Transcript 0:00 [upbeat music] If you don't know your numbers, you don't know your business. 0:15 I learned this the hard way with my first company, so I created NetSuite, the world's number one cloud business system. 0:23 NetSuite gives you the visibility and control you need to make smart decisions and grow with confidence. That's why companies who use NetSuite grow three times faster than the S&P 500. 0:35 Schedule your free product tour right now. 0:48 Hello everyone, and welcome to this week's Digital Leaders show, where we talk key issues in news at the intersection of business and technology in hopes of making you a better digital leader. 0:59 We're live every Wednesday at noon Eastern, 9:00 Pacific on LinkedIn and YouTube, and the audio recording is available on major podcast platforms the following week. 1:07 Today we're discussing data economics in the age of AI with former Dell EMC's CTO, plus other business headlines for the week. Thanks to our sponsor, Oracle NetSuite. 1:17 Their unified cloud business management solution helps organizations operate more effectively and provide real-time visibility into operational and financial performance. 1:26 ERPA AI members and followers get preferred pricing of 40% to 45% off NetSuite. Just go to erpaai.com/netsuite for more information. I'm Dan Goodstein. I head up ERPA AI. 1:37 It's the Institute for Robotic Process Automation and Artificial Intelligence, as well as the Digital Enterprise Institute. 1:42 We're a professional association, media, and peer-to-peer research network providing education, networking, and advice to over 130,000 members globally all around automation, enterprise AI, and digital transformation. 1:54 My co-host is Chris Sirdak, senior AI advisor to ERPA AI, author, consultant, former White House Chief Transformation Office advisor, partner for digital transformation at Gartner, and practice lead at EY. 2:07 Chris, good afternoon. How are you, sir? It's been a while. It's been a while. It seems like it's been a while. I, I, I- [laughs]... 2:14 don't know if it's an AI algorithm or what, but I keep saying this, the, the avalanche of headlines, uh, each week seem to be getting bigger and bigger. 2:22 So I don't know if time, more time is going by or less or what, but, uh, lots of, lots of things to talk about, and I'm, I'm, I'm really pumped for our, our guest today to get his, uh, his two cents on, on some of these things. 2:35 I think it was Lenin that said something to the effect some s- in some decades nothing happens, and in some weeks, entire decades happen. So I kind of- Yeah... feel we're in the middle of that. [laughs] Yeah. 2:45 I tell my, my colleague here, uh, you know, I had a, a busy month today, [laughs] Yeah... so it seems like that kinda, that kinda thing. 2:52 Well, a couple headlines that, you know, got my attention other than the US election, uh, but, um, you know, from a technology business standpoint, NVIDIA surpassed Apple as the world's most valuable company, uh, over the last week. 3:05 Uh, three point, I don't know what it is today, but, uh, I think it was like 3.4 trillion, 3.5 trillion market cap. 3:12 Um, so and that's, you know, largely fueled based on, you know, what we're talking about today, AI and, and, and all the, the compute that and the economics that goes along with that, right? 3:23 I, I'm just glad I got my GPU about a, a year before all this insanity happened. I, I don't need the extra capacity and I don't, you know, the waiting list has gotta be astronomical now. 3:32 Yeah, absolutely staggering the amount of money that's being put into this, and then I've also seen here and there reports that because the, the chips are being so used, like the compute on them is so heavy, that they may only last a year or, or 18 months. 3:48 So, a good time to be, uh, in the fab business building, building chips, I guess. 3:53 Well, the o- the other part of it is, you know, there's a lot of conversation about, you know, and we, and we're guilty of this too, you know, the, the pros and the cons and the risks and the security, whatever. 4:02 But, you know, one of the things that, you know, I think we've touched upon, you know, a couple times before, uh, at a high level, uh, but it kinda came out today is, is the energy, the electricity required for the d- these data centers, and, uh, Amazon and Meta are now facing, you know, more regulatory challenges to add to their, their list that we've talked about over the last few weeks now, uh, around, uh, their nuclear power ambitions and, and just other areas that they're just trying to figure out how to, how to fuel all this. 4:30 Well, I mean, Sam Altman flat out said about a year ago that we would need to double the electricity production capacity in the United States just to keep up, and this while we're trying to green everything. 4:42 So I mean, talk about, you know, a push me, pull you situation. Um, so these things are fantastic. 4:49 And, and, and then Sam, I think was the last week or the week before said, uh, that, you know, ChatGPT or, uh, OpenAI will be profitable as long as we don't count the cost of training. 4:59 Well, [laughs] okay, that's like 90% of the cost basis. I could be profitable too if I elim- if I took 90% of my cost structure off my books. 5:07 So, um, yeah, the economics of pursuing this, this particular approach are definitely getting interesting, I would say. 5:15 Um, and, and particularly if we have that much interest in, you know, climate change, global warming, et cetera, um, they s- seems to work very much, uh, in opposition to each other. Oh, I'm glad you used that term. 5:27 I'm a big fan of segues. That, that was a good segue into t- talking about opposition of each other. 5:33 So in the same week, I saw headlines around the US government signing up for OpenAI, ChatGPT licenses, uh, for the, for the federal government, um, agencies like NASA and IRS, you know, looking to harness the, the capabilities. 5:52 Um, uh, a- a- and, and, and also, and also Meta. 5:56 Uh, Meta's open source Llama models, uh, is gonna be, uh, also gonna be accessible to US government agen- Is, uh, in particular around the defense and national security of... 6:06 And then you see this other headline that the Chinese military is believed to have hacked Meta, uh, with its [laughs] new chatbot. 