AI Adoption & Implementation in the Enterprise

Most companies are somewhere between excited about AI and quietly stuck. In this episode, George sits down with Walt Carper, Founder of Cogsentia, to dig into why so many AI initiatives stall before they ever produce real value, and what it takes to move from pilot to production.

Walt brings 25+ years of hands-on experience leading large-scale technology transformation for Fortune 500 companies and the federal government. The conversation covers the real reasons AI projects fail, how to build a business case that survives the C-Suite and what Human/AI Hybrid Teams look like on the ground.

If your organization is navigating AI adoption, this one is worth your time.

Cogsentia
https://www.cogsentia.com/
https://www.linkedin.com/services/page/a818a033a20a087267/

Keystone Technology Partners
https://keystonetechpartners.org/

Transcript

SPEAKER_00 0:20

Welcome back to the Keystone Technology Podcast where strategy meets execution, buzzwords have to earn their keep, and shiny new tech has to survive contact with reality. I'm your host, George Adair. We are addressing the elephant in everyone's boardroom today. AI and the massive gap between the hype and the reality of getting it to work inside organizations. And our guest today has spent his career living in that very same gap. Walt Carper is joining us today, who's the founder of CogSentia, a firm that built specifically to help organizations navigate this AI adoption, cut through all that noise, and deliver solutions that really produce uh sound ROI. Walt, welcome. Yeah, would you mind starting out by taking us, taking the audience through a bit of what you're up to and how you go to where you are today, how you got to where you are today?

SPEAKER_01 1:22

Yeah, that'd be great. Thank you, George. I appreciate the opportunity. So uh Coxentia is a company, our name literally means knowledge and experience. If you go back to the Latin roots of where the name came from, that's really what we're built on. And the philosophy that we're approaching the industry is that we want to cut through all the things that you just mentioned, in terms of cutting through the hype, cutting through the noise. How do we find real productivity and do adoption of these leading-edge technologies like AI safely and efficiently? So, what we see today is you know, firms kind of fall in really one of two extremes. And it's it's unusual that you really don't see too many in the middle, but you either see firms that are very loose in their AI and people are kind of just experimenting, and you have maybe pockets of productivity where some people are very productive and know how to use AI very well. You have other people who are reluctant or latent adopters who really struggle to use the technology and they don't communicate well. So, you know, you run the risk of like those firms could be in a position where maybe you've got these early adopters when they're flying high, but maybe not flying very safely. And then you've got the opposite extreme where you've got firms that are completely like, oh my gosh, what are we going to do? We've got to stop everything. We can't do any of this. And I have seen, you know, one of my customers today, they won't even allow grant relief to be used, they won't allow open AI to be used. You know, so I mean it's very, very, you know, two opposite extremes is what I'm seeing in the industry. And both problems are bad, right? Because if you're flying high and flying unsafely, then you're exposing the organization to risk. And if you are flying conservatively and deciding, hey, I'll I'll never let it get off the ground in the illusion that's providing safety, guess what? You're going to be marginalized. So where we're trying to go is we want to meet the industry where it is today and find that you know middle ground where we can say we provide a structured, safe, confident methodology where you don't have to be an expert in AI to find the productivity. So we want to democratize that, but at the same time, we want to do it safe. So that's really kind of the the thinking behind all of our platforms and how we work.

SPEAKER_00 3:43

Oh, that's great. Thanks for that overview. Um, and you know, you really set up where we're gonna go here because I think every organization is looking for a cogsentia, but they don't necessarily probably realize it. You know, they're trying to leverage their internal executive teams, maybe even business, you know, business owners and functional business lines are doing it themselves. Uh, and they're trying to accomplish it in in these silos. Um and what's happening is I've got a little stat here. Gartner's saying that around 85% of these projects are failing, and that's a huge number, right? We've been in the innovation cycle plenty of times. We've seen things come and go, uh, but I think this is a much uh bigger you know endeavor that we've taken on, and we're noticing that the failure rate is much higher. Um, so so really, what how do we get through that, Walt? How do we get organizations like yours into uh the right places to help out these people that are trying to take it on themselves to get knocked down this 85% and get it to a meaningful rate? You know, so what what are some of the things that you're seeing that's engaging the organizations?

