to like all all of a sudden see everything on one page. I'm like, this just seems so powerful for for bringing people together, for ⁓ delivering things. More efficiently, et cetera, et cetera. So I kind of like put my stamp on it at a UI. I'm like, this is this is my thing. I'm I'm gonna lead this thing. So did a did a bunch of projects, started a process excellence practice because I realized that business capability, and I might be getting into the weeds a little bit, doesn't go deep enough that you need to know like what processes are actually happening on the ground. So ⁓ just did a bunch of projects ⁓ with with cus ⁓ with clients, transformation strategy, risk and compliance, you name it. But every every time I applied this framework to it, it made things much more organized, much more straightforward. We hit our deadlines, everything was typically ⁓ bright green. Well and that's that's sorry to interrupt you. I that's one thing as you're saying that I'm just curious, like, you know, of all the firms that you were working with from a customer perspective, was like there a common theme? Like when you came in and were like, okay, let's talk about what you do and how you do it. Was anybody like on their game and had their stuff figured out? Or was everybody like, ⁓ my gosh, nobody's really ever asked me. Like was it a common denominator across these firms when you started ⁓ implementing this? Well, yeah, ⁓ have their stuff together. It's probably probably good to define it. Like obviously people know how the business works, but the problem is it's in like organizational knowledge is in pockets. Mm-hmm. Like, you know, y everyone's been in an organization where, you know, we can't let Joe play the lottery because if Joe leaves, like who knows what the hell's gonna happen, right? So but it's not so they don't have it like documented comprehensively and in act. And that's that's kind of what what I do. ⁓ because then if you're if you're driving like say an SAP implementation, well you got different players within that project. Like you'll have the business who knows what they want, you have technology who needs to know how to configure it, you have testers, you have risk people, et cetera. And if they don't have a map of the world, then things become very expensive, very long. There's a lot of discussions back and forth where there's misinterpretations, etc. ⁓ so it's a way to kind of smooth all of that out. Gotcha. Interesting. So left EY, then I went to Citibank and their context there was City was in twenty twenty was hit with a four hundred million dollar consent order. So basically the regulators coming in saying, your control environment's a mess. And they specifically pointed to their risk data as being a problem. So they started this massive ⁓ consent order transformation program to fix it. And a lot of that work is I was inventoring all the processes for the the US retail business and then taking their risk and control data and mapping it on top. And it was very valuable to see where they had gaps, overlaps, where people were like, I have no idea what this risk. even is ⁓ I don't know who owns this so it was a way to just clean up their risk data. ⁓ so I left there to write my book. I'm like I I need to I need to get this out there. ⁓ and I've been doing consulting ⁓ with various clients who are who want to implement this framework. And then most recently I've I started late last year but I'm building an AI digital twin organization platform. And essentially what that is is AI is trained on publicly available data, ⁓ but it knows nothing about any specific organization and how it operates. So, in order to bridge that, like if you want to s truly scale AI, you need that data infrastructure. So AI knows knows the ins and outs of of everything you do. ⁓ and that's essentially what it is is ⁓ I build that inventory of processes, ⁓ Marry it with some key other operational data. Like every organization has operational data, say your system repository and your service now CMDB, your RISC data in a GRC or RISC repository, your org enroll data in an HR platform, performance data in various other places. But if you could take that data, migrate it into my platform into libraries. Now you could do the association between what the process is. And the data that describes it or the resources that it's using. ⁓ and then that's all kind of in a structured relational database. And then I have AI on top ⁓ through an MCP interface. So right now I could do anything that you need to do internally to manage your organization. So if you're doing strategy, I could say, tell me my strengths and weaknesses so that I could identify my next set of projects for the upcoming year, so that I'm putting my funding so my funding. Yeah, I think probably I would think a lot of companies, you know, and certainly you're you're probably working more at the enterprise level, but you know, kind of in the mid-market SB level, a lot of folks are trying to home grow their own thing. So, you know, I'm gonna MCP Claude into SharePoint and into Salesforce or into Net Suite, and so it has context of what I'm doing and try and you know bring all that information together. But I it's still it's still a bit homegrown and and has its ⁓ has to be self-down. Self-maintained and if you're missing information and and so on and so forth. I can easily see how that would be about, especially for a large organization. Yeah, and and I talked actually talked to a large bank. ⁓ I have a friend who actually showed me a tool they built there, and it was it was pretty impressive for what they had. They have this application they built, and there's a drop-down box. You could pick Opus or any other, there's a bunch of LLMs in there. ⁓ but they had connected to it their their RISC data, their system data, their JIRA, their confluence, like they had a bunch of data sources. Well okay, I'm impressed. One thing that they're missing though is the the semantic layer. So what's the things that tie tie it together because how