6:13 So, you know, m- maybe that's why, maybe there's that natural, uh, as you say, push and pull, but you know, really, uh, interesting juxtaposition this week, uh, seeing those, seeing those headlines in the same places. 6:25 Yeah, I think in the last two to three weeks I've, I've been pushing out on LinkedIn, um, there's been a number of studies just talking about the number of exploits that are inherently embedded in ChatGPT, and that you cannot... 6:37 You can't identify them until they execute, and even when they do execute, you may not even know it. 6:42 So the notion that we're gonna take, I mean, anything more than what would be on the public internet from any of our government agencies and, and put it into ChatGPT is, is truly terrifying. 6:54 But, um, you know, it, it's, it's sort of that, that wave of momentum driving in that direction. 7:00 We see this trillions of dollars going into this technology, so it's sort of inevitable we're gonna go there, but that's where I b- you know, put on my, my mission success rocket scientist has- hat and say, "Okay, 7:12 what could go wrong, and how are we going to protect ourselves when it inevitably does?" Um, my, my concern is that I don't think we're putting enough energy into that discussion. Yeah, just, it struck me as, you know, 7:27 be- being, being as critical of government in general as, as at least I've been, you know, about how slow and conservative they are, um, this seems fast to me. [laughs] You know, I, I'm, I'm not surprised- Yeah... 7:39 to see, you know, startups and, and mid-market companies doing it. Um, I, I evangelize a lot of the large enterprises, Fortune 1000s, to get into it. 7:49 Um, but to see the government embracing this, you know, to m- in, in my perspective, pretty quickly, uh, certainly in government terms quickly, uh, was a little surprising. 8:01 It, well, and it, it's complete back to opposition to what you're seeing in Europe, where they're, um, almost over-regulating, they're being almost over conservative. 8:11 Um, and that leads me to that old saying, you know, the US, the US, uh, invents it, the, the Europeans regulate it, and the Chinese or Japanese copy it. 8:20 And I think we're seeing a der- degree of that now, but to your point, the speed at which all this is happening, both in terms of controlling, from the European perspective, and embracing, from the US perspective, is actually pretty stunning to watch. 8:32 Um, it'll be interesting to see how it all shakes out in about 18 months, which I think is the time horizon we're looking at. 8:38 Yeah, and to, and to your point about European regulation, we, we covered, uh, kinda cell phone, Surgeon General's warnings around cell phone and social media use for, for kids a few weeks a- back, a few weeks ago. 8:51 Uh, back a few weeks ago, the UK is considering a new smartphone ban, uh, uh, for kids now, um, on the, on the European side. 8:59 On the US side, there was this, um, contract that the United States Immigration and Custom Enforcement signed with an Israeli spyware company that's gotten held up by the government, um, uh, maybe violating a, a- an executive order around privacy. 9:16 So, you know, really, and then you, and then you, and then you throw in just, you know, again, politics aside, but you throw in the, the, the election results, uh, into this conversation, and, um, I don't know. 9:27 Se- seems like a very volatile environment, but the, at least, at least the stock market disagrees with me. I think it's up 3, 4% already today so far. [laughs] Like, that may all be on the backs of Nvidia, I don't know. 9:38 Um, yeah, we're, we're coming into an age where our Trojan horses have Trojan horses with Trojan horses, um, and we're just hoping that, you know, no one sets a match to it all. I... It's gonna be absolutely fascinating. 9:50 The thing, the interesting thing with that cell phone ban is back in my compliance days, we always said the worst thing, the, the worst thing... 9:58 There's nothing worse than having no policy, than having a policy that you know you can't enforce. That's completely unenforceable, um, and so why even put it out there? So that, that's, that's an interesting one. 10:08 But it is another way of putting out penalties and collecting revenue, so, you know, there is that, I guess. [laughs] Yeah, interesting. Interesting. Well, our, our main topic today is data economics in the age of AI. 10:20 I'm very excited to welcome Bill Schmarzo. He's the former CTO at Dell EMC, and a globally recognized big data AI and data science influencer, educator, and practitioner. Bill, thanks for being with us. Thanks. 10:34 Thanks for having me. I, I couldn't hear Chris on the session, but I can see him now, so that's good. Did you pay extra to be in the mirror? I didn't pay anything for the mirror. Oh. [laughs] No, interesting. So 10:47 give us a, a little bit of a setup, Bill, uh, for those who don't know you, you know, a little bit of your, your background and, and experience, uh, at, at, at, at Dell EMC, but, uh, also just i- in general. 10:58 And then I wanna ask you about this, uh, this term kinda data economics. Sure. So, um, background-wise, I think anybody who knows me knows that first and foremost, I'm a teacher. 11:09 Um, I, I've been very fortunate throughout my career to be involved in a lot of different situations, so, um, I'm, at my heart, I'm a teacher. 11:18 I teach at, uh, I'm teaching again this fall coming up at Iowa State University. I'm an honorary professor at the National University of Ireland in Galway. I'm an executive fellow at the University of San Francisco. 11:29 I've written several books. But first and foremost, I'm a teacher. 11:33 And, um, even my role at Dell Technologies, where I'm on their AI spear team, allows me to work very closely with some of our customers in trying to figure out where and how do we apply this AI, gen AI, all these different tools to create new sources of customer product service and operational value. 11:51 So, um, that's my background. 