SPEAKER_01 4:58

Yeah, that's a that's a great question, George. And I think I'd have to set that up a little bit by saying that, in my opinion, and I I would say that of every project that I've ever worked on that was in crisis, you know, they have these called type projects. Probably 100% of the time. In fact, I can't think of a single time where the problem wasn't human. Right? It's always the human factors that are approaching the project. It's either emotional, it could be you know perceptions, it could be all these different cultural different things that are coming into play. And certainly AI adoption is exasperating a lot of these factors that have always been there. So I I think the first thing that when we approach a problem like this in terms of you know, why is my project failing, is well, you know, did you actually set your project up for success? So the things I look at in terms of like when we start looking at is say, well, first of all, are we solving for technology or are we solving a business problem? Because vendors want to sell you technology, right? That's you know, that's their bread and butter. OpenAI wants to sell you their their models, Claude wants to sell you co-working, all these different things that are out there today are really vendor-oriented solutions. And I think that one of the first things you have to look at when you're doing any new adoption is you really start to go back to what is my definition of success? And this is you know, this is not an AI specific thing, this is literally what I built my whole career on, is like when I would go and talk to people and they would say, Well, we're doing such and such, and I would ask them why. Tell me why. And it's the same thing applies here for AI. Okay, tell me why you're doing it. And you'd be surprised at the number of times I done actually get an answer to that question. And so that hasn't changed in terms of my whole career, but I can look at it and say that you know, the biggest challenges that we have in terms of those failed projects is we have to address the different factors, and those can really run the gamut from I'm either not confident in the technology, I'm afraid of the technology, or it could be uh things like I'm just not sure that we can do this safely, you know, all over the map. So we we have to look at it on a case-by-case basis. But I I think that my my premise is that if we're successful, we go in and we coach the human factors, and we bring in our platforms are really designed to make sure that the human factors and the fear factors come down because we're doing things safely, we're doing things with proven methodologies, we're just basically optimizing the AI for the type of work that it can do well versus the type of work that humans do best. And I think that's really the fundamental thing we have to look at is when we look at the failure, we have to address the the reasons for failure. And in my argument, it would be 100% of the time, those reasons for failure argument factors, not the technology.

SPEAKER_00 7:55

Yeah, really good points. And and something I want to touch on that you mentioned there, trying to get out of those stage gates. You you said, you know, making sure that the AI is safe and secure. And I I'm I'm pretty certain most, if not all, organizations are thinking of that. But the thing that I often hear is well, safety means we're not going to get anything done, or safety means we're not gonna get out of the wait and watch, or safety means, you know, X, Y, and Z. And it has this negative connotation, I think, at organizations. So talk to me about how you get around that, and you know, what are some things you can do to ease executives' minds by addressing the security, but letting them know that, hey, we're we're ready to improve your organization, get get that product to market quick, you know.

SPEAKER_01 8:47

Mm-hmm. Well, uh, there's a couple of ways I approach that. And sometimes really have to educate the executive team on what's really going on. Right. And sometimes that those can be things like shadow IT, where you know, maybe even within the organization or sometimes you know outside of the organization, if people are desperate to get things done, they want to make sure that they're not getting left behind in their career advancement and career development. They understand, you know, intuitively at some level that you know if they can use AI effectively, then it's going to elevate their standing and put in the company. So you have to and so for people who are of that mindset, these are the ones I call the high flyers, they are they are going to be doing this anyway. And I think the first thing you have to under help the executive suite understand is that even if you take the most extreme position, say, well, I'm going to sell my foot firmly on the break, so I make sure that every every I is on it, T is crossed, and you know, that I'm a hundred percent absolutely secure safety. Well, guess what? You're not because people are going to be frustrated and they're going to take it outside, and they will either engage with outside vendors or they will go and try and do things on their own. Fundamentally, the need is going to get met. Whether you step up as an organization to meet the need, or people try to self-fulfill the need, it's going to happen one way or the other. So that's key key learning number one is you have to address the elephant in the room. The second thing is that when you look at the things that people are advertising in the industry and saying, well, you know, my my frontier model gave me confidently wrong results. Or my coding platform, like there was one recently where some guy literally wiped out like three years of development work in about 30 seconds because of the something of that a you know coding interface, it wasn't clawed, I think it's a cursor, but bottom of the line is you see all these big headlines that are out there about all these massive things happen. And and realistically, you have to point out to the executives that the reason those things happened were not limitations of the tools, those were process figures. And if you have the process case in place and you're doing the work safely, and quite honestly, the work doing getting done safely isn't terribly different than what we've done to do work safely for the last 30 years. It's just you know, maybe modifying some of the approaches that we've always done. But the idea is, you know, you know, common sense and good practices and solid engineering doesn't go away because we have AI. If anything, we have to do this more rigorously than we did before. We just have to do it in a shorter time frame.