you describe require requirements in JIRA may be completely different from how you describe your risk and your risk platform. Right. How do you know that those are the same thing? Yep. That's is that ontology? Because I I I met with somebody yesterday. Ontology, yeah. Canonical, canonical and ontology. We had a big conversation about that, ⁓ about that yesterday in a big AI workshop. So exactly. Yeah. Well, it's and it's interesting, you know, one of things I I'm gonna ⁓ come back to your book because I have a question about just writing the book, but I do want to just talk about that kind of that build versus buy ⁓ methodology, because that's something I think a lot of companies are are challenged with today. You know, the the vibe coding experience, should I just build my own you know ⁓ layer or should I do things myself but but first off going back to so the book I'm curious just individually did you did you always have this in the back of your mind were you like gosh I I I like I really want to write a book or did it kind of hit you one day like what what was that process like just in in identifying you wanted to write a book then you wrote it and then you got it published like just as a as a human wanting to do that like what was that process like like I want I never was like in my twenties saying I want to be an author. It was more I developed the framework, I ⁓ deployed it to a lot of clients, a lot of different organizations at work. I'm like, holy crap, like the world needs to know about this. Yeah. So ⁓ I it's the like when I was at I think UI, I was I was like, I need to do this. So I started kind of outlining the chapters, but I didn't really do anything with it because I was I was kind of busy. ⁓ but then when I left ⁓ city, I was like, all right, I gotta make this happen. So I put a lot more time into researching it. And there's a a good website, ⁓ and it it's it's escaping me right now, but it's ⁓ basically a website that's a a resource. For authors. It tells you through like every step of the publishing process. And basically, there's a three step process for publishing a nonfiction book. So you you create a proposal ⁓ which lays out like your table of contents, sample chapters, competitive titles, like there's a bunch of stuff in there. You send it to an editor to clean it up, and then that editor will send it to an agent. ⁓ I actually kind of lucked out, so I I wrote the the proposal. proposal, sent it to a bunch of editors, and one of the guys is like, hey, I'm a acquiring editor for publishing house. We want to we want to publish your book. I'm like, sweet. Yeah. So ⁓ yeah that that was fun. And then you know he's like what's the timeline? I'm like I've never written a book. I have no idea. Yeah why don't you tell me right? Yeah. So I I said it and ⁓ yeah it was it was a lot of work. Yeah. Was it fulfill fulfilling though? I mean I it's probably more work than you're expecting, but I mean was it was it good to kind of see the final product and to get it out there and ⁓ yeah. ⁓ the the the most exciting point was I when I looked on Amazon and saw it, I was like, Wow, I'm on Amazon. I'm on Amazon, baby. Yeah, that's awesome. And what and what so what's like at a high level, what's the core foundation of the of the book? I mean digital transformation success. I can hedge on some of it, but basically it's it's in three parts. One is the first part is just description of well the the case of why this is needed. And and I always use a stat. ⁓ so every year companies spend $3.4 trillion globally trying to transform themselves. ⁓ yet 70% of those transformations will will fail. So I frame it as it's it's a big problem that everyone has. so the first first part is the the framing of the problem and why this is needed, and then just what is what is the ⁓ the framework. So I described kind of the nitty-gritty kind of ⁓ technical aspects of it. Then the middle section is all about how it can be used. So what are the specific use cases? Yeah. And there's six broad ones, but it's strategy definition, transformation, ⁓ change management. So how do you do automate your SDLC? through AI, ⁓ operational excellence, looking for waste inefficiencies, ⁓ automating through a gentic AI, et cetera. ⁓ The fifth is is IT. How do you design, manage your IT environment so that it's agile, more closely aligns with the business? And the last one is risk and compliance. So how do you know what you what laws, regs you have to comply with, what op risks, etc., and make sure you have that strong control environment. So what so what's the ⁓ like looking looking back at your experience there? I mean, just thinking about large companies, small companies, so this trillion dollar ⁓ spend on transformation and 70% fail, like what what's the number one comma denominator? Like if you just have to sum it up, what's where where do most companies fail? Where are the seventy percent normally failing as it relates to transforming their business? ⁓ it's just it's complexity and chaos. Okay. Okay. And everyone I talk to, you know, when I when I talk to crowds that are in this world, like everyone has a story to tell. Of just something that just went sideways. ⁓ but but basically it's when you when you're trying to run something big, you have so many different practitioner types in there. You have business people and you have technology people and data people and risk people, testing people, etc. And they all have to come together and and understand kind of what the current state of the world is, like at a broad and a detailed level, and then come together and say, how do we need to change this? And To to meet our needs. And if you think about a transformation, like a transformation is a portfolio of projects that's typically multi-year tens or hundreds of millions of dollars. Like there's so much coordination that has to happen. But each of those different practitioner types have a different understanding of the organization, have a different perspective, and even have a different