Transcript 0:00 [upbeat music] If you don't know your numbers, you don't know your business. 0:11 I learned this the hard way with my first company, so I created NetSuite, the world's number one cloud business system. 0:18 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. 0:31 Schedule your free product tour right now. Hello everyone, and welcome to the Digital Leader Show. 0:47 We're live each Wednesday at noon Eastern, 9:00 Pacific, talking key issues in news at the intersection of business and technology with a special focus of automation, AI, digital transformation, in hopes of making you a better digital leader. 1:01 Today we're discussing enhancing automation outcomes with digital workers, with industry veteran and former PwC Intelligent Automation leader, Mike Engel, plus headlines for digital leaders. 1:10 First, a quick thank you to our sponsor, Oracle NetSuite. 1:13 Their unified cloud business management solution helps organizations automate core processes and provide real-time visibility into operational and financial performance. 1:21 ERPA AI members and followers receive special pricing and resources. Go to erpaai.com/netsuite for more information. I'm Dan Goodstein, I head up ERPA AI and the Digital Enterprise Institute. 1:33 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. 1:46 Check out erpaai.com for more. 1:49 And my co-host today is Chris Serduck, an author and advisor who formerly worked in the White House Chief Transformation Office, was Partner for Digital Transformation at Gartner, and Automation AI Practice Lead at EY. 2:00 He currently leads our fractional Chief AI Officer offering. Chris, welcome, good afternoon, and, uh, let the games begin. Lots going on this week again. Amen. Yeah. 2:13 Uh, we all survived May the Fourth be with you, so that's good. And, um, more, uh, more AI news to go over. Yeah. 2:22 I, I'm, I'm s- I'm trying really hard, as, uh, a lot of people may know from, uh, past shows, to, to find digital transformation news that is not AI related [laughs] but I think it's, I think it's quite impossible, uh, these days. 2:36 I mean, I mean, part of it's probably that the AI algorithms have, uh, adapted to what I search for, but, uh, you know, if you look at headlines, uh, it's, it's part of, uh, most of them. 2:47 Uh, so couple, couple that I highlighted this week, um, uh, new, uh, research from Microsoft and LinkedIn's Work Trend Index says guess what? People are using AI at work whether their bosses know it or like it or not. 3:03 Uh, there's some, uh, food delivery wars are heating up with Uber and Instacart. Apple debuts its new chip in iPad Pro, no surprise, geared towards AI. And, uh, Amazon is, uh, working on a, a chatbot. 3:19 Actually, they've made their, their Q chatbot available, uh, this week for AWS. So lots to unpack there. Where do you wanna start? I, so I actually looked at the premiere for the, the Apple event. 3:33 And, and I also think that a key part of that, you know, great, so we have a slightly thinner iPad, um, I found it interesting that they're doing $100 billion buyback of their own stock. 3:46 I don't know if you caught that as well. Um, and you really, it really does bring the question, is the company capable of innovating any longer? And, and how long has it really been since they have innovated? 3:58 Um, so I, I, I think it's gonna be interesting to see what Apple does and how it turns out five years hence, um, 'cause it really feels like they're, they're just incapable of innovating any longer. 4:10 I mean, how many people are, are you seeing wearing the Vision Pro thingies? Practically no one. 4:16 Well, and I think part of it is, you know, uh, we, we, we always kind of lived with this expectation, I think, that the, the Googles and the Amazons of the world, the Metas, you know, were, were the ones that were gonna innovate, right? 4:30 The most, you know, the last couple iterations of the internet and advertising and everything related to it have been, you know, innovations at Google. 4:39 Now, by some accounts, Google's kind of behind the, the eight ball and, you know, it took a startup like OpenAI to, to really kinda, uh, you know, shake things up. 4:48 But I, I think part of it's just this, you know, it's, it's democratized to a, a level that, um, you know, we, I, I don't think we've ever seen before. 4:56 I mean, I mean, this, this, this report coming out of the, the Work Trend Index said that, you know, 73% of Boomers are using AI at work. Whoa. It gets more as you go Millennial, Gen Z, and, uh, right? 