SPEAKER_00 11:32

Yeah, no, very good point. And I'll have to say, you know, um, I I think the common sense might be uh I don't want to say, I don't want to use the heavy word flying out the window or the heavy phrase flying out the window, but it certainly is taking a little bit more of a back seat in some organizations. And I think those are the ones that are that are probably spending um huge amounts of capital, and their expectations are so high that they're willing to reduce the process to get get it to you know a place where they're seeing marginal improvements in their revenue. Um, and that's just a an assumption. I just want to point out uh a stat that I have here. Global AI investment has hit 91 billion in in 2024 and expects to hit, I'm sorry, 2023, it looks like, and expects to hit 200 billion by 2025. That's this year or neck last year. So we'll see what those numbers come out to to be once they have the figures. But uh, it does seem like the investments are going to exponentially grow and the amount of uh capital that's gonna be set aside for AI is is going to just increase dramatically. So, how do we keep that sensical uh behavior, that process rigor that's been in organizations intact uh while C-suites and boards are saying, look, you're behind, we want to see the investment, and and we don't care how you got there, just tell us uh when we're successful.

SPEAKER_01 13:04

Yeah, boy, that boy, you've really touched on something that's very sensitive to me there, George, is that I see this so much, where there is so much enthusiasm on saying, okay, we are using this technology. And there's a term in the industry right now, and I'm not sure if you're familiar with it or heard it, it's called AI washing, where you're basically saying that uh, in effect, whatever inefficiencies I have in my organization, you know, either maybe it's staffing or else. Well, we're gonna do our layoff cycle, but now we're gonna say do our layoff cycle and say, well, AI has bought us so many efficiencies, we don't need these people. And you see it all over the headlines. I mean, big companies, small companies, it's just making massive headlines right now. And I'm just thinking, that is so incredibly foresight. And the reason for and and it's just you know, the value of the AI is loud, let the AI do the mechanics, you know, let it sort through data, let it take all this noise and extract signal. That's what AI does really well. Whereas the humans and the organization, they provide the experience, the judgment, the ability to make sense out of conflicting data, you know, that's the human value. And if we let our human capital go in order to justify these massive expenditures on AI, it's a very short-slighted game. And so to me, the biggest thing that really upsets me the most in the industry today is this whole concept of AI washing. And I would be willing to bet a lot of these fronts really haven't achieved anything, or very little as far as efficiency with AI. But whatever they're basically going to sugarcoat and tell their board, well, we save so much with AI that we don't need all these people anymore. And for a quarter or two, that may work. But long term it's not. And I think we're going to see a reckoning in the industry going in through 27, where you know a lot of these firms that did these massive lay-and-offset intellectual capital, that intellectual capital is going to be picked up by their competitors. And I think it's going to be the detriment to some of the people who I think were very short-sighted trying to do their short-term justification to offset these huge expenditures because the board expected me to tell me, tell them what we're going to do with AI. So back to your original question, I think what you have to do is really come back and say you got to ground through all of this. You know, you don't design solutions around the technology, you design the solutions to solve the problem. So our basics blocking and tackling, George, we start out with like, what's your problem? Define the problem. Are we going to use what we have available today in our toolkit to solve this problem? You know, it doesn't have to be all AI. It can be AI augmented, but things that we already do. It could be RPA, could be you know traditional you know, systems development, whatever the choices are, is all just because we have a new tool in our toolbox doesn't mean that the screwdrivers and the renters and all the things we've used the last 30 years are suddenly obsolete. They're not. And that's really how we have to approach it.

SPEAKER_00 16:11

I I couldn't agree more. And I think that the humanistic uh gap or just lack there of understanding is might be the the biggest, you know, we're gonna look back in the future and we're gonna say, wow, did we get that wrong? And and what we got wrong was the human element of it, that that there was a component that we should have looked at during this time of, you know, as anybody that goes through innovation cycles and innovation seasons knows that you're gonna build up and take your team with you. And therefore, if you do that, you'll know that you'll have a sustaining effect in the long term, where not only do you have good loyalty from your staff, but you also have them now at the levels and the understanding of that product, that innovation that you're trying to put in, that they'll they'll be with you for the long haul. Now, what you're gonna get is this distrust for your organization, this this unmet expectation that you know the every human is looking for, that tie-in, that um, hey, they they've got my back. I'm gonna go learn and and put my extra effort and time into this new technology, and therefore I'm gonna give it to this company. That's not gonna, I'm gonna guess there's gonna be a a lack of of that loyalty uh from both sides of the fence, and that's gonna create a big gap. So, yeah, so I I that really wasn't, I I think that's a great thing to to continue to talk about. It wasn't in some of my stats, and I and I think that's a miss as far as these uh how we get this AI adoption to a sustainable state, and we have to bring in that human element, and not a lot of data really pulls into that. Um so I think that's uh that's something we could probably find some more data on and as we go forward in this new AI world. Yeah, that's the hardest thing to measure.