11:53 I think what's, what, what I find, um, interesting as I share my background is I've, I've found myself transitioning from a data scientist to a data economist and also to a humanist, and we can spend more time talking about what, what the heck that might even mean. 12:09 Yeah. No, interesting. So, um, so, so set it, so set it up. When I, when I hear data economics, you know, my oversimplified brain is, is this about data, or is this about the economy, or is this about revenue? 12:22 Uh, so, so, so give us a sense of how you look at it. What's, what is this, what does this topic really kinda get involved with? So economics is about the creation and distribution of value. It's a value conversation. 12:35 It's a very broad conversation, more than just, you know, finance, more than just ROI. When you think about, 12:42 uh, value creation from an economic perspective, you've got business and operational and customer value creation, but you've also got areas such as cl- you know, suppliers and partners and communities and ecosystem. 12:53 You've got, um, things around, uh, diversity and society and, and, and ec- and environment, and even, obviously, ethics. So when we talk about economics, we talk about value creation. 13:05 And, and people get economics cr- co- confused with finance and accounting all the time, and let me give you a really simple example. So pretend you've got me 13:14 and an Uber driver s- standing right here, and we both go out and buy a $40,000 car. From a finance or accounting perspective, the value of that car is $40,000, right? 13:25 From an economics perspective, it couldn't be more different. My $40,000 car sits in my driveway. It gets snow on it. It gets leaves on it. It's not worth $40,000, right? Probably value is probably around 10 or 12,000. 13:39 However, the Uber driver's using their car to drive people around, make money, to deliver goods, make money, and when they're not using it, they're leasing it or renting it out to other people to drive it around and make money. 13:50 So from an economics perspective, the value of my car is much less than 40. From an economics perspective, the value of the Uber driver's car is much greater than 40. So 13:59 first off is we need to have a mindset shift around how do we create value, right? 14:07 Once I have that in place, and I don't mean to keep droning on here, if we think about value creation, there's a couple of aspects around data and AI that are very unique. The first one is around data. 14:18 Data as an asset never depletes, never wears out, and the same dataset can be used across an unlimited number of use cases at near zero marginal cost. 14:26 I call that the data economic multiplier effect, that I've got this very unique asset, that we've never seen an asset like this in our lives, that I, the more I use it, the more it's still there. It doesn't wear out. 14:37 And so as organizations think about data as an economic asset, the ability to tie that back to use cases and drive value for the organization's business, operational, customer, community value becomes very distinct. 14:51 So data has this data economic multiplier effect. 14:55 But we also have with AI the ability to build assets using AI, reinforcement learning, generative AI, analytic, you know, all the different kinds of AI tools, that actually the more that they're used, the more reliable, the more predictive, the more safe, the more valuable they become. 15:10 Just look at a autonomous vehicle. It's a learning vehicle. It's, every time it's driven around, it's getting more responsive or more predictive and more reliable. 15:19 And so we can use AI to build assets that continuously learn and adapt and become more valuable the more different scenarios and more different use cases they see. I'll stop for a breath. Does that make sense? Yeah. 15:33 Great, great. I- Chris, anything to add? Y- yeah, but I, interestingly, I made sort of a, a counterpoint to that in my second book, Jerk, from 12 years ago, and that was that, 15:44 um, we're shifting from a capital-based value generation in our society and value measurement to an information base. 15:52 And one of the fundamental changes and differences as we shift to an information base- based wealth creation is that money's fungible. You know, I can, I can trade a dollar for a, a won, for a, a, a euro, whatever. 16:05 Um, and $1 is really indistinguishable from another. Data's not like that. 16:10 Data's actually contextual, and the example I'll give is, like, if, if I know IBM's stock price next week, that's worth a lot more than knowing their stock price from last week. 16:20 And I think one of the challenges companies have in terms of governing data and managing data is we're still using money-based governance structures and rules against something that doesn't behave like money. 16:31 To your point, data's very different, and we need to treat it differently. You have... I mean, does that kinda resonate with you? Bingo. You, you said something really key. Data is contextual, right? 16:41 You, one cannot determine the value of one's data in isolation of the business. Yep. 16:46 And so in order, in order to understand context, you must start by understanding from an org- from your organizational perspective, how do you create value, right? How do you... What are your business initiatives? 16:57 What are you trying to accomplish, and what are your desired outcomes, and what are the KPIs and metrics against which I'm gonna measure those desired outcomes? So you're spot on is the data is contextual. 17:06 That's why, that's why you, distinguishing signal from noise in data is hard if you don't know the context of the problem you're trying to solve, and in particular, how I'm gonna use that data to create new sources of value. 17:17 So that word contextual is spot on. I, it, well, and, and that piece of information in possession of one department in your organization used for one use case creates value. 17:29 I switch it to a different person in a different, um, part of the business, and now I have a, a $10 million lawsuit 'cause they're not supposed to have that. 17:38 And I need to be in not only in, in constant track of what I'm doing with the information, but where it is, how it's being used, and most organizations are still struggling with that. Would you agree? Yeah. 17:49 Well, yeah, I agree. It's the, you know, the data governance is a problem, and part of the problem, in my humble opinion, is that we apply data governance to everything. It's like data management, right? The... 18:01 If you wanna, you wanna get a chief data officer all bent out of shape, ask them what their most important data sources are, and they'll look at you like you got lobsters crawling out of your ears. 18:09 "Well, I don't know," right? Right. 