Dissecting the AI Hype Cycle—with Joshua Greenbaum (Enterprise Applications Consulting)

What if AI is just the latest Blockchain?

In this episode of Futureproof, Prakash Chandran sits down with Joshua Greenbaum of Enterprise Applications Consulting to explore the AI hype cycle. Josh reflects on his 30 years of technology consulting and examines whether AI is following the same trajectory as other technologies, where true value can get lost in the froth and frenzy of investors and founders trying to capitalize on it. Together, they explore the reality of whether SaaS really is dead, the criticality of standardizing data in the AI era, and the central question that all AI companies should be able to answer. 

Topics covered include:

• Technology history repeats itself. While technology itself changes constantly, the way it is received in the market doesn’t. From dotcom to Blockchain, the hype cycle has a pattern.
• The hard parts of SaaS. SaaS is certainly changing, but reports of its death have been greatly exaggerated. AI that can build prototypes is far from replacing companies like Salesforce, which have learned from years of on-the-ground work with real customers and real problems.
• The importance of data standardization. The value of AI will come down to how well it can access the information it needs. Standardization of data and logic is critical to this outcome, and one that companies have to get right before they can succeed with AI.
• The lone wolf developer. Developers aren’t going away, but developers that work in a vacuum may be. Building is a team sport even more than it was before, and the developers who can see and understand business problems are the ones that will build the future.
• The real question all AI companies need to ask. No business leader is waking up in the middle of the night thinking, “I need a large language model!” All companies, and AI companies in particular, must remain laser-focused on the actually important question: “What business problem am I solving?”

Episode ID: 18906617-dissecting-the-ai-hype-cycle-with-joshua-greenbaum-enterprise-applications-consulting

Subscribe to Futureproof wherever you get your podcasts.

From Xano - The fastest way to create a production-ready backend for any app or agent. Xano unifies AI speed, code control, and visual clarity, so you never trade reliability for velocity. Sign up for free today.

Transcript

SPEAKER_01 0:00

If we're gonna make these giant investments, let's make sure we can justify it by solving million-dollar problems, not just 50 cent problems. A business stakeholder wakes up at them at 2 a.m. in the morning in the cold sweat because she's got a business problem. Do they wake up in the middle of the night and go, oh my god, I you know, I need a large language model?

SPEAKER_00 0:19

No, they don't have a welcome to Future Proof. I'm Prakash Chudrin, CEO of Xano. Today I'm joined by Joshua Greenbaum. Josh is an enterprise application consultant and one of the longest-tenured independent enterprise technology analysts in the industry. Josh started his career as a programmer, moved to Europe in the early 90s, where he wrote the first technical analysis of SAP's R3 for the American market, spent years as a tech journalist, and has been an independent analyst for over 25 years. He's covered it, called it out every major hype cycle from dot com to blockchain to the current AI wave. So nothing is safe or off limits. We'll talk about that.

SPEAKER_01 1:13

Safe or sacred, yes.

SPEAKER_00 1:15

Thank you. Well, I'm I'm excited for this conversation with Josh because he's not anti-AI, but he is anti-hype. He's watched the same patterns repeat for three decades. Promising technology gets over promise. Organizations skip the hard work and projects fail. His consistent message is that technology alone doesn't solve anything. The real bottleneck is always people, process, and change management. And right now, as AI promises to accelerate faster than organizations can absorb, that message has never been more relevant. Josh, thank you so much for being here today. I'm excited for this conversation.

SPEAKER_01 1:48

Thanks, Prakash. Thanks for having me on board. And and uh I the people listening on couldn't see the couple of eye rolls I I made during during that introduction, but thank you very much. It's great to be here. Uh great to have this discussion and great, you know, future proof is I listened to a couple of the podcasts, more than a couple.

SPEAKER_00 2:04

Yeah.

SPEAKER_01 2:04

I'm excited to join enjoin your your your uh your your your posse of uh friends and family you showed.

SPEAKER_00 2:11

Yeah, of course. Well, I think you have a really unique perspective and uh and just a lot of experience that I think most people uh don't have. They don't have that muscle memory and they don't have like uh the years uh being an enterprise software that you have. So I uh that's actually where I wanted to start. Just talking a little bit uh about your career. Just give us a quick overview, a quick summary, and how you became kind of one of the longest uh running tech analysts in the market today.