5:13 85% of Gen Z-ers, right? Uh, and you know, the other, the other kind of tidbit that w- that came out of this was, you know, two-thirds of business leaders said they wouldn't hire somebody without AI skills. 5:24 70% said they'd rather hire a less experienced candidate who has AI skills than the alternative. That's a crazy paradigm shift, right? Especially if you're in the- Yeah... 5:36 interested to know what industries that was, right? But- Well, and, and, that, that brings to mind the conversation that you and I had with some executives yesterday talking about this, where, 5:46 great, we want someone with AI skills, but what are those skills? And how would you know that they have them? 5:53 That, you know, so if, if you haven't gotten at least some of that experience yourself, you're in no position to judge whether or not someone else has the necessary skill 6:02 Which kind of cracks me up when, you know, you'll see the job posting, "We wanna hire an A- an AI engineer with 10 years of experience with LLMs." Okay, they've only been around for seven years. 6:11 [laughs] So, so you're kind of off to the wrong, o- off on the wrong foot to begin with. So I, I think understanding what those skills are is gonna be even more of a challenge. 6:22 And, and I, I think an inevitable, and I can't wait for our guest to get into this a little bit, um, inevitably what more and more automation does is it makes the need to do the fundamentals, the hard parts of work, even more emphatically true. 6:40 Um, you know, so the, the more we avoid doing the hard stuff like writing requirements and doing adequate testing and so forth, um, these technologies just mean that those fundamental pieces of work are even more important going forward. 6:54 Yeah, and I think you have a, an AWS situation happening with, you know, AWS [laughs] but also with the others, right, where, you know, people are just kinda, you know, like, you know, years ago, right, it was, you know, CIOs just swiping their personal credit cards to get access to the cloud because they didn't wanna deal with procurement and, you know, jump through all the, the hoops. 7:15 I mean, you have that now, and I think this, this announcement by Amazon around, um, you know, uh, this kinda enterprise-focused GenAI, uh, becoming available this week. 7:25 It's aimed at both business leaders as well as developers, different versions, and the prices start at, like, $20 a month. 7:31 I mean, can you imagine, like, if, if, if you were, if you were putting cloud services on your credit card, $20 a month? I mean, obviously these guys are hoping for some, some cloud lift, uh, as a result of this. 7:42 But, uh, you know, we, I think we talked last week that, uh, like, AWS is up, like, 20% year over year, um, largely because of this, this AI boom. Well, and, and, and I think 7:54 it, it takes me back to a lot of the compliance work I've done for, for many years. And, and, you know, some companies are saying, "We're not going to allow the technology," et cetera. 8:03 The worst thing than not having a policy... There's no nothing that's worse than not having a policy at all, and that's having a policy that you can't enforce. 8:11 And so I, I think recognizing that AI is here, it's coming, there's really not a lot of ways to control it. Um, you know, uh, uh, we're working on AI governance and what that means and so forth. 8:24 This is among the challenges that we're trying to address. But it, it's not going away, so then the question is how do you properly manage it? Yeah. Yeah. Well, and I... 8:33 And, and the other, the other shocking headline to me was that the, uh, the, the, the US government is looking into ways to use digital twins to help supercharge the manufacturing, uh, of AI chips. 8:50 I mean, there was a time where, you know, you didn't hear any of these, these buzzwords coming from the government. [laughs] Yeah. And I know you spent some time there, right? 8:59 I mean, I mean, I, I continue to be shocked on a weekly basis of some of the things that you see coming out of the, the White House and the, the administration. 9:07 Uh, you know, words that, um, you know, te- uh, some, some technologists weren't up to speed on. And is it possible that the government is leading the way with some of this stuff? Uh, 9:23 so in 1992, when I went to work for Lockheed Martin building spacecraft for NASA, we were building digital twins, right? 