18:10 If we know what our most important data sources are- Then the data governance, data management, all the other data enrichment, data ma- metadata manage- it's all the kind of things that we do to round data to make it more valuable becomes much more, becomes much easier to do when we know which of our data states or data sets are most valuable and which ones aren't. 18:29 But we don't do that process, and so what happens is we peanut butter data management across all these data sets, and I'll be honest with you, 10% of your data sets probably account for 90% of your value, which means it's an ordinary- And 90% of your risk [laughs] 18:41 Yeah. Yeah. So, so focus on what's important. And how do you know what's important? You gotta be able to tether it back to what's important to the business. You ha- data does... 18:50 You said it earlier, data's only valuable in context. And so if we don't understand the business, we don't know which data sets are most valuable. 18:57 And if we don't know what data sets are most valuable, we're gonna peanut butter our resources across all these different data sets, and this is why no o- well, this is why all the business executives cringe at funding, funding a data management initiative. 19:09 But Bill, isn't that- Yeah... isn't that exactly the problem, though, right? I c- how do... It's kind of a, a, a chicken and an egg situation. 19:15 How do you know which data is more valuable unless you start uncovering that data? I mean, especially with organizations that are super siloed, you know, I'm sure you guys went through it. 19:24 You got lots of a- mergers, lots of acquisitions, then they divest certain companies, spin off certain companies. 19:29 How do you have a, a, a visual, you know, a visual of the entire s- you know, stack, all the silos of data? So s- don't start with data. It's like the conversation with AI. Don't start with AI. 19:41 Start with understanding how your organization creates value, right? Data and AI is not our endpoint. Data and AI is a means to our endpoint. 19:48 So start the conversation understanding what's important to our organization, what are we trying to accomplish? 19:52 So I talked to a couple hospital chains, and they're all about how do I improve patient treatment effectiveness, right? How do I become more effective in treating patients? 19:59 I deal with banks, and banks are all focused in on customer acquisition, customer retention, customer cross-sell, right? 20:04 Understand, and they can tell you that, by the way, if I can get my customer upsell from where it is to industry average, it's worth $65 million. 20:10 I like talking to customers who have problems, who have, you know, an M with a bunch of numbers in front of it, right? 20:15 So start by understanding what's important to the organization, what they're trying to accomplish, what are the desired outcomes, what does good look like, how are they gonna measure it, 20:23 and then figure out, well, what data do I need, right? Problem, like I said, is probably scattered all over God's green earth, right? 20:30 Well, now I've got a strategy that says, okay, I need these data sets that are scattered all over the place. 20:34 Maybe I build a data lake, data lake house, data department store, b- data outhouse, whatever you're gonna build. 20:39 But now I need, I put in place an architecture that allows me to bring that data together in support of that. And by the way, 20:46 knowing what problem I'm trying to solve means I'm not trying to build analytics for everything. 20:50 I'm trying to build analytics around predicting who's my most at-risk customer, what's the most likely response to them, what's the right, best promotion, what's the best cross-sell, what's the next best action? 20:59 And I can get very, very granular in my decisions with data and analytics by knowing where it is I'm gonna go. And we're just stealing from Stephen Covey, who many, many years ago, hey, begin with an end in mind. 21:10 Well, the end in mind for most organizations, whether they're profits or nonprofits, is how do I create value for my constituents? [coughs] So how do you- So pretty much- How do you quantify it, Bill? 21:24 You know, if you look at- Look at-... you know, data in general- Data in general... what's, here's the economic value of, of data. 21:31 I, I even heard a story recently, I'm not sure if it was from you, Chris, or someone- From Chris, yeah... where they said, you know, after 9/11- You know, after 9/11... 21:37 insurance companies didn't wanna, uh, insure data, they didn't wanna insure IP. Uh, how do you put a tangible number around that? Again, it gets back to understanding what the organization's trying to accomplish. 21:49 I mentioned one customer is trying to improve, uh, cross-sell effectiveness. They estimate it's worth, you know, $60 million a year for them. Okay, that's great. What data sets do I need to support that? 22:00 And I actually outlined this [laughs] a very boring book, right? The Economics of Data Analytics and Digital Transformation. 22:05 My children said, "The only way that I could have made this book more boring is to somehow put the word Brussel sprouts in there, Dad." [laughs] Right? 22:11 But it, but it outlines a process that says, okay, if you understand what it is your organization is trying to accomplish, what's important to them, what are their strategic initiatives or strategic, strategic opportunities, what's the value of those, right? 22:25 Reducing, um, unplanned operational downtime if you're a theme park. You know, they can, they can tell you what it costs on an hourly basis to be down. 22:32 Okay, so we start with understanding how it is the organization is defining and measuring value, right? 22:39 Then we look and figure out what data do we need and what analytics we need to build, and the analytics we're gonna build is around the data, the decisions that we're trying to make at that higher level. 