SPEAKER_01 18:03

Is how do you measure you know someone's you know where somebody's state of mind is. But you know, the things that I typically see George are, you know, uh first of all, you know, there's this saying in the industry right now that you know that you have to earn trust, and you know that you can you can have a hundred interactions with an AI model, and it can give you a hundred right answers, and you can feel really good about that. But the minute it gives you a wrong answer constantly, which it can do, all of a sudden trust is gone. All right, and then people get this foundational. And you really have to start thinking about it in terms of we are accustomed in the computer in the computer industry, especially, is we want a deterministic world. And up until very recently in the FNF AI, everything that we did was certainly deterministic. You know, we we built calculators and they basically can do math far better and far more accurately than we can. We can build computer software that can do deterministic problems faster and more reliably than email. The downside is now with the AI world, we're moving to our probabilistic model. And you know, and it's a probability and statistics. And 80% of the time it's right where you expect it to be. 95% of the time it's going to be within this band of rightness or correctness, whatever's left. Guess what? It's wrong. It's completely hallucinating. And then that basically is why you can say, okay, well, you know, I can how can I trust something that I don't know if it's going to give me the right answer? And I look back to it and say, Well, how do you trust so-and-so to give you the right answer? Or how do you trust any human because we're probabilistic systems? And so guess what? Now we have to re reframe our methodology into thinking that you know we are going to be introducing a probabilistic system, and you have to engineer for that. And so the idea is, first of all, it's not a system, or it's not an AI, it's a legendic system. And so what we do at Coxantia is, for example, you know, very simple examples. I might say, well, if I have if I'm using Claude to develop something because Claude is very good at reasoning, but I'm not necessarily sure that Claude's going to express it well. Well, I'll have a you know open AI rewrite it. All right. And the idea is you start tailoring the models and you never let one frontier model govern your entire solution. So that's like one of the design principles that we do to make sure things are safe. You know, is that you know you gotta think about well, how do I learn to trust what I'm getting back? And the idea is I never trust any system 100%. I always basically get a second opinion. It's like getting a second opinion from your doctor or your co-worker. You need second opinions on your frontier models. You know, that's just that's just reality. They're probabilistic systems. You know, if you build enough checks and balances into them, your confidence can be extremely high. And I think that's part of the education process. We need to get out there. These are wonderful tools, but they're probabilistic tools.

SPEAKER_00 21:05

Yeah, it's such a good point, and that's certainly something I hear a ton about. I I just had a great uh breakfast with a CEO or CIO, sorry, and and he was telling me about uh his organization that recently got bought out by this uh this organization, this uh very large PE organization, and they brought in this philanthrop or this this uh billionaire tycoon individual who uh lives and breathes AI. And his his primary focus is to come in and reinvent organizations with technology. Of course, his big thing right now is is AI. And so he was telling me about how uh you know he he spoke to the board and the executive team around how he sees and views AI. And what what he told them repeatedly was he doesn't trust one AI over the other, and he puts them up against one another, you know, using MCP technologies to tie the them in and and test them against one another. Um and so I think that's a big I think you're on to something there, Walt, where when you speak to organizations and you let them know, hey, you know, trust is a big factor and and we put humans to the stones when it comes to trust. Well we we can put the AI to the to the same rigor and the same testing end that that we did, you know, let's say the human by having multiple people view. And so I think there's there's an opportunity there when we talk about the AI gap and getting these things across the line, uh, there is a lot of distrust. There is a lot of scare tactics, you know, when you talk about hallucinating AI or LLMs that eventually start to provide inaccurate data because they've been overused, there's a way to kind of present that. Do you find that a need when you talk to your clients to really drive that home? Or does that come up?