SPEAKER_01 2:37

Yeah, dubious distinction award. Thank you. I um I mean, uh why did I become the longest uh reigning enterprise tech analyst, uh stubborn uh uh perseverance, I guess. I I um I was very lucky that I was able to find this niche back uh starting really in the 80s, uh late 80s. And I think one of the things that that what really struck me is I was I was a tech journalist, I was writing about the software market. We didn't really have an enterprise software market per se at the time. We had a lot of development tools, very interesting, because we're sort of that's the cycle that keeps returning. The next platform, each new platform has to grow its tool base and and grow its uh the the folks who are are expert at it and and start really applying those tools to the next to the next business problem. And this was this was the dawn of the client server market, relational databases, uh a lot of a lot of net, you know, a lot of new tools were were showing up, development tools, software development tools, uh networking tools, et cetera, to really help build those applications. And um I was very fortunate to have sort of caught caught a wave of interest in taking these tools and really building neck next generation robust enterprise software. Oracle actually started it with Oracle Financials um and sort of ran with that. Uh obviously the Oracle database was foundational to all of this uh new technology. And um I um had gone to school in Europe and was sitting around in in um here in California in the uh late 80s, early 90s saying, you know, there's all this action going on in Europe and uh no one knows anything about it. I'd love to go back there. So I put together this this freelance little business of mine and popped over to Paris in the end of 1991, and and you know, luck favors the prepared mind. So I was there prepared, but luck really landed in my lap. SAP announced the launch of the what was then the the biggest piece of client server software ever, R3. R3 really kicked off the modern ERP era, and I got to have a front row seat. And that my background and understanding what Oracle had been doing and the tools and the database business really let me move into that. And suddenly, you know, there's this ERP market that that really really took off the big, you know, the big global systems integrators, as we call them now, consulting firms, as we call them then, really jumped on the bandwagon and we were off to the races and kind of been there ever since. So it's been it's been a fun ride. And thank you for you know the the muscle memory comment because you know, as as you know already, you're you know, anyone who's been as I always say anyone who's been in this industry for more than five minutes sees that it's very, very cyclical. We we tend to do the same sort of things.

SPEAKER_00 5:38

Well, that's uh that's actually what I wanted to ask you next. Um, you know, you've kind of described your career in in uh preparation for this conversation, you've described your career as moving in circles rather than a straight line. And I was wondering if you could elaborate on what you mean by that.

SPEAKER_01 5:56

Well, I think, you know, the industry moves. I mean, you know, if we want to get metaphysical physical, this the circle is a fundamental, you know, organizing sort of structure of the universe. Everything is in circles. And we talk about the circle of life as well. We have the circles, circles of technology. I think that we always have business problems to solve. We always have things that we want to build, and we always have new new ways in which we want to do them. And it's it's that it's that marriage of new technology and new ideas that's cyclical because there's always going to be this this this similarity between it's whether, you know, whatever kind of new new set of tools and new set of technologies, we have to learn what the we have to learn the limits, we have to learn the possibilities, we have to live through the hype cycle and kind of sort of move through the hype to get to the reality. We have to teach ourselves and our business stakeholders value, and then we have to really deliver it. So it's it seems like these, you know, every every every new cycle feels similar because there's still these basic criteria about what it is you're gonna try to do, how you're gonna do it, and how do you get people to change? And you mentioned, you know, change management is really fundamentally the biggest question. We we want this to work, and that means it can't just be the technology saying, you know, hammer on the head, you're gonna do this. We've got to actually bring our stakeholders along for the ride and make them make them valuable.

SPEAKER_00 7:25

Yeah. So one of the things actually uh regarding kind of the cyclical comment, I totally think that you're right. Just as, you know, uh the industry kind of goes through these different cycles, you know, you've had the privilege over 30 years to look at dot-com, to look at IoT, to look at blockchain, to look at uh now AI, right? And I'm I'm curious, just at a high level, you know, we're we're going to get definitely into the change management discussion, but what is like something that you just commonly see, having covered all of these different cycles, what is something like it always feels like this is it, right? This is going to change everything. And there feels like this like wave of things or patterns that we see over and over. Can you maybe speak to some of that?

SPEAKER_01 8:10

Um, yeah, and unfortunately, it's it's not sort of maybe you know the most positive vision of our industry, but we we really become enamored of the of the this shiny new object. And uh and there's an immediate uh need to overpromise. There's an immediate need to build a hype cycle because for better or for worse, you know, the financial side, the venture capital side, the private equity side, money drives this industry. And um, unfortunately, along with that money side, which can be good, right? Because how else do we build new companies? But with, you know, we take equity and capital and do that. There's there's there's unfortunately a real, you know, a real tendency to try to get one over on the small investor. I hate to be really blunt, but I've seen some really ugly things go on where where we talk about something that's very exciting and from a technology standpoint, a practitioner standpoint, we get it and then we look around and see how it's being applied. Uh blockchain's a very good example. I would say, you know, the the it there were some really interesting ideas, but there was a lot of just plain old hype. And it was one of those situations where if you actually ask some of these, you know, investors in some of this technology and some of the things we were doing, you know, to to to you know, with putting value putting blockchain into a into value, um, it was it was completely overdone, over, overhyped. And and this is my main problem with with you know AI today is that there's so much promise and so much capability, so many really great, honest practitioners doing really good work. And then there's the froth and and and frenzy of the market and uh how people really trying to gain over you know, overdo it in terms of hyping it up and trying to make, you know, frankly, make a little more of it than than is deserving.