language. Like when I was in technology, like the business would would use different terms that I didn't really understand. I'm like, just tell me what I need to build. Right, right, right. So so how do you how do you bring all of that together is a challenge and and a lot of organizations they either hire outside consultants to help do it, ⁓ or if they do it themselves, but they don't have like the the core skill set in driving large change. And that's that's what what gets them in trouble. I yeah, it it it seems like at least the the experience that that I've had also is it's it's always, and I think maybe this is what you you said a bit ago, it just it's it's more complex, it's going to take more time, it's going to take more resources, likely more capital, you know, deployed than you probably expect. You know, and I think even if it's a small scale, you know, ERP implementation or full scale Transformation of the business. I just to your point, you know, that complexity, all the people in the room, unless you have a really good guide, right, or consultant that's really helping you and you trust the process, right, to follow that consultant. It just seems like I know even from our own experience, like we've just always underestimated what's going to need to happen ⁓ in order to successfully get this across the finish line. Yeah. I I I've been using a lot this ⁓ just to make it real, this ⁓ example of Zimmer Biomet. They're a medical Device manufacturer. They paid Deloitte a total, and there was a bunch of change orders in there $94 million to implement SAP. So they were taking nine existing ERP platforms and consolidating it. The GoLive was such a disaster. They couldn't ship products, they couldn't generate invoices. Basically, the company was shut down. They lost $2 billion in market cap, and now they're suing Deloitte for $172 million in damages. So, like that is that's an extreme, but that's that happens at many different Different levels of scale across organizations all the time. When you're a company, you can imagine spending a hundred million dollars, right? To deploy and that I, you know, and obviously that to deploy and consolidate all this, and then the amount of time and energy and effort that went into that for then it to not be successful, right? It's not like a six-month rollout. This thing probably took years to deploy. And ⁓ know, you just think about the business disruption. And so it's like that's that's one of the things I kind of think about is it's just you know, for people. People out there, anybody that's listening and thinking about transforming their their business is just making sure you are well prepared for the level of effort and and ⁓ that it's gonna take and capital and time to truly transform. And I think to your point also is probably just making sure you find the right partner. You know, there's just there's so much noise out there of people that have expertise and can do different things, but but ultimately you're hiring somebody to guide you through the process of of how to transform your business. And so that you've got a major reliance on that partner. To ultimately guide you to success. And there's there's risk in that too. And this happens in the in the SAP or the ERP world all the time. You hire someone like a Deloitte or or whoever, and then you realize like just time keeps slipping and they keep asking for more change orders and more money, but you're so stuck into it and there's such a a knowledge gap between what they understand and what you know that you can't switch horses. You just have to say yes to the change orders. Right. Yeah. Because if you say if you say no, then the project stops and you know And then you you have to go through your your vendor ⁓ selection process again and really start over. So how all right, so so let's I'm gonna pivot you a little bit to think about I had I had two questions, but I might reframe my question for you in ⁓ like how does all that change In the future. I think about like AI. My my first question for you was kind of like go back to your Accenture EY days. Like if you had the AI tool sets on at your fingertips that you do today, but back when you were working in these large organizations, like how fundamentally different would it be? And then my second question, kind of diving into that, is like, you know. If you put yourself in the hands of the customer, the customer is is expecting, like you want a partner that's gonna guide you to success. You know, you don't wanna be beholden to their change orders and their their hourly pricing and hourly billing. Like you want them to take some risk, you want them to be leveraging technology and AI to accelerate implementation and deploy best practices. So you know, maybe the first thing just kind of a general reflection, like what would your world have looked like had you have had the AI tool sets ⁓ that are available today? ⁓ it things would go so much faster. Yeah. But but I think there's still a lot of room to grow and that's that's why I built Insight Twin. ⁓ but the just just even like, you know, a vibe coding. Like You know, I would have to hand code everything that I did. And now now it's a state where you could have clawed code or whatever agent you're using, create the code, ⁓ but you have to still be on top of it and make sure that what they're doing is right. ⁓ 'cause I think humans have to be accountable. But but ⁓ where I think so things will go faster, but I think where things could really take a leap, and this is this is kind of the the the insight to its story is You still need that data infrastructure because AI, as I mentioned before, is trained on publicly available data and doesn't know, like in the zire in the Zimmer Biomet instance, doesn't know about the nine ERPs, what those specifically do, where they're deployed, what what people leverage ⁓ use them, etc. So if you give it that data infrastructure, now AI can could really run the entire program. ⁓ so th this is something I I it's on my roadmap. But like the