SPEAKER_01 22:59

You know oh absolutely it's in fact I have that conversation almost on a continual basis because really you find the people who are practitioners, you know, I can make these comments to the practitioners and they're going, hey brother, this is exactly what we this is our life. All right. And then when you start picking it up sometimes it's just you know it's interesting as like when you get one or two levels materially removed from those who are in the trenches how different their perspectives are. And I think a lot of it is really being driven. We're in a height cycle you know we're at the Gartner height cycle and we are very much on that essential part of that Gartner hype cycle where you know there's uh you know we're still everybody's on the you know the pinnacle of the hype and then there's going to be a trough of disappointment in the plane of productivity. What we're trying to do is really you just want to skip that trough of disappointment goes for productivity. But the idea is that you know you really you have to educate people on it is and some of it is attacking both ways. I mean you have to look at you know there's this you know advantage that you know if I'm the expert and suddenly there's a system that becomes an expert why am I happy you know there's a fear factor there. There's a fear factor that's coming from you know people who observe the hype and that's being pressured from the board maybe the from Wall Street or maybe even from customers and say well why am I still paying you for this when I can I hear everybody's doing this a thousand times cheaper than AI. And so you know there's this radical there's so many pressures from both the bottom end and the top down. And I think we have to go back to our roots George and go back to fundamentals of engineering practice and discipline will get us through we solve the problems we use the tools and we educate and set realistic expectations then I think we're going to get around the 85% failure rate. But if we fail to do any of the things that we've learned how to do over the last 30 years we're going to continue to fail.

SPEAKER_00 24:57

That's my yeah yeah I think I think you're absolutely onto something there. And and let's let's tie in the next topic here which I think is is uh meaningful when we think of how not to fail in the future so some of the uh you know early stage AI is really tough to quantify right um it's it's tough to build out solid ROIs that CFOs demand and and a lot of people tend to say well if we get a good pilot out there um you know and specify the cost after the pilot you know we might be able to identify you know what we're actually going to get from this particular um this particular off-the-shelf ROI or I mean AI uh so talk to me about how you approach the conversation uh with CFOs in particular and or I guess board members that are really financially positioned and they're looking for that you know near you know accurate near 100% accurate ROI um how do you frame that what steps do you take that's a great question and and it's really there's not a one size fits all approach but I will tell you that in the past um I've certainly heard the argument about saying well maybe we should do a prototype like first and and I think that can be valid but my advice when if a customer is coming to me and they want opinions on a prototype is they'll tend to they'll pick a low risk uh easy to measure use case and say okay we're gonna prototype that and the problem I think that that eventually leads to is it doesn't set you up for success.

SPEAKER_01 26:39

So the first thing I would do in terms of people who say prototype first is okay but let's prototype something hard and let's and let's really get to the to the nitty gritty of it because if investors are going to be coming in behind us or we have to present this back to the board we just can't you know we can't give it a softball pitch. We're gonna have to get something that's real and expose here's where our you know area of uncertainties are and really try to frame the conversation around that. Second thing is in terms of educating up is you're never going to be 100% certain on the RRI. And third and this is my pet peeve you've probably heard him say don't make your RI about labor savings. Okay because I've heard this before I've heard this about RPA I'm you I'm sure you live the RPA dreams as well where people would say yeah well we'll we'll talk about hours back to the business. Okay that was that was a great term all right but how how many organizations really shrank because of RPA coming in probably none because the reality is once we develop a new way of playing the productivity the bar and the expectations always go up. They never come down they never stay the same. And so you know that that labor gets absorbed somewhere else so you got you can't fall into that trend. Again it's it's uh it's the same song that you and I have heard sung many times before we saw it in RPA. We probably saw it you know going back maybe more maybe than you about the early days of the internet the idea is that I've seen this pattern before we know how this movie ends so let's start we can think about let's pick really good prototypes let's set the expectations that necessarily ROI is going on is going to be a ban but it's not going to be necessarily disparately quantifiable. And then if we can get the CI uh CFOs to buy into the fact that we're expecting this range in productivity this is what we expect to see and really attack where we think we can gain market share or gain you know a branch and expand our business versus trying to say how we're going to attract the cost. So I'd like to I'd like to address the the demand side potential more than the cost side potential. And I'm not saying it's irrelevant but I think it's very important because if we get too focused on you know what I'm going to take away you s you lose sight of what I'm going to get. And so I like to find the conversation that this is what we're going to get and I do that managing up to the C suite. I also do that down because when you start managing it down you know people who are the people who are afraid of AI was going to take the job or whatever else you really have to say okay it's not about what the AI is going to take it's going to what's going to allow you to do do you really want to you know spend hours and hours editing documents do you really want to sift through you know you know spend five days you know trying to sift through these reports to gain insight or why don't you just give that to the machine and then you have brilliant the ability to bring the brilliant you know execution. So the idea is you know that you start you gotta start looking at what it's going to cost and what it's going to take away both the good and the bad start looking at this is what I'm going to get. And I think if we can focus people on this is what I'm going to gain, then I think we have the opportunity to really turn that around.