SPEAKER_00 10:01

Well, one of one of those uh the pieces in the froth and the hype uh that I know you and I were talking about earlier was just this whole like SAS is dead, right? Like you see this, everyone is talking about this. Uh there's been articles written about it, um, but it's so much more nuanced than that. And I remember um, you know, a month ago, I think you shared an article from um your friend Thomas Otter uh that kind of like, you know, spoke just some of the more nuanced things that people need to think through before making statements like that. I'm curious if you could share with the audience around like when you hear statements like that, what are they not necessarily thinking through? And where's your mind? Like from a month ago when you were thinking about it, where are you now? Like how have you kind of like codified your thinking around that statement and what people need to think through?

SPEAKER_01 10:46

I think it was Mark Benioff when he said software is his dead and he had that you know started salesforce.com that we had this sort of let's let's kill, let you know, we have to we have to kill our it's a very Oedipal sort of kill our father and you know to move on to the next generation. So yeah, SaaS is dead, and frankly, it's SAS isn't, you know, every every time that statement comes out, it's I I take a deep breath and say, no, it's you know, we we want to be able to figure out a way to move forward, but we're not really, you know, really have to again, you know, kill what came before us to do it. And yet there we here we are again. SAS SAS software is has so many incredible attributes that that to say that this the this next generation of AI is going to replace it is really to completely misconstrue what what the software market does. Uh and first of all, SaaS is a label anyway. So let's peel back the onion of what are we talking about when we say SaaS software? What we're really talking about is this the current, you know, the current functional generation of software that's actually driving a huge amount of business value across the global economy. It's it's being delivered in SaaS because that's the delivery model that makes the most sense. And at the end of the day, however much you want to look at new development technologies and new uh new ways in which to rethink user experiences, which I would say you know, fit into the new AI models directly, you can't get away from the fact that inside the some of this existing SaaS software is a tremendous amount of thought and technology and coding that really allows companies to actually function and run in very complex business environments across a globe, a very complex global economy. You're not just going to replicate that in you know, coming up, you know, with a bunch of smart programmers sitting around doing whatever it is you can do with the latest technology. That doesn't happen. That there's a lot of hard work and a lot of complexity that goes into building those applications. And that you can't replace that over a knife.

SPEAKER_00 12:48

Yeah, and I think that's something just worth expanding on. That like, you know, I think uh especially last year, um, you know, just like the speed in which AI demonstrated that you were able to create the first working draft of something is, I mean, it is truly remarkable. But when you think about, for example, like even a company like Salesforce, think about the number of nuanced edge cases that they need to be able to handle, the amount of bugs that they have had to fix. Um, that just doesn't get baked into, you know, like uh a spec in a system prop to then just be spit out. You know, when you go to enterprise, when you go to production, when you're handling all of these nuanced use cases, it is uh a product of people working out in the field, coming up with these weird edge use cases that then Salesforce has to bake into its product. So, you know, as much as uh monolithic SaaS software uh that Salesforce is, that's just an example of why it is so hard to disintermediate these big companies, these companies that really focus on the customer, these nuanced value here uh in a way that AI just can't uh can't do.

SPEAKER_01 13:57

It's easy to build a prototype, it's easy to build an MVP, it's easy to build something exciting and new. It's hard to maintain it. It's hard to take it out of the laboratory and into the real world and say, okay, so now we're sitting inside a complex global logistics supply chain, and we have these requirements about risk management and compliance management, accounting, tax, you know, tax and regulatory um compliance and all of those, all of those nuances, and you know, they're not even educations. That's the norm, that's the standard. Business is complex. So, and and the you have to really account for that when you start taking things out of the laboratory and into the real world. Um, one of the things that SaaS does so well, because it's a develop once deliver, once deploy many kind of model is compliance. And that's one of my favorite uh aspects of going to a larger SaaS provider who's got a team of people who are dealing with compliance and dealing with it at a global level. It's really fun to build that MVP and make something that looks 10 times better than a Salesforce or an SAP or whatever. Adding that additional value and maintaining it day in, literally day in and day out, year in and year out, that's a whole nother level of work that I think is is lost on the on the kind of the develop, the development first kind of crowd. This stuff needs to have a life cycle that's measured in the value cycle of a company, not in the hype cycle of a new technology. And I think that that being able to being able to bring that mentality into the new AI world, I think is really essential. It's kind of what I'm trying to do is say, hey, this stuff can be cool. It has certain applications, it has certain abilities, but let's ground it in reality. And if we're going to make these giant investments in technology and in hardware data centers, whatever, let's make sure we can justify it by solving million-dollar problems, not just 50 cent problems. And that is my other concern is we see a lot of the hype around AI, particularly generative AI, is about the the this consumer experience that people are having with Chat GP. We're solving, you know, that's not solving big problems, that's solving small problems, which is great, but is that commensurate with the investment we're making? Is that commensurate with the potential environmental damage we're doing to the planet by by by doing all this? I I want to question that and make sure that we're all questioning that. So we're not just jumping off a cliff literally into a brave new world we don't really understand.