the entire SDLC could be in a long running work workflow. And then you could have specific agents that do things like generate my scope, create my requirements, do my design, do my test plans, do my people change management plans, ⁓ all within a within a workflow that ⁓ gives you status kind of automatically. But then you do have, I'm a big believer in this, you need human and loop control. Just because it could create a scope document doesn't mean anything until there's a ⁓ human account. accountable that has to sign off on that and say yes I agree with this then now that could be used for the basis for everything else that it's dependent on so do you do you think I mean just philosophically because this is something I talked to a lot of our guests about and I spent a lot of time you know researching and and trying to understand this I mean when you think about Accenture you think about EY just you know we've got a consulting organization in our business you know you do consulting as well like is there is there a massive disruption and reduction in in workforce right There now there might be short-term disruption, but like long-term, is there a massive disruption to attorneys, to consultants, to these labor-based models because AI is able to augment, or does it do what technology has done in the past to us and over time it just enables us to solve more problems faster? And you know, we actually might be able to grow that industry because now we're able to do things at a more rapid at a rapid rate. I think pot both are potentially true. On the on the first part and the disruption, the way consulting works, it's ⁓ every engagement's a pyramid. So you have the the senior or the partner on top, you know, senior managers, and then you have the the bottom of the pyramid is the the analysts and they say they're two to four years out of college or whatever. That's where you make the margin because their salary is is low enough where if you say double the bill rate, then you're you know, you're getting a lot of margin. But what traditionally what that that bottom layer has been doing is been kind of the groundwork. Like they they create spreadsheets, they create decks, et cetera, et cetera. But now with like Claude Cowork and some of these other tools, like I Claude does my decks. It can create spreadsheets for me. So is that layer needed? So if you start picking this ⁓ picking off like those more junior people, now the entire kind of margin model like breaks apart. So I think that's where they're str and and I think a lot of customers are like, why am I paying, you know, a hundred and ten dollars an hour for this kid two years out of college when I could just have my people do it. Yeah. ⁓ so I think that's that's a root of and and it's been documented like the I think the the people most at risk are the ⁓ are the younger kids that just don't have the experience, don't have the knowledge yet. ⁓ so I think that's disrupting a lot of it. I think there's a prior to the show we we we were talking about about but can you can you take some of that greater knowledge and kind of force it down the pyramid? I think that's happening. Like I read an article that ⁓ the CEO of McKinsey said they have 60,000 employees and 25,000 of our A are AI. ⁓ And if you think about it McKinsey they have a lot of frameworks they have a lot of like a lot of assets that they deploy to clients and it I'm sure and I don't know this intimately that they've taken all that knowledge and put it in kind of one data set that That AI has access to. And if that bottom of the pyramid now has access to that knowledge through AI and can provide a higher level service, then that's probably the future for them. Yeah. Yeah, and I think that's kind of like in our warm-up there. We were talking about that. And you just you kind of philosophically think about it, like, okay, if AI takes the low grunt work tasks and wipes out kind of that entry-level ⁓ job to say, then what happens in 10 years? How do how do entry-level people become more senior level people ⁓ if if some of those tasks have been put away? But then by the same token, so so then you have kind of a very top-heavy organization. You know, top heavy, expensive, experienced individuals with with kind of that bottom layer of of AI. But then you could also think about it completely in in reverse, is you could also use AI to be the knowledge, the experience, the best practices, and you might actually be able to leverage ⁓ in empower a more junior level resource. with a more experienced AI agent. So it'll be, you know, and we're talking about this in a in our practice as we think about how we continue to to to grow and and how we service customers is you immediately go to, okay, well, the the the button pushing, the data consolidation, that's now being done by AI. What's gonna happen to these junior level resources? But then you're like, well, wait a second, what if I document all the industry best practices and context that I know and then enable those people to maybe step up and perform at a higher level. So it's gonna be really interesting to kind of See how the staffing model shifts. And I think, you know, fundamentally, you know, if you kind of follow the follow what history has told us, is that you know, this should be able to accelerate our ability to solve problems, but you know, there's definitely going to be some impact on the on the labor pool. I mean, I'm talking to you know, kids coming out of college and you know, in marketing and and you know, you know, entry-level business and stuff, and they're they're struggling to find jobs, you know, because I think every business is trying to figure out what this economy. system is going to look like. And so I talked to an attorney that has a, you know, I'm I I don't know, ⁓ I want to say they're like a thousand attorney, two thousand attorney, big, big, big company. And, you know, they said they're they're reducing ⁓ by between between 10 to 20 percent their annual entry level attorney hiring, you know, and