SPEAKER_00 30:00

Yeah you you're spot on. I think um you know I read somewhere McKinsey Global uh institute said that uh AI enabled process automation can reduce operational cost by 20 to 30 percent in targeted workflows obviously you're specifying what you're going to you're going to do but I think that's the the useful I think you're spot on that's the useful conversation for the CFO. It's not hey we're going to you know I know you have a you know 12% EBIT uh you know increase this year and you're trying to gain more market share that may not be where we want to position our you know the the today's most revolutionary technology that is just beginning we might want to look at more of our process automation and how do we clean up some manual tasks instead of trying to go after this massive overhaul of revenue and business of you know uh drivers that that redefine who we are it's just uh to me I think you're right it's it's a it's a aggressive approach for some organizations and then on the other side of it it's there's organizations that aren't moving at all and they're sort of sitting back and saying hey I'm gonna watch as this unfolds and you know I might wait five years or plus uh so what why don't we real quick talk on that I've got a a question for you so this is an interesting one that you know from a from a you have a business mind you run your business and your purpose is to find clients and and help them out uh but let's say you come into an organization and you know they're genuinely asking are we AI ready? You know are we capable of taking this on uh you do an assessment uh early on and you find out you know they're really not they're really not AI ready but you you think you can probably get them there at some point but how do you address that from a business standpoint? How do you look back at the company and say you're just not ready.

SPEAKER_01 32:03

You know that's that's funny. You know it's almost like the you know the questions would be has anyone ever told you that they're not ready? Yeah right and I would say nobody that was being honest with me. You know it's uh so yeah that's absolutely true. I think every organization is a complex amalgamation of human opinions, of human uh desires and motivations that none of which are probably exactly right. So I think the first thing to understand is we've got to have an objective standard by which we measure ourselves against and that way it can kind of take some of the opinions out of it. And you you alluded to this earlier about people all agree safety with slowness versus you know and and there's maybe some some merit to that but the idea I think you really want to come back to George is that if we have an objective standard and we have the ability to apply a strict lens to our objective standard then we can start taking some of the opinions out of the way. And it doesn't really become a question of are you ready or are you not ready? What are you ready to do is really the better question, right? Because yeah because anybody can do something right off front right you know you can say well based on you know maybe your your current uh data uh you know data protection policies maybe you have existing uh data sharing agreements whatever else you can say well I think you're ready to do this and it's not really a question of are you ready for AI or are you not it's just what's appropriate for your organization now. And I think that takes a lot of the a lot of the emotion out of the conversation and I think it's really what you have to do when you're faced with these is like you know no nobody's either black or white. I mean it's it's it's always somewhere in the middle and you have to really start thinking you start bring the conversation around I think you're ready to do this and this is why and that these are the things you can do and then next step looks like this I think that becomes a much easier conversation to have than to say you're not ready stop you know because that's never really the case. And but then again never as bad as some people think and it's never as great as some people that's right that's right.

SPEAKER_00 34:14

Yeah and and certainly to to close out that thought is you know technology is uh has has always been ahead of organizations for many many years and is certainly ready for an organization it is the people it is the desire it is the you know the goals the missions like you said what is our character who are we and how do we determine what our truth is um so I yeah I like the way you put that um and I think organizations need leaders like you to come in and and help them understand that truth you know and get them kind of unstuck out of their their ways. Well let me move into kind of a a final topic that I think is really interesting and I thought I'd position it in a certain way for us here. But uh there's some stats out there in the World Economic Forum that says by uh 2025 uh 85 million jobs have been displaced by AI and 97 million more new roles are projected to emerge requiring you know human AI collaboration. You mentioned uh there was a couple phrases you mentioned here as well as in uh at Cogsentia's website using the phrase human AI hybrid teams you know as a core part of what you do and I think a lot of people they might hear that and they might uh either picture robots taking over or you know uh or they might just overall dismiss it as some marketing language uh but what what does a for from your perspective what does a human AI hybrid team actually look like in place well that's a great question and I'll give you an example of how we work with clarity.