22:48 It's not a very hard process, but it requires us to do something that we don't like to admit, that the most important part of our job isn't the technology. It's what we're doing to provide value at the top end. 23:02 We, we love the fact that we can make this gen AI tool, you know, speak to me like Elvis Presley, or read a Schmarzo introduction in the voice of Dr. Seuss. That's really cool, right? 23:11 Well, it's kinda like making a dog walk on its hind legs, right? It's interesting, it's fun, but I don't want that dog bringing my beer to me, right? 23:19 So not knowing what it is you're trying to accomplish means that we- we're not clear in where and how we use all these powerful tools to help us to achieve that. 23:29 And yet everyone's suffering- And yet everyone's suffering... from FOMO that I better use it or else, and I, you know, I might be missing out, so let's go adopt this stuff. 23:37 And, and I, I liken it to what we tried to do with RPA about 10 years ago. People didn't get value out of RPA, and now they're moving to AI. 23:44 I'm like, "That's like trying to learn how to juggle with chainsaws instead of red rubber balls." Yes. Yeah. It's, it's, we, we let the AI tail wag the value dog. Mm-hmm. Right? And we, we get so fixated... 24:01 And again, data and AI, for the 99.999% of companies I talk to, is not the endpoint. It's a means to an endpoint. 24:11 So if we clarify up front with them how, you know, that endpoint, what you're trying to achieve, what are your stakeholders trying to achieve, what are their desired outcomes, right? 24:19 There's a, a process, I call it the art of thinking like a data scientist. I teach it at the university. 24:23 It's a very simple process, but you spend a lot of time upfront understanding the business before you ever put science to the data. 24:35 And would you agree that that's largely a context setting, context discovery sort of exercise? Because what are you even trying to achieve with this? And what does goodness or value look like? Bingo. 24:47 It's, it's all, it's all contextual, right? What, what is also interesting, if you talk... If I talk to three different or four different or five different banks, they're all trying to do different things. 25:00 They're all in the same industry, but what's important context for bank number one may not be relevant for bank number two, bank number three. They're all different levels of maturity. They're in different markets. 25:08 They got different skill sets, right? So you, the, the idea that I can build a banking solution that addresses all of them is, is a, is a, is a farce, right? 25:21 You, you, you need to be able to understand what it is that bank, that hospital, that manufacturer, that theme park is trying to accomplish. What's important to them? And then we, we're carpenters. 25:33 And we, once we know what we're trying to build, are we gonna build the Eiffel Tower or build an outhouse? Then I know what tools I'm gonna need, what wood I'm gonna need, what data I'm gonna need. 25:42 Bill, you know, I think back- Amazing... sorry, since Chris, you mentioned RPA, you know, I think about RPA as, you know, a lot of people thought it was gonna change business models. 25:52 Really, it kind of optimized, in, in a perfect world, it optimized existing business processes maybe not as much as we wanted it to, but that was, that was kind of the idea. 26:02 Do, do you see this kind of new wave of AI more of the same optimizing business process or is more about really changing paradigms and creating new business models? 26:12 Well, I think Chris sets the topic very well is that, is that what we're doing with AI is, uh, here, I'm in Iowa, right? And they're paving the cow path, right? 26:21 We're taking existing processes and just making them faster, maybe not better, right? And the need to stop, pause, understand what we're trying to do allows us to actually re-engineer that process, right? 26:37 And this, by the way, this is where a lot of design thinking tools can come in to really understand the customer journey, what is it they're trying to accomplish, how they're gonna measure success, the key decisions they're gonna make. 26:46 And then figuring out if that is what they're trying to do, I've got all these great new tools and capabilities available, I can re-engineer the process. 26:55 Maybe skip some steps entirely, eradicate some steps, combine some steps, right? 27:00 But no, the easy thing you wanna do is I got an existing process, I'm gonna, I'm gonna try to make it faster, and we know that faster does not equal better. 27:07 This is the whole gen AI conversation that Chris and I have been having on, on social media is that just because I can do, uh, a productivity focus on gen AI so I can do things faster is missing the bigger prize. 27:21 It isn't about being faster. It's about being more effective, right? It's about being, it's about doing more with less, not just doing more. Yeah. 27:34 I mean, if, if you're doing something incorrectly and you're doing it faster, you're doing it incorrecter-ly-est or... I mean, and, and, you know, I did re-engineering back in the '90s and, and we hated it. 27:44 We didn't wanna do it, but that was, that was when we were doing things at a thousand times less fast with a thousand times less capability. You've gotta re-engineer this stuff to make any of this stuff worthwhile. Yeah. 27:57 It, you're exactly w- w- you've got new tools. It kind of reminds me of the NBA bringing in a three-point line, right? 28:06 When they brought in the three-point line, it took a few years for companies, for, for teams to realize, "Wait a second. Making a shot two or three feet back is worth 50% more?" 28:15 I mean, the economics of the three-point shot were right in front of us, but yet it took organizations, it took professional teams a lot of time to realize, "Wait a second, we need to reframe how we think about this from a value perspective." 28:28 They, they re-engineered their businesses, their, their plays they were running. 28:31 They re-engineered the players they were drafting and developing to take advantage of the fact that this shot back here that's a few feet back is worth 50% more. 28:40 And we're seeing this, by the way, with the Boston Celtics who are, who are on a, on, on... And we saw this with the Warriors too, the Golden State Warriors on a... 28:47 They, they launched three-point shots like they're drunken sailors, right? Because, but they're, they're not doing it just randomly. They're not shooting just more. They're shooting them more effectively. 