you just generally think if that if that is long term based or short term based, I mean that's a massive disruption to a lot of these kids come out of college. That's a real risk. I have a eighteen year old and a twenty year old. So yeah, you know, I worry about them for that. ⁓ yeah, the advice I give them is just learn how to how to work with AI. Yeah. ⁓ because it's ⁓ You know, it's it's more than just prompting it for like, hey, help me write this email. But like really understand how to how to work with data, how to how to ⁓ set up your structure so that you can really leverage it. ⁓ but yeah, it it you know, I think it's a real risk is like what happens ten years from now when those senior people retire or move on and then there's not that next layer that knows. Yeah. ⁓ but but I think you're hitting on a an important point. It's like it's all about data. Like Even that that knowledge, like getting all that knowledge together requires a structure so that it it makes coherent sense as a whole. ⁓ but then you could dive into specifics. Yep. And that's kind of what what we're doing with Insight Twin is, you know, a lot a lot of these senior people like in a consulting engagement, like you got a guy who like really knows payments or really knows finance or whatever. But if you could document the finance environment. But then those junior or the payments environment, then those junior people could know just as much as they do. Right. Yeah, and it's it brings up a great point. I can kind of maybe a pivot into Insight Twin is you know the the AI is as good as the context that you that you give it, like the information that it has access to. And I think you know, as as humans, we're we start seeing, ⁓ my gosh, look what Claude or Gemini or you know, OpenAI, like look what it can do for me, look how fast it's working. It must Must be right. It must know everything. I'm going to place all of my trust and faith into this, into this tool set. And I'm going to, you know, it's going to guide my entire life and my entire business. And like I'm seeing like my kids, I mean, I got 15, 17-year-old. I mean, you know, they're placing a lot of faith, you know, in it. I have friends diagnosing their own, you know, whatevers based upon what Claude says. And I think like as a society, we're putting so much ⁓ individual ⁓ weight on what AI tells us. And even though it's a great guide for us and it's a it's a but it's it's not always ⁓ as accurate as we all ⁓ want it to be or think it's gonna be and so you think about that probably in in business context is like we have a lot of proprietary stuff inside of our company we we do things a little differently from time to time and so you know it has to have that that underlying context and so you know if you're implementing AI in your business and this is where I I'm curious to kind of learn more about about Insight Twin but you know if if I'm just using Claude and it doesn't have context or knowledge of you know our ERP or our our Salesforce instance or you know how we do things or why we do things or what our employees look like or skill sets and others, you know, it's just giving me the same advice it would give any organization. And so that's probably I think where I'm kind of seeing where where you're going probably with Insight Twin is yeah use the public data that's great but but what's more important is to ground it in in your own data, your own processes, how you run the organization. That's right. And and it's ⁓ so Insight Insight twin is actually ⁓ like a SaaS platform, obviously it's technology. But I think a bigger part of it, and it goes to like the a fool with a tool is still a fool, is you and and this I think this is an opportunity for the consulting firms too, is the data just doesn't appear. It takes a concerted effort to to think about how am I going to structure this data, how am I going to collect it from the various sources, how am I going to get the right people that need to say that that what I assembled is correct is is accurate. ⁓ You need governance on it because that data will can will continue to change. So you need processes to say, how do I revalidate what I have on a periodic basis? Or as the data changed, where I have hooks that say, you know, update the data. But that's that's a bigger part of it. So Insight Twin is a technology, but it's also a playbook of like what what is it you need to do to make sure that you could trust this context data so that so that it's giving you those more accurate responses. It's also You know, th I think the the accountability model needs to shift in some of these organizations too, where You know, a lot of the doer tasks are are, you know, especially at the junior level, are gonna are gonna go away. But I think it turns more into like an ownership model where you break down the org into specific areas of ownership. And then those areas of ownership have to be responsible for making sure that the data that's captured in that context repository is accurate. ⁓ and that when when you generate new knowledge based on that data, that they're the ones signing off on it, that that that that's accurate. Yeah. It's it's funny you say that 'cause we're going through a process right now, you know, where You know, you've got these these various people, right, that that report to different leaders, right? Yeah. But they have common responsibility of certain ⁓ either processes or ⁓ kind of pods or center of excellences. So like for us, like integration is is one. So we've got we've got developers on on one side that know integration, you've got kind of the the marketing business owner on a on another, you've got the consulting product. I mean, you have all these various skill sets that are across the organization. and they all report up to different ⁓ individuals, but they're all accountable to that that one kind of pod. And so what what we just what we've not done like in the past two weeks is to is to really formally establish, and I'd be curious for your advice on this, because