SPEAKER_01 35:59

Now Clarity is one of my platforms and uh just for the purpose of the audience Clarity is basically designed to solve a specific challenge around how do we get good requirements for our systems and do it in a much more efficient way than we used to do with traditional elaboration. So if you look at what Clarity does today we have a genetic elements that are built into Clarity where uh our process would start with maybe interviews we do basically uh jad sessions or you know similar type sessions things that are of the that are very tailored to human to human interaction but all that data comes into the agency part of the platform and the agents are really taking all this massive amounts of data and extract signal and then basically organizing things well that would take a human first of all you know most humans wouldn't really have the ability to deal with large amounts of data efficiently and certainly would require teams to spend a lot of time trying to sort and reconcile and what we can do with a human hybrid genetic team is we can say agents deal with data agents basically collaborate collate and queue things up for human insight and observation. And all of a sudden that's like taking our humans and giving them a superpower. So it goes back to that earlier conversation it's not about what we're taking away it's what we're getting and the idea is we want to give the humans the ability to do what humans do best. The judgment the assessment the ability to deal with complete conflicting points of information and make a rational code and decision that's what humans do best. So when we look at our platforms that's what we design for the other thing is we really have to democratize the AI too and I alluded to this a little bit earlier where if I were to say you know for example someone asked me well why can't I just give my team chat GPT and let them go wild with developing you know their their safe natural artifacts and I would my response to that is really comes back to as well first of all your team are not at the same level skills. Some are going to be you know traditional analysts who may have very low skills in terms of using AI. Some of you less experienced analysts I maybe the more younger ones might have far more experience with the AI but they lack the judgment. So the idea behind having a structured AI agenda platform like Clarity Brings is all of a sudden we democratize everything. So it's a very structured process in terms of we walk the entire team through the entire process. And so you don't have to be an expert in AI to actually enjoy the benefits of the AI. You don't have to worry about am I prompting model right? Is my contacts window too large? Did I poison you know poison it is a model trying to anticipate what I want versus what I actually need. And all these things are very real if you allow people to freeform with the AI. So I think the real value is we can democratize the AI but we can also democratize the productivity where people regardless of where they are on the adoption spectrum can all get the same benefit out. But we're not throwing away our safe and you know well driven methodologies. So we take a very structured well understood methodology we automate it we make it efficient we bring in the AI to do pieces that AI does well and I think everybody enjoys the benefit. And in fact I've seen it on the projects where we're using Clarity today. Our requirements elaboration time went from literally months to weeks and our iteration and our I mean it used to be you know we would go through you know epics in you know two or three months now it's two or three weeks right so we've seen a tremendous shrinking in the time to value as well as the the value delivered for the level of investment. So these two things are really powerful if we hit that right let the AI do what it does well let the humans do what they do well I think we have a very synergistic platform that provides immediate benefits that everybody appreciates both up and down.

SPEAKER_00 40:12

Yeah that's right everybody appreciates everybody gets to you know still have their core competencies used in organizations what they do best while you know taking off that low-hanging fruit you know AI handling the high volume stuff the low judgment tasks and then the humans right working the exception handling and and relationship working I mean and beyond that it's I think it's going to be much further. I do think there's going to be a realization in the future that we sort of look back and say we're we're really glad we we took this next step. And it didn't necessarily overhaul the world like we thought it was going to you know it it it augmented us in a way that allowed us to be able to do more uh you know with the the reduction of staff we already have I think people uh cease to understand that in good well-runned organizations even the biggest ones you still don't have enough people to handle the work that's right and your staff said it exactly right it's like maybe it leads to disruption on 80 some million whatever it was I think you said it was like 95 97 million created that's a wonderful topic because you know that's a net game yes some people did have to transition skills but you know the the other adage you hear a lot about is will AI replace me and the answer is no that somebody using AI probably will.

SPEAKER_01 41:39

So I think that's the message we have to get out to the to the people out there is that you know AI is a free tool and just like word processes were a great improvement at the time like Allie this is this is this is just another evolutionary step.

SPEAKER_00 41:55

Lean into it embrace it but you know I I think the the potential is far outweighs the the potential short term losses I think we have so much that we can do and I think it's going to be the world's going to be a much better place because we took this step than if we had to it is it is and I think the the caveat here is what you just said 12 million net new jobs 85 were lost 97 were were gained so we have 12 million net new um yeah that's a win in my book. Absolutely uh yeah so let me sum this up uh thank you so much Walt this was wonderful so you know when we think about the AI adoption the enterprise need uh we're really thinking of processes and core competencies things that we do as an organization or or say that we do that will help to enable this new technology to get past the line and so it's not necessarily about having to make wide sweeping changes to the organization. It's definitely not about having to wipe out FTEs in order to pay for it. It should be more aligned to what you what you how you operate and how you inject and take in new innovation using those lines of protection and governance to watch over what's coming in. And so you know CogSentia and if you want to sum this up please do but Cogsentia can come in and help uh organizations understand the way they operate where their business struggles are their biggest gaps and to be able to lean into their probably their most manual and arduous processes whether it's with software development or product enablement and many other things that really enable AI and their organizations to take their organizations to the next level.