28:56 They're, they've re- re-engineered their offenses to think more effectively about how do I create space? How do I have to... I gotta have at least two, if not three shooters on the court at all times, right? 29:06 So again, there is, I think the initial rush is we take these new technologies and we try to just force fit them into our existing models, and they, we, you know, sometimes it gives us faster activity, but it doesn't re-engineer and give us better. 29:18 It doesn't give us more effective. 29:19 So I think that that's probably the problem with RPA as well is we just rush to it without stopping to think, "Wait, how can I use this differently to achieve maybe better outcomes more quickly, more effectively?" 29:33 Bill, is there any suggestions you have as far as frameworks or models as organizations are looking to integrate AI and, and leverage data for more competitive advantages? 29:43 Well, it, it, it's, again, I don't wanna sound like a broken record c- 'cause I, I probably am, but the AI journey has to start by understanding what it is you're trying to accomplish, right? 29:52 What is, what are your desired outcomes? What are your business initiatives? 29:55 Start there, and then once you have that, then I can start thinking about, well, across all these different AI options and, you know, machine learning options and deep learning options and all, the plethora of technologies available to me combined with the plethora of data that I either currently have or data I could get now allows me to think about how do I orchestrate those in a way that creates value? 30:18 The, the data science process is critical to this, right? 30:20 You've gotta have a data science process, but you have to know where you're going in that process, and going, knowing where you're going, by the way, means you need to understand how your stakeholders, how your constituents define and measure value, which is why my methodology, I blend data science with design thinking. 30:36 My data science teams were a lot more effective in developing models when they knew how the stakeholders were gonna actually use those models and operations to make better decisions. 30:46 So to me, again, it starts, it always starts with understanding how do we create value? How do our stakeholders measure that value creation effectiveness? What are the key outcomes they're seeking? [coughs] Excuse me. 30:58 [coughs] I'm back. [laughs] It's that old saying, if you don't know where you're going, you don't know where you're getting to, right? Yes, exactly. 31:10 That's- People are, they're out in the middle of February and they're like, "Let's go somewhere. Where are we gonna go?" Let's buy me some AI, baby. [laughs] Oh, all right. So great segue there. 31:22 Uh, I know you and I have been heavily engaged in a variety of discussions about the, the wonderfulness of generative AI, and I get accused of being like a, a technology Luddite a lot, and I would argue I'm the exact opposite. 31:34 I'm trying to help people get something useful out of the technology 'cause I love it. 31:38 Um, you know, what are your general thoughts about where we're s- where we're seeing a generative AI particularly going, potentially valuable use cases, and maybe some cautionary tales based on a lot of, you know, I've gone back and forth on that. 31:52 Well, first off, Chris, I, I have great admiration for the fact that you're willing to be a naysayer in a world of lemmings, um, where everybody wants to just jump out on this stuff and, you know, whether we're gonna implement ChatGPT or Gemini or Copilot, whatever it is, right? 32:07 There's this rush to really put this stuff out there and as you have articulated it so well, so well, is that GenAI is a rush to average. It's a rush to regression of the mean, right? 32:19 It's just that the technologies that underpin that, I can't explain it as well as you do, basically leads us to average. And so we're getting maybe higher degrees of average, but we're not getting innovation. 32:31 And to me, the real potential for GenAI 32:33 is not just making us more average, but taking us and leveraging some of the human characteristics around curiosity and exploration and creativity to really leverage GenAI to drive new sites of innovation. 32:46 So to me, there is that, that there's an opportunity to drop, to jump over this chasm, leap this chasm where we're so focused on productivity that we're not thinking about how, how do I use this tool to, to be more creative in how I'm approaching markets, how I'm reaching customers, how I'm expanding, how I'm building products, how I'm minimizing risk. 33:05 I mean, think about all the difference. I had a list of what I call the nine innovation prompts that you should think about when you're engaging. 33:11 Like things around, you know, these tools are great at if you ask it about, well, what am I not considering? What risk am I exposing myself? What are some of the potential unintended consequences? 33:20 It's gonna come up with all kinds of ideas. Some of them may not be useful, but some of them will be. 33:25 And being able to, to move the, move the needle from productivity, doing things faster, to innovation, doing things more effectively, becomes, I think, the real winning prize for this GenAI tool or tools. 33:40 You bring up a critical point. Some of the responses will be less ideal, some will be more ideal. 