you you've obviously dealt with much larger organizations than than we are, but you know, is is when you said ownership, I immediately that brought me back to what we're doing is we can have this cross-functional ⁓ aspect of what we're doing, but who who owns it? And then who owns that center of excellence? Who is responsible for for making sure that that connected process works? And then how do we then begin to build context and best practices and you know, get get the thoughts from this person or this team into this person's hands and eventually just establish what would be a common framework for deployment. It's it seems like kind of what you're a bit what you're saying. That's exactly what I'm doing. So what my model is is I follow the the org hierarchy. ⁓ so you have a center ⁓ integration center of excellence. Well that reports to a leader. And what what my process right now it's a it's because ⁓ the challenge is a lot of this knowledge is in people's heads, so you have to talk to them. So it's an interview process. And I start at the top of the hierarchy and say, describe what your organization does. But if I went to the leader of the COE for the integration, it'd be like, you know, document interview. integration requirements or you know test integration, build integrations, whatever those processes are. ⁓ but then you know I keep walking down that hierarchy until I get all the processes and then I go back up the hierarchy and have people give me formal approval that what I captured is accurate and complete. Because I want that on record. So no you know if someone says well this is wrong but well you told me the data ⁓ of of what you do. ⁓ But yeah, then then ⁓ Yeah, then then once you have that. ⁓ the the other thing I wanted to say is that's kind of at an organizational level, but then there's kind of a cross-organizational process as well. So the center of excellence is obviously one part of developing or ⁓ delivering change. There's probably business creating requirements in another org, there's a maybe a testing team in another org, etc. You want to create something like a value stream that connects those individual organization processes together. could see where all the handoffs are. And you could have an owner for that. So who is responsible for our our change process? And then they have to do the coordination with all the different bits, but like but at ⁓ but now they're armed with specific details of what those processes are, who owns them, and maybe some details about how they how they work. Yeah, interesting. And then how so we we'd kind of played around with this topic earlier. So, you know, I kind of in this build verse buy methodology is maybe something I want to kind of dive into. So you built a platform, you know, it's it's it's it's probably purpose built for the industries and the types of organizations that you work with, and you know, you start deploying best practices that you're learning across other areas. But you mentioned earlier, like there's companies out there that are just funding their own vibe-coded, you know, solution to connect these dots and and know, it's it's a conversation we have a lot with customers around this kind of build versus buy methodology, you know, and and you've always had the ability to build your own technology. You just it now you can just do it a lot faster, you know, and a lot less expensively because of because of AI. But the the common the the root reason why you typically don't build your own technology is you're gonna have to maintain it. It's it's only proprietary to what you do. There's no peer learning industry best practices. You know, you've got to keep the integrity, the updates, the data security. And so I think, you know, in the past when a lot of companies looked at this build first buy, It wasn't that it was going to be so expensive to build. It's just that they don't have to maintain it and manage it and keep the upgrades. And I think a lot of companies, you know, I'm thinking probably for customers of yours, you know, they're they're not hiring you to just repeat what they already do today. They're hiring you because you probably have insights and best practices that you can bring to deploy to their business to ultimately help. So I'm kind of curious what your thoughts are just on this like build verse buy methodology in the new in the new age of AI. It gets into the I'm sure you've heard the term SaaSpocalypse. ⁓ yeah. ⁓ we talk about it all the time. Yeah. I think I think they the the distinction is like f if you're a SaaS company, can you be vibecoded away? Or are you infrastructure for for AI? and a lot of these companies are I you know at risk because you know I you know I I I I started playing around with a free instance of HubSpot and I tried to do something very simple and I couldn't figure it out and I couldn't figure out if it's you know the licensing level or whatever, but I didn't do it, but I was like, I should just create my own because I know what I want. This is easy. Yep. So like companies like that are at risk. Whereas whereas Insight Twin I think is, you know, it is building the infrastructure that AI sits on and there's a whole process around collecting knowledge and organizing that. So I think it's a little bit it's a it's a little bit safer from that perspective. Not immune, but but a little bit safer. ⁓ but yeah I think that's that's a real thing. I think a lot of companies now have to make a decision like, do I buy this or or do I build it? And when you say the support costs, yeah, they don't want to do it. But if the per seat licensing model is such that it's cheaper to e even build, you know, s stand up your own support organization, then then maybe that's the the route you go go down. Yeah. But it's it's interesting you say that and just you start thinking about it. Obviously we're in the SaaS space and you start thinking about, you know, where where are we at risk and what value are we adding To the customer base, and you know, how are we innovating at