SPEAKER_00 16:40

Yeah. So I think there's a couple threads that I want to pull on there. But you know, one of the first is this like notion of like once that prototype is built, what does it mean to maintain, govern, enforce compliance over a project to kind of enforce that company's value in the long term or for the long term? Um, you know, I think that uh, you know, I think you you may have even said it, but they have like this statistic around uh X percentage of uh IT projects that fail, and 40% of AI projects and agentic projects that will be introduced into IT uh will fail. And I'm curious, like in that, buried within that, is not necessarily, I think, a technology problem, but a people and a process and a change management problem that I want to start talking about. So, you know, when we think about, I think the world of last year where AI was on everyone's tongue, it was a mandate from the top. Start building. We're not building fast enough. To I what I feel right now is everyone's having a little bit of sobering and trying to kind of suss out, okay, well, where do we normalize? How do we move at this AI speed with control? What does that really mean? And I'm curious as to with your experience being within these enterprises, what are you seeing? Why are AI projects failing? Why are they hitting their heads against the wall? And what are things, what are things that people aren't doing that they should be doing to deploy AI properly?

SPEAKER_01 18:06

Well, you know, you hit the nail on the head, Prakash, with the first part of that, which is that, you know, the failure is the norm, unfortunately, in IT projects. Um, it's not necessarily the dominant model by any, thankfully by any stretch of the imagination. But when we we look at the history of enterprise software implementation, particularly the this generation that started in the 90s, we have these enormous failure rates. And failure, you have to define it carefully, but so I'm gonna do it. It's it's failed to achieve the expected results within the time frame also expected. So this so yeah, we eventually can get there, and yeah, we can eventually get there, you know. But did we did we actually uh did we achieve what we wanted to do, but B, did we do it in a in a in any in any way in a cost-effective way, in a way that met our needs in the timeframe we needed and didn't didn't you know destroy our budgets and and our and our company in the process. This has been a a major problem since since the dawn of this industry and has not changed. We thought SaaS would change that. That that's one of the ironies of this is okay, we were doing implementations the hard way, we were doing it in the data center. It was everything was a one-off. It was, you know, it we'd have to get the army of consultants in, they you know, park the bus out front and take them off and get that. That that was supposed to go away with SaaS, it didn't. And and we're seeing this this latest, this latest report on this 40%. I think that's optimistic. It's probably a little more like 50% because that's been the norm. 15% of these pro agentic projects aren't meeting their expected you know time frame and value. And that has to do fundamentally with a lot of several different factors. One is that we we tend to really misjudge how technology is supposed to be applied to an actual business problem because the business stakeholders aren't usually in the room when these questions are being asked. It's the technologist second guessing in any in certain ways, it's it's people who are who are you know well meaning in in a certain extent, but they're not actually the practitioners. They run off, you know, they they run off with a great idea, and then you know, and then you hit reality and and it fails. The other, and even if you have, even if you start change management, right? You say, okay, let's get, let's get this, let's get our our primary stakeholder in the room. We're gonna change how we how we sell something, we're gonna change how we sell products, let's get our sales people in, we're gonna change the user experience so that we can sell faster and better. That's great. And that's a you know, that that's a good start. But what about your supply chain? If you, if you're if you're changing how the user experience is for buying, are you changing how the how the back-end experience is for delivering? If you're not thinking through the whole end-to-end process, you're gonna miss, you're gonna miss the opportunity to succeed. That has that problem is not changed as we look into AI and agenda because the the reflex is to jump fast and move quickly and do these things that that that look like fun and look cool and are gonna be impressive, hopefully to your board who's hammering on you to do the AI projects. But but you you know, we live in this extremely complex world and we cannot afford to build something that doesn't really take into account the complexity of the underlying enterprise and its technology. And when we try to do that, the the standard way, which is the way we've been doing it for 25, 30 years, we find out, whoops, you know, we we have these failure rates because we're still not thinking through the end-to-end process and the people that need to be involved.