33:45 But you as the consumer of those and evaluator of those ideas, you need to be smart enough to know the difference, and that's the challenge. The vast... 33:53 In, you know, we're not, all of us aren't smart about everything, so not all of us are gonna be able to make that evaluation at the time you get a response. Is this a good response or not? 34:03 And the example I, I give is all these lawyers that were, you know, getting case law and stuff for their depositions in court, thinking that this thing was spouting off actual case law when it was just making it out of, you know, thin air. 34:16 And some of those people are, you know, enjoying, uh, no longer being attorneys as a result of that. [laughs] Yeah. So, um, we, we didn't educate people very well, right? We didn't... 34:30 We, we failed in letting them know what the risks were. We unleashed this tool on them. 34:36 We gave them a really powerful, you know, powers tool to use without any sort of safety on it, there are people are chopping off fingers and cutting off toes and such. Um, 34:46 it worries me, Chris and, and Dan, because even now, now that we know the horror stories, we still haven't prioritized literacy, GenAI literacy and data literacy very high. 35:00 We as organizations spend hundreds of millions of dollars on technology, and they spend a fraction of that on educating their people. We've, we've got this, this balance is totally off skew. 35:13 And as you said so well, Chris, if we aren't training people how best to use these tools, we'll end up with all kinds of bad decisions, unintended consequences, confirmation bias. We end up with all kinds of bad things. 35:26 It starts with education and awareness, and we as a society, we've never... 35:31 Think about al- almost any organization you've been a part of, education has always been a small fraction spend versus the money they're spending on technology. It's a great point, Bill. 35:41 I, I didn't think of it like that, and you, and you're right. Um, you know, so, so much of the traditional IT has been to stop, block, slow down innovation. 35:53 Uh, back to the AWS days where, where CIOs, CTOs were using personal credit cards to buy cloud space, right? 36:00 Instead of, instead of educating people on how to do it and, and maybe managing it and governing it, they were just trying to, you know, block it. 36:07 Now you have, you know, AI democratized to a level where, yeah, you can, you can create a video of, what was it? Your, your dog walking on its hind legs. I like that one. 36:15 Um, so yeah, yeah, and, and, and it, it's, you know... I, I, I said to somebody, you know, recently it's reached the point of inevitability where, you know, your employees are using it. 36:25 It's like bring your own device, right? They're using it. It's, it's not a matter of if, it's a matter of when, so why not teach them how to do it? What I also think here is really interesting, um, Dan, in this topic. 36:40 So I was, I, I'd been working with a couple of local high schools On their curriculum for their teaching. And these schools are going through, are gonna go through a major, major re- renovation, re- re- retransformation, 36:58 'cause I believe that the days of measuring student success by how much they can memorize and regurgitate is over. We'll never be able to memorize and regurgitate. I got my own GPT, the Dean of Big Data GPT. 37:13 It knows about blogs I wrote seven, 10 years ago. I was like, "Oh, wow, I wrote that? That was pretty good," right? I'm never gonna beat it. 37:19 It's like I've got a, I got a $1.99 calculator that does square root faster than I am ever gonna do it. Right. Right. I've given up trying to deduce square roots. 37:27 It's like the Einstein thing, that he didn't know his own phone number. I don't know if it was true, but I always heard that, that story. 37:31 And I, I remember being in, you know, physics class, and them handing out a, a reference chart on the first day, and I was like, "Wait, what? [laughs] You're gonna give us all the equations on day one? 37:41 We can use this on a test?" I remember it just boggled my mind because we weren't, we weren't educated that way for so long, right? You're spot on. 37:49 And so the, the education system has to focus not on memorization and regurgitation, but on application. Teaching people how to apply this to deliver more relevant, meaningful, responsible, and ethical outcomes. 38:05 That is, the training is about how do we apply. Yeah, you're gonna teach them how to use a tool to do that, but you're... It's not about memorization. 38:13 It's, I like to say, and this is probably a, a statement that's probably not true, but what you know isn't nearly as important as what you're learning, 'cause the world is changing so quickly. 38:24 And so that is not how our school systems work, and so we've, we're, we've got high schoolers, middle schoolers, even college students who have been taught an old-fashioned way using standardized curriculums and standardized books and standardized class size with standardized testing. 38:40 Everything has been standardized, right? And in a world of standardization, from a human perspective, it kicks our butt. 38:45 So our education system needs to do a total transformation before, as a society, we're gonna be able to benefit from all the things that these different AI tools can do for us. 38:56 Bill, I've had a couple of, uh, international universities approach me and say, "Hey, you know, we wanna build an AI master's program. You know, what classes do you think we should have? 39:04 What, you know, should we base it on Gemini or ChatGPT? What, what courses should we have?" And I've said, "Well, you need to have philosophy. You need to have civics. 39:13 You should be teaching the Socratic method," and I couldn't care less what platform you choose. I couldn't care less what technology you choose, 'cause we'll have a different one in six months. Just wait. 39:22 But to your point, the, the, the more we automate away our humanness, the more desperate we are gonna be to, to hang on to- Right... 39:29 the fraction of our humanness that remains, and that's gonna be the most valuable part of what each of us brings to the table. And I think that- Yeah... 39:35 focusing w- then, and that emphasizing we see more and more as the days go by. Amen, Chris. Amen. Now, think about what it is that makes us humans. We've a natural curiosity we're born with. 39:48 I watch my one-year-old grandson taking apart things and playing with things, right? 39:52 I remember when I was very young, I took apart my dad's radio and put it back together, and I said, "Hey, look, Dad, there's two parts you don't need." Oops. [laughs] Guess who did, right? You broke it. 40:00 But we have that, we have that natural curiosity. Don't stifle it by throwing us into standardized classes and standardized curriculums. Nurture that curiosity. 40:09 Nurture that ability to imagine, to explore, to, you know, to, to drive that, the creative- creativity necessary to drive innovation. But so our school systems have to change, I think. 