a faster pace that might be able to ⁓ outpace what a customer could vibe code themselves, right? Because certainly if if we're static and not innovating at a fast pace, a customer could very easily, with enough money, you know, vibe code some level of connected ⁓ solution. So it's been it's been an interesting self-reflection for us to say, you know, are we moving fast enough? Are we solving problems? Are we making Making our customers better at a faster pace than they could do themselves. And then I love what you said is are we connecting, are we connecting the dots? Can everything that we that we do ultimately begin to provide deeper context ⁓ for AI and workflows and agents? But yeah, I think about a lot of tool sets that that are out there and and I know a lot of people in the space, you know, and and you know, companies that started in the past handful of years, you know, maybe a couple million bucks or so in in ARR or or higher and yeah, you know, that that that are seriously at at risk and and and so are the big so are the big folks. You know, I mean the big Well i and and there's another angle to it too is the the per seat license model. ⁓ like what if that the What if the user's not a human? Like can you still charge that per seat license model? Like there's a lot of things that they need to that the whole industry, SaaS industry needs to figure out. Yeah. Yeah. There's a whole you know, we're there's a lot of c conversation around consumption modeling and then obviously outcome based modeling and you know, ultimately creating employees that you that you charge out. You know, a a SaaS company like us would create employees that we could charge out like we would a traditional labor pool, but it's skilled and trained in certain workflows and yeah, it's it's I you I've I've never you know, I'm 45 and you know I've been in the in the space for you know 20 something years and and but still it's just the the pace of change is is you know and the noise and the FUD and the you know the experimentation is just like it's so exciting but it is just so scary as well you know and and you're just waiting for somebody to announce something major every day and what's gonna happen here and how are we gonna do this and you know I think the thing that we keep telling ourselves is we we have to we have to pay attention to the market. We have to understand what's going on, but you have to be careful about allowing all the noise to to derail your efforts or fragment fragment your efforts. You know, there's kind of stay focused on your vision and mission. What are we doing? And you know, think of this thing as it's a an enabler, but you need to be smart about how you leverage it. Yeah. I don't know if you saw that there was an article that Microsoft like deployed ⁓ claude code to all their developers and then they looked at their token costs and they started like pulling a lot of them. We were like, wow. Yeah, you guys are killing us. Yeah. ⁓ it's amazing. I mean, you you put it in people's hands and I mean we do it in our organization too. You just you wanna you wanna give people the tool sets, but then you're like, holy cow, like this thing starts really running up. And you know, if you keep your your your labor model the same and you also put productivity tools on place, you gotta really, really watch it. But I just I think we're at a stage of still, you know, it's becoming a little bit more mainstream, but it's still just rapid experimentation. I mean, there's people that are at at different levels of of technology adoption and We've got people in our company that are just still trying to figure out where it fits. ⁓ we have people in our company that are high level power users doing amazing, amazing things. And you know, the challenge that we have for our company, and I would just say for anybody that that we talk to, listeners or whatever, is you just gotta get started. You gotta get you gotta get your head in the game. And just to your your 18 year old and 22 year old, like look. Be enabled. Learn, get trained, get skilled, play, practice, you know, do that because that will increase your chances of ⁓ of success. And you know what? If if AI wipes us all out and we're just run by robots and and computers and we all gotta just live on the farm, then we got bigger problems, right? So but but you know, there's another stat out there. ⁓ MIT has a report that says that ninety five ninety-five percent of AI initiatives ⁓ show no return. And I think they're really Specifically talking about ⁓ agentic AI workflows. So while it is there's a a lot of excitement, there's a lot of like movement. I think people the the problem with like an agentic workflow is it looks great in a pilot, it wows people, but then they hit a wall before they they move it in the production. That's because they just haven't solved the or really thought through the what's the interaction model between machines and humans? Because you can't ⁓ you know I believe you can't make AI accountable. Like like if it if it makes a mistake, which it will, you're not gonna be like, Claude you did a bad job. I'm so mad at you. Like it's gonna like whatever what's the next prompt? Yeah, right, right. Yeah. Used to be able to used to be able to rough chat GPT up. You can't really rough Claude up. When I when I first started working with ⁓ ChatGPT, I realized that as I got more frustrated with it. it would, you know, it was, it would do a better job. And so I would kind of lean into like, you better not screw this up. You're really gonna make me mad. And it put like this whole level accountability, Claude doesn't care. Yeah, Claude doesn't care. Claude doesn't care. At the end of the day, it's a machine. It it doesn't have emotions. It doesn't care. Yeah. ⁓ there's a lot of people investing in AI, like at a at an enterprise level or a a business level, a lot of people investing in AI. But I think they're they're challenged with how do I actually make this thing a success? Mm-hmm. ⁓ and I still think it comes down to