SPEAKER_00 21:44

Yeah. I think that's just so important to highlight there that it's not necessarily a technology problem, but it's that the people that need to be in the room that understand the holistic business objectives, the process, the The data, the governance, all of those pieces, it has to be there in order for an initiative to uh to be successful. It was interesting. I was having a conversation with someone earlier today around um actually something that AI is doing. Is it because it makes it so easy to build, um people are like less and less stopping and asking if what they're building is the right thing. So it's almost exacerbated the problem of prototyping. Like the actually the friction that you had before with kind of slowly building gave you time to be methodical, to have those conversations, to bring other people in the room, to put together the requirements and the processes that were needed. But because you're building so fast, there's a plethora of all of these things being created that it like may not necessarily even be in line with what the business outcomes are. Are you seeing this as well?

SPEAKER_01 22:53

Absolutely. And and you know, and this is where, you know, really being cognitive of the sort of the cultural sociological aspects of change management is super important. Humans don't like change. We really prefer to do things the way we were comfortable doing that. Um, I mean, and that and that's true, true human nature. Changing too fast, particularly forcing change, is a is a sure way to get people's backup and to and to get resistance. So, so right away, you know, if it moving this concept of moving too fast is is is is is is a key part of the problem. The other problem is that, you know, and I alluded to this, when we're really looking, we need to more and more look across the enterprise at end-to-end processes and really enable technology change to allow the business to function in this much more efficient, productive way. There's a real problem on the enterprise side. This is an artifact of how enterprise software has been sold. There are very few stakeholders, particularly in the C-suite, who really have that end-to-end view. The CFO certainly sees order to cash as a as a financial problem. You could say that the head of supply chain sees that as a as a supply and logistics problem, but the two have to come together in a way that most enterprises aren't prepared to do from a from a strategic and from a uh uh uh um a stakeholder standpoint. So so you've got this problem on the on the comp on the corporate side, on the enterprise side, where we don't really understand how our processes are supposed to work together. You've got the technology side saying, hey, we've got this really cool technology, let's start building stuff. And you're almost doomed to failure because you're not, you you've got you've got a problem that's hard, that's not well articulated and well understood, and you've got a technology that's that's still sort of nascent and trying to trying to find its way and trying to figure out its limits and its possibilities. And it that's almost too theoretical to apply to an enterprise and be successful. So yeah, so it we're we see a lot of these mismatches happening. Um and and the other biggest problem is that the larger the enterprise, and we really can start talking at the 150, 200 billion, 200 million mark on up, the more the complex more the complexity of the of its technology infrastructure comes into play. There's so you know, there's thousands literally of applications at play in a very large company, hundreds in smaller companies. There, the the the the problem of interaction between all these different elements, the problem of common data models, the problem of of protocols and and standards is so endemic to this. And that that you you want to, you know, you want to solve these problems at the business level, you have to start at this deep technology level. And that sometimes is really just having a standard of what is the customer record look like across a complicated industry. That's a huge issue. Those are battleground problems because every department has a different view of what the customer should look like. And and they have a, you know, and and in the past they lived in a silo, they had, you know, they had the right to do that. And now we're saying, no, we need to have a standard. That's a political problem. It's not even a technology problem.

SPEAKER_00 26:22

Yeah. I'm curious, like I'm sure people listening to this are going to be thinking, okay, so you're you're surfacing all of these very good points around things that we need to be thinking through uh reasons why it's so difficult for uh AI, let alone any technology, to come in and disintermediate things, especially within larger enterprise organizations. I'm curious, in your experience, uh especially recently, who have you seen start to tackle this well? It may not be the fastest, but doing so in a methodical way that does take into account the people, the process, all of the things that we're discussing here today, who do you feel like is a uh their approach feels like the right way to start adopting this? Just like as you've worked within these enterprises or just talk to vendors that might have an outlook or a framework around how people should start thinking about this, um, you know, just like as you look to someone as an example for someone who's saying, okay, they're doing this in a sober way, in a realistic way. Um, I guess that's what I'm looking for.