SPEAKER_01 43:48

I couldn't have said it better than Snowdoor so thank you.

SPEAKER_00 43:51

Excellent all right well thank you so much for that that ends our topic section we're gonna jump into our rapid fire I hear we're gonna have a few minutes. Again, short answers, first instinct, Walt. Uh no wrong answers at all.

SPEAKER_01 44:08

Okay. Fire away.

SPEAKER_00 44:10

Firing away here. All right. So one word of phrase that describes where most companies are with AI right now.

SPEAKER_01 44:17

FOMO. Fear of missing out. That's probably a phrase more than a word, but it's an acronym. But it's I think it's it's exactly where people are. One way or the other, there's a lot of fear of missing out.

SPEAKER_00 44:29

I loved how you you went outside the lines instead of saying early stage or or piloting, you said FOMO. That is exactly right. 100%. Uh biggest myth about AI implementations you wish would just go away. AI is going to take our jobs.

SPEAKER_01 44:45

That's yeah, we've talked about that a lot today, but that's you know, AI is going to change your job, but it doesn't have to take your job. And I think the when we can educate and get people comfortable with the concept, it's going to be truly transformational.

SPEAKER_00 44:59

Yeah, couldn't agree more. All right. AI agent, RPA, or chatbot. If a company can only deploy one of this uh this year, which one?

SPEAKER_01 45:09

That's a tough one to generalize. And I think if you were, I would have to answer it by saying if you're starting from zero, start with a chatbot, because that's pretty low risk and pretty easy. But uh if you're somewhere else on the adoption, the answer could change. But yeah, let's say if you're starting from zero and you're just making first tone of water, start with a chatbot. It gives people a lot of exposure to you know the good and the bad in terms of AI adoption, and it's a great training ground.

SPEAKER_00 45:38

And it's been around for a lot of years, so I think there's opportunity for that particular you know AI to have developed a lot faster than others. So Cog Centia is built on knowledge plus lived experiences. What's one moment of a lived experience that fundamentally shaped how you work with clients?

SPEAKER_01 45:57

Well, uh, I said it earlier, and I'll just kind of repeat it here. It's the fact that projects don't fail because of technology, it's because of humans. And when you understand the human factors that are pushing the project off the cliff, that's when you start transforming and bringing it back. So it's not necessarily an experience, but it's a repeated experience. When you realize that projects fail because of humans, then the solutions become absolutely all right.

SPEAKER_00 46:27

Last one here five years from now, what does a truly AI-first organization actually look like?

SPEAKER_01 46:35

I I think it's going to be, first of all, incredible time to value and incredible value for investment. I think those are gonna be the two main characteristics of any AI-first firm is that you know we're really going to redefine the new normal as a result of this. And those who embrace the new normal are going to be our future leaders. And those who are lagging behind may see some erosion of the markets here.

SPEAKER_00 47:00

Yeah, yeah, I couldn't agree more. All right, Walt. Well, this has been a great conversation. I'm so glad to have you on the show today. Uh, thank you for coming in and providing our listeners with some really practical approaches to transforming their organizations uh with AI and just getting past all those difficult setbacks. Uh, for anyone who's listening who wants to connect with you or learn more about CogSentia, uh, where does uh one where does one go?

SPEAKER_01 47:30

Well, website's a great place to start. It's www.constentia.com. And or you can find me on LinkVID. And uh hopefully we can put some links to both of those in the uh podcast as you know we go into hosts. But the idea is you know, find me there. I'm happy to have a conversation.

SPEAKER_00 47:48

Excellent. Yeah, we will put all of these notes and uh how to get a hold of Walt in the show notes. Please look there for any information. So thanks for having another thought-provoking episode in the Keystone Technology Podcast. If the podcast meant something to you, share it, please. Forward it to an AI leader, CTO, CIO, VP director, anybody in IT who's trying to figure out their roadmap. Send it to the team that really needs to hear that the pattern has played out before and there is a path through it. That's how these conversations just get further along. All right, until next time, build with purpose, keep leading with clarity, and keep pushing past the buzzwords toward what actually works.

SPEAKER_01 48:33

Thank you, George.

SPEAKER_00 48:34

Thank you all.