40:20 And even our systems of tra- of teaching inside of uni- of organization, but just in the school system though, we need to change in how we teach people to leverage those things that make us more human, 'cause we can't compete with the, the $1.99 calculator that can do square roots faster than I can ever do it. 40:38 Nor should we, right? 40:41 So Bill, I'd be remiss if I didn't ask you the, you know, the obvious question about balancing data privacy concerns with the potential upside of, of, you know, e- e- economic potential with this stuff. 40:54 Where, where, where are you on this? If I had to put you on a, on a spectrum of super conservative or super aggressive or somewhere in between, where are you on that, uh, spectrum? I'm a moderate. 41:05 I think you find it that the challenge that we face as society is that we have a hard time doing this. We, we get ourselves into this or that conversations, like monetization or privacy. No- Either way, yeah... 41:23 it's, it's an and conversation. It's monetization and privacy. Drove me nuts during the pandemic, for example, where the conversation was healthcare or the economy. W- why do I have to choose? You're right. Yeah. 41:37 Why can't I have healthcare and the economy? How, why can't I bring these two needs together? 41:42 This is, by the way, this is where design thinking techniques are really valuable in helping to do, to bring together conflicting perspectives and come up with something that's even better than what you had originally. 41:52 So I, again, I, I, I find myself... You, you, you've gotta have monetization, you gotta be able to create value, and you gotta have privacy. 41:59 It's kinda like [coughs] I wrote a blog to talk about, from the AI perspective, which master do you serve, shareholder or stockholder? Or, or, or stakeholder, right? Shareholder or stakeholder, right? 42:11 Do you serve the, the ROI, the financials? Do you serve the ac- the, the community? And the reality is you have to serve both. 42:19 I gotta serve the, the shareholder who wants a return on their investment, but I also need to share the stakeholder who is gonna try to benefit from this. It isn't either/or, it's an and conversation. 42:29 And we as, whatever it might be, we struggle with these and kinda conversations to bring together very different perspectives. 42:36 People, having conversations with people you don't even, may not even like, but understanding what their rationale is and how do I blend different perspectives together to drive something that's more creative and more innovative? 42:48 If you play every game like it's zero sum, every game you play will be zero sum Well said. Exactly. It, it's all a, it's all a question of your approach. Yeah. 42:57 If you wanna, if you wanna change the game, change the frame, change how you're gonna measure value, change the conversation. And then, and, and, you know, we, we wanna touch a bit on ethics. 43:07 That's also the risk question. If someone thinks that they're going to get a high value return on any of these investments without having attended risk associated with it, that defies economics. I mean, it's... 43:19 If, if, if you could get value without making any kind of risk, then you should have already done it. So, you know, I'm, I'm certainly not suggesting people don't adopt the technology because there's risk involved. 43:29 I say know what the risk is so you can properly manage it. I mean, is that kind of along your thinking as well? Bingo. Yeah. You, you gotta... 43:37 Part of the exercise we go through when I work with clients is to not only understand when we, when we target a business initiative, we talk about the risks of failure, right? What are the ramifications if things fail? 43:47 Because if I can identify those brainstorm- those risks, I can start to identify KPIs and metrics around which I can monitor those risks. I can also start to brainstorm unintended consequences, right? 44:00 What happens if this is really successful? What are the unintended consequences? I was working with a large construction firm. They wanted to increase the number of design wins, right? They wanna increase it by 20%. 44:09 That's great. What happens if you're successful? Do you have enough people to handle all the work? Are you gonna find the quality of the work without enough suppliers, right? 44:17 You start thinking through all the ramifications of unintended consequences, and if we can brainstorm them, then we can start identifying the KPIs and metrics around which I can manage and monitor those. 44:27 So yes, we, we have to balance risk and reward, benefits and impediments. That's just the way it is. And by the way, when you do that balance, when you look across those synergies, that's where innovation happens. Cool. 44:42 Very cool, Bill. Thank you so much, Bill. I really appreciate you joining us. I think we could go on for, like, four or five hours, so maybe we'll, uh, we'll schedule a follow-up with you if you're, you're open to. 44:52 But, uh, it's been really helpful. Thanks for joining us. Thanks, Dan. Thanks, Chris. Thanks, Bill. Great conversation, Chris, and, uh, I didn't realize you guys actually worked together, like, 20 years ago. 45:02 Is that true? 20 years ago, yeah. We, uh, uh, it was almost like being in prison together. No. 45:07 [laughs] Yeah, so we go way back, and it's great to see, um, you know, uh, both of us progress along the lines and sort of it's the five blind people touching the elephant. 45:17 They all s- kind of s- think they're touching something different, but it turns out we're all just getting a better understanding of the same problem. Oh, yeah. Great. Small, small world. 45:24 Well, thanks for joining us, everybody. I hope that was helpful. We're back Wednesday at noon Eastern, 9:00 Pacific. 45:29 In the meantime, if there's anything you're working on and, and want some free advice or just to learn some art of the possible from our team of SMEs, go ahead and schedule a free networking call at erpaai.com/networking or go to erpaai.com for more resources. 45:43 In the meantime, have a great week, everybody. And Chris, good seeing you, buddy. Take care. Have a good week. 45:48 [outro music]