SPEAKER_01 27:23

You know, you know, and let me let me maybe answer that in slightly, slightly different than than the way it was posed posed, which is that we're where's, you know, if you look at what is the foundation of a successful project in this in this world, and and particularly one in which we can say, let's apply some of this new new AI technology to it. The the the projects that have the best chance of success are the ones that start at the data level. Because to me, the real you know, and I've seen this time and time again, we run into a a, I mean, supply chain, uh particularly supply chain planning is a wonderful place to apply AI technology because these are complex problems with lots of variables. And wouldn't it be great to be able to both predict, you know, what to understand how to reconfigure your supply chain as the world of of tariffs and supply chain interruptions uh becomes the standard and the norm more and more in every every domain. Wouldn't it be nice to be able to do that and and to be able to add some predictability to a very complex process and therefore increase the productivity of how you how you order supplies? The problem is that in order to think about that and to really model that, which is what we want to do, we have to have consistent data. And that that is that that is that is absolutely the the fundamental starting point for any of these projects. And it's it's it's it's remarkable how often I've seen these projects get started, and then they go, okay, we've got we we think we understand the model, we think we want to do something really cool. Let's go find the data that we can use. And then you start looking around, and the data is so incredibly inconsistent and so incredibly non-standardized that you know they that you've just wasted all this front-end time, and now you have to go back to the the really fundamental. So I look at uh uh the the way in which both practitioners and enterprise software vendors are approaching data as this fundamental starting point. Let's get the data right. Any investment in data standardization, any investment in building a unified sort of data model for your core processes is going to pay off incredibly in every any kind of next generation project, in particular the ones that are AI-based and therefore need all this incredible data.

SPEAKER_00 29:51

In solving the data problem, or even not solving, but standardizing the data, you actually, it's like mandatory that you connect with like different people. You understand the processes that create that data. You it it's basically this um uh this forcing function to make sure that everything is correct before you start layering things on top of it. So that that totally makes sense to me. Um another another piece that I wanted to talk about is you know, you you wrote that um no technology succeeds alone anymore. It takes a technological village, not a bunch of overconfident tech bros, even if they are well-meaning. Um I am curious when you talk about village, like what does that look like uh in a company that is uh trying to adopt AI?

SPEAKER_01 30:38

In the beginning, the enterprise software community built software and try to sell it the way they built it. So they built it in this silo. There's an ERP product here, there's a CRM product here, there's an HCM product here, there's a supply chain product here. And those they they that's how they built it, that's how they sold it, that's how they delivered it. And the flip side of that, then that that model is recapitulated inside the enterprise. So now we have a silo buying, CRM asilo buying, ERP asilo buying, HCM. And and that that created, you know, it's a village is the wrong model for that. It's more like the sort of a you know, a medieval city-state model where every every every organization, you know, every organization has its own little fiefdom, literally, and they're sort of fighting for scarce resources and and trading with each other only as necessary or only as advantageous to themselves. The village idea is that, okay, you know, we've got to really break that down, get away from this competitive city-state model, and say we've got to actually function as a cooperative village. So, yeah, in order for us to advance our customer experience, we need to make sure that the head of supply chain understands that at the end of his or her supply chain is a customer buying a product. If we have to make sure the head of sales understands that the other end of what their value that they're providing is somebody making a collection of small bits and pieces that are gonna be assembled into a final product, we have to build that village mentality because without it, we're not gonna make those processes efficient. We're gonna have this fundamental disconnect where we try to we try to put a band-aid over each of these poorly connected components, these poorly connected silos, and hope that there's enough band-aids that we actually connect them well. And that that will never work. So, so when I look at how do we make change happen in the enterprise, I say I start with you have to take into account the reality of the fact that there is no monolithic enterprise at all. It's a collection of produ of tools and products. And the goal for change is to bring them together and to work with them. That's that's every, you know, the AI tools that that depend on a single solitary fundamental foundation of product and infrastructure will fail because that's not what's going to be in the real world. So I I think you know, this this idea that that complexity in infrastructure has to be dealt with as a fundamental aspect of success is one that's you know, I want to see a lot more of more and more the big enterprise software companies are saying, hey, you know, you don't have to be wall-to-wall SAP to be successful, wall-to-wall, Oracle to be successful. We will work with you, even if you're running Salesforce, even if you're running Workday. But that mentality is relatively new. Yeah. And and it's welcome, but it's it's still hard to change.

SPEAKER_00 33:48

Yeah, yeah, I can imagine. I think, yeah, so much of this, there's a very just interesting cultural shift just amongst that culture that a lot of uh leaders and IT leaders need to go through or application development leaders. And I find it interesting because in my conversations, you also have what I call legacy IT leaders who really haven't had or been had their hands on uh new technologies in a long time. They kind of have been, you know, um, I guess graduating through the ranks through some other means. Uh, and then you have the ones that are are just more forward-thinking and want to enforce change, maybe faster than they should. So trying to get practical, you know, for a moment. One of the things that we discussed, which was a key insight, is starting with the data, right? That's like one practical piece of advice that a technology leader that wants to adopt AI responsibly should start with. Are there other non-hypey things that you think people can start doing to lay brick by brick the foundation of adopting AI responsibly, especially in this like trust incremental way that I know you're advocating for?