context data. Yep. Is what are you feeding it so it actually knows who you are? Right. And then I think to your point is how how do you make sure that it's accurate, right? So when it when it when it when it gathers insights for you that you're either human in the loop or you've got some QA QC agents, whatever it is, to to make sure because I've you know, even connecting to our systems, you ask Claude questions or context and it doesn't get it all right, right? So we've got a kind of that trust but verify ⁓ methodology. Well i and I saw this ⁓ like I read this BCG as this 10 20 70 rule, but basically Basically, they say 70% of why AI fails in organizations is because of people and processes. And I'm like, hmm. I would have said data, but then as I think about it, I'm like, well, that's true, because data doesn't just magically appear. It takes people and process to make sure it's it's in the right state. Right. So you know, I I I don't think we could come at this as this is a technology problem. It's a it's still at the end of the day, it's a people process ⁓ to problem to Get the data to enable the technology. Well, it probably goes back, I would just think in general, you know, the the Amazon-like experience, the instantaneous ⁓ desire for us to have like instant gratification, you still have to put the work in. And I I go back to your examples of these large consulting projects, is you know, when you hire the right partner and you dedicate the time and the resources and the investment to fund, you have a much higher likelihood of success. And it's kind of the same thing with AI. If I just just if I just hop in one day and try and vibe code a quick quick solution and then start using it as an enterprise solution, I'm probably gonna be, I'm probably gonna be met with dissatisfaction. But you know if you if you implement something VAI and making sure you have the right people and the processes and the context and the connectivity and maintenance schedule. And so I just think so many humans are just pre-programmed to this like easy button and it's just it's just not it's just not. Yeah, no. Everyone's looking for an easy button. They are, man. You know what it takes? I just I I just want to come in and be Hey I would just will you just you just do my job for me today and then I'll be back tomorrow morning and then it just magically happens. So I think that's that's some security for all of us to know that you know AI and robots are probably not going to take over the world anytime soon. We ⁓ there there's one feature that I'll that that I'll be working on this year that's not it's in not in place yet, but I mentioned the interview process to identify what process is e you know, following the org hierarchy what processes people do. Well I I think that could be ⁓ exchanged with a bot whereas you get an email saying hey Chris would like you to you know do this interview ⁓ and then a bot comes up and it's preloaded with you know what what we think your your processes are and then you're just interacting with the chat bot saying no I don't really do this change the name of this combine these two whatever it is yeah ⁓ but that that way you could I mean if you think about it it's the interviewer and the interviewee I I I can can automate the interviewer but I can't automate the interviewee because I need to to get what's in their head. Yep. So are are you an AI are you an AI Avatar? Is that what you're are you telling you're you an A? I am I have no emotion. Do you need to yell at on Yeah that's right. That's right. I've been stoic stone cold. You can't rough me up. ⁓ that's awesome. Well Mike, I mean I tell you it's been awesome to chat with you today. I just as as maybe a closing thing, I I mean we've talked about a ton and You know, you've been in the consulting space for years, you've built, you know, technology, you've written books, and you know, I just think about our our listeners, you know, the CEOs, executives, leaders, you know, peers, people in in a variety of industries that, you know, ⁓ listen, ⁓ my parents, you know, my ⁓ but but like what what in your mind you know, with with with what you have knowledge on, like what would be kind of the one thing you would leave folks with and just say, you know, as you think about as all of us thinking about leading teams and companies and people and and organizations, like what what what would Michael say that, you know, we need to really make sure we're thinking about over the next, you know, year or two to increase our chances of success? You know, ultimately I think it's the it's the context data is, you know, these things are smart and are trained on generally available data, but you know even with like cloud cowork like the reason why it's amazing is because I could I could point at the files in my in my ⁓ hard drive. ⁓ you gotta you gotta think through your data and then you know back up a few steps if you're gonna do that. Like what is what does that mean? What does that mean for people? What does that mean for how you validate accuracy, like there's a whole there's a whole bunch of things you need to think about. ⁓ so yeah it's it's not it's not an easy button. It it does take does take some work to enable it. But once once that works in place, I think it could do some amazing things. Yeah. It's awesome. Perfect man. Well dude I I really enjoyed catching up. I'm so glad we got to catch up today. And then I and you're right up the road in ⁓ in North Carolina and so you'd love to I'll hit you up next time I'm in Greenville. Yeah man and same thing. If I'm up that way I'll let you know about Appreciate you joining the pod today, man. Yeah, thanks. Thanks for the time. Okay, thanks for joining me on Built with Purpose. If you have enjoyed today's episode, subscribe, share, and leave us a review. It helps more bold leaders find the show. For resources and show notes, visit our website and connect with me, Chris Faye, on LinkedIn. Until next time, keep building great companies, cultures, and legacies with purpose.