SPEAKER_01 34:56

Well, you know, I now we now we may get into some somewhat you know side sidebar conversation because AI is such a big concept and it encompasses so many different technologies. So, you know, and and this is this is something, you know, and I caution your your your viewers and listeners to to think about this in more in the in the sense that AI, AI, you know, there's lots of AI around. There's, you know, you can I I have a I have a list here in my my cheat notes of 10 different versions of what is AI. We tend to focus on AI as as being the you know the the LLM kind of modeled AI of today that that is so over, frankly, overhyped. And we forget that so much of what we do on a daily basis in our enterprises sitting at our desktops is is moderated and improved by various other kinds of AI. This this web webinar podcast we're doing right now, all that image processing, all of that you know, correct correction that's doing that's going on in terms of in terms of sound and whatever, that's all AI. We have computer vision, natural language processing, and predictive analytics that are based on different AI. So I I I try to bring that sanity, that that concept down downline a little bit, because not every enterprise problem needs a large language model to solve it. Some of the best problems in the world can use other kinds of AI to do them. We look at logistics, we look at some of the issues again in supply chain with compliance. You're not gonna, you can't let a large language model keep you compliant. There's too much hallucination and too much uncertainty about what you will get out of that. So you have to you have to use a different kind of AI to do that. So I also really try to work with my clients to say, let's, let's take this, let's take apart this mandate for AI and really say, let's call it productivity or automation. Where can we improve the business and what is the most appropriate technology for improving the business? Not let's find a let's let's find a nail to beat this beat on with this new hammer we have. Um so so I I try to ground that in reality, and in particular because that's what you want to do when you sit down with your business stakeholders. A business stakeholder wakes up at the at 2 a.m. in the morning in the cold sweat because she's got a business problem. I always tell my my vendor clients, okay, so here's your here's your buyer, needs to go back to sleep. If they say, oh my God, what I really need is insert product name here, is that is that really what's going to solve the problem? Or is it, you know, do they wake up in the middle of the night and go, oh my God, I, you know, I need I need a large language model. No, they don't. And and so we we have to really also make sure that we are thinking in the, you know, putting ourselves in the shoes of these buyers, of these people with real business problems, because very often our conversations about AI in particular are about doing cool things and not about solving genuine problems. So I try to ground the world in that let's let's solve the problems that you know that that are really meaningful for the business, not necessarily meaningful for you know for the AI hype cycle.

SPEAKER_00 38:29

I think that's such relevant advice. And even though it sounds very simple, it's something, it's like a trap that everyone falls into, just like whatever the new thing is, being like, how can I apply this as a hammer and everything is a nail? Um, I'm guilty of this too. So I'll do it.

SPEAKER_01 38:46

We all do that. You know, it's natural, you know. And and look, we're all technologists. I, you know, I sit here complaining a little bit about how this stuff is applied. I love new technology. I I that's why I'm doing this. So I I find that that seduction to be as as as powerful as it is to anyone else. But then, you know, but but it's just years and years of seeing how it's been misapplied, looking at, you know, blockchain. I ran into a vendor who wanted to, you know, we were gonna we're gonna change the the sea to table supply chain by by going to this Vietnamese fisherman and tagging that fish with a with with a tag that gives it a blockchain identifier so that when it ends up at the on the table in New York City, you know that it's this particular fish that's worth the $60 a plate they're playing in. You know, no, that's you know, that's not gonna be the way to solve that problem. That's that's you know, there's so many things wrong with that model. But but that was that was great. We have a we have a business case until you unravel it by looking at what's the reality of teaching that fisherman how to do that and making sure that you you actually that tag actually stays on that fish, making sure, yeah. So so there's a lot of there's a lot in all of us that says we'd love to make this cool new stuff. And I always sort of make a career out of, yeah, okay, that's nice. I love it too, but let's let's be real.

SPEAKER_00 40:13

Yeah. I think in this conference, and I I could talk to you for hours, um, it's so funny because like one might not see that, but I can see you're you're a you're a technologist that loves technology, and you you just want, I think, so much to make sure that it's applied in a way that's going to be useful, right? That's what technology's highest and best use is, and there are certain realities around accomplishing that. Um, you know, as we start to wrap up, just a couple questions that I I had for you. Um, one is I I'm just kind of curious, like the the notion or the definition of a developer, I've asked this a couple of times, like is is changing. Yeah. What what do you think when you hear about like a developer? And like obviously more people have access to build than ever, more people have access to vibe code than ever. How do you think about a developer and who who is and should be building software today?

SPEAKER_01 41:07

Wow, just throw me a softball, why don't you? Um that's a that's a really, really big fundamental problem. And obviously, because you know, we've moved into this no-code world full, you know, full bore, we've we've we've got vibe coding going on. And yet, you know, the question is, is that is that is that enough? Is it enough, particularly again in the enterprise, to to hand, to hand this, and this is a CIO would certainly say this, you know I'm not gonna let those people touch, you know, do something that's gonna touch my back end and interfere with my security and interfere with with my maintenance and whatever. But I think I think at the more and more when I look at when I look at successful projects, it's it's a team approach that that gets things done. So there it's almost like there shouldn't be an individual developer. There should be a team that builds something because if you really, you know, and this is this is a very simple management, you know, one-on-one kind of thing. You really want to understand how your business runs. You need to, you know, walk around and talk to the people actually doing the work. And we and in the factory floor, that's a great place to understand what's wrong with your factory floor. In your service model, it's a great place to understand why you're why you're having retention problems uh with your customers. Um, so so that you you wanna, you know, you almost you really want, it's kind of like it's almost like, what's the metaphor? It's almost like putting on a play. Okay. You need a playwright to write a really great play. But in order for it to stage it correctly, you're gonna have to have this team to really put it together to make it something that's truly compelling to the audience. Um, so I I say the developer, the lone wolf developer, I'm not, I'm not really fond of anymore. I want the I want the development team and I want that that stakeholder engagement component. Now, and one of the things, you know, one of the things that more and more I'm involved in is trying to get these these enterprise architects and these developers to really think and talk the language of business. Because I think they they need to make that, they need to be that translator that I've been my whole career talking business and talking technology. And so the develop to me, the the perfect developer is someone who really has that curiosity to find out what is what really makes the business side tick and work more importantly than what's you know, what's the latest software tool I'm gonna master this week? Be cool.

SPEAKER_00 43:41

Yeah, I think that's uh fantastic insight. Um my my second question to you is um, you know, you obviously we talked about you having that love and that seduction for new technology. One of the questions I like to ask is what's in your AI toolbox? Can you share kind of what tools uh uh you're tinkering with today?

SPEAKER_01 44:02

Great. Thank you for asking the question. Yes. And in fact, it it fits perfectly to with my my the thing we were just talking about. I'm working with a with a with a company that they're they're a client of mine, Basis Technologies, and they built actually an AI uh an LLM-based tool. So there I am, you know, I'm I'm not just saying it's all bad. That really allows a business stakeholder, I'm sorry, an enterprise architect to to query the tool and say, I've got a business problem or my company has a business problem. How can I, how can I how can I engage my business stakeholders in the right way in order to understand what they need in in order for them to be become part of the process? And it's a brilliant little tool because you just literally say, okay, I've got a supply chain problem. We have a problem with inventory turns. It comes back and says great. Okay, so inventory turns, well let's start with you you're gonna have to talk to the you know the head of supply chain here's the questions you should anticipate, here's the answers you should be able to provide. And here's the three other stakeholders you're gonna need to talk about. And eventually this tool really builds a a web or a network that allows you as the architect or programmer to understand the business story. So I I find this is a a really you know this is a brilliant application of of of of the of new AI technology and one that that's in service to this exact problem of how do we engage technologists in the business side and vice versa frankly I I'm kind of excited about we're we're prototyping it at a couple conferences.

SPEAKER_00 45:36

I love it. That sounds like a really uh useful way to integrate AI into everything that we've been we've been discussing. And then um you know my final question to close is you know you've watched enterprise tech for 30 years and you've seen every hype cycle come and go. If there was a lesson from these past cycles that you wish AI builders or AI leaders of today understood right now, what would that be?

SPEAKER_01 46:00

I'm gonna go back to what I said earlier is that you know you have a you have a buyer you have a customer sitting up at 2 a.m in the morning in a cold sweat are you really solving their problem or you're solving your problem and and and and if you're solving their problem are you doing it in a way that's that's cost effective and that's engaging to them without that you're just playing around um but having that you know really grounding what you do in this practical world of I'm trying to solve a real problem than a problem that that exists today is going to to me be much more meaningful than let's let's figure out you know we're we're gonna be we're we're gonna be playing with this cool tool because it's the latest cool tool. So maybe that's that's sort of me reiterating that statement that let's be practical, let's be real about solving business problems and not just making making cool technology do cool things.

SPEAKER_00 46:52

Well I think that is uh the perfect place to end. Josh, thank you so much for your time. If people are interested in learning more about you and your thoughts and uh where they can find you where where should they go?

SPEAKER_01 47:03

Start with LinkedIn because that's that's my my main social media platform Joshua Greenbaum LinkedIn and you should be able to find me that's not too hard. And by all means drop me a line if you have anybody who wants to hear more talk more always always interested in these kind of conversations