9:32 We, we literally had in the computer the structural model of the spacecraft that we were building so that we could analyze it. So 9:40 a lot of this is just marketing, and, and, and I get that the marketing is how you kinda help sell this stuff, but using simulations, we've been doing that for decades. Um, 9:51 and, you know, it, it, it's great that people are doing this sort of simulation work. But then depending upon and relying upon those models depends a great deal on the fidelity of both the model and the data that's used. 10:04 And, and so my earlier point where AI doesn't, AI doesn't take away the, the need to be technically astute and to do good requirements and good testing. 10:15 It ju- it actually makes it more emphatically necessary that we do it correctly. Yeah. Yeah. Interesting. Well, I wanna use that as, uh, our segue to, uh, introduce today's guest. Today we have with us Mike Angle. 10:29 He's a longtime friend and contributor to Earp AI, formerly intelligent automation leader at PWC. Uh, he's an advisor and, uh, currently the president of the International Digital Worker Association. 10:42 Mike, good to see you, my friend. Ah, it's good to see you, gentlemen. I, I gotta tell you, I was sitting there in the green room just chomping at the bit. You had so many meaty things you were talking about. 10:52 So I hope we continue that. I love the government thing, by the way, and, uh, you know, kinda leading the way. So we'll, we'll see where that goes. 10:59 I, uh, you know, I'm surprised at how little hate mail I've gotten from members that are in the, in the [laughs] government space. 11:05 I've, I've hit on this a couple times over the last, uh, few weeks, but I'm excited about it. 11:10 I mean, it's something that I think, you know, my first reaction even with some of the regulatory stuff around AI was like, "Who, who within the government's actually gonna be equipped to, to do this?" 11:19 But, uh, you know, they keep surprising me- Mm-hmm... which is, uh, good. 11:23 So Mike, for those that, that don't know you, if you don't know Mike, you should know Mike, but if you don't know Mike, tell everybody a little of your, your background and, uh, what you've been up to. Yeah, absolutely. 11:34 Thanks for letting me do that. So I've, you know, I've been in consulting my whole life, so fact of the matter is I've had no job in the last 30 years, just been a consultant. 11:41 Uh, worked for some of the best companies, I believe, on the planet in this space. 11:45 Uh, but the last, I don't know, call it 10 or 11 years I've been working with these two guys and, and a handful of other folks really trying to drive this whole, uh, automation space, intelligent automation specifically. 11:55 And so that's l- led me to... I- I left the firm last year to do some other, to do some other cool things, including starting this digital workers association. 12:04 So I've been s- been spending most of my time doing that, thinking about the ethical and responsible ways of being able to deploy, uh, digital workers while at the same time being impactful. 12:15 So this has been really interesting as it brings together some of these discussions on government as well as- Mm... you know, regulatory bodies in terms of how are we gonna manage these things once we have them in place. 12:25 Uh, something, guys, as you recall, this is something we didn't really talk about when we were talking about RPA 12 years ago or whatever it was. We kind of skipped that part. We had no standards. 12:34 We didn't even say the same words when we were talking- Yeah... about the same thing, right? Uh, so that's what the organization is really about. It's a not-for-profit. 12:41 We'd love to have everybody join us, um, and we'd love to hear your voices, so. No, great. Great. So, um, yeah, lots to talk about. I wanna ask you about the, you know, some of the regulatory stuff. 12:52 But, yeah, but let, let, let's start with our, our theme, right? Uh, d- digital workers. 12:56 You know, at one point we talked digital workers, at one point there was, uh, this concept in automation of a, a bot for every employee. Now the- [laughs]... now the buzz term seems to be, uh, co-pilot, assistant. 13:10 W- are these all the same thing or, or what do you see as the differences and, and when you, when you think about digital, uh, you know, digital assistants and, and, and- Yeah. 13:19 Yeah, I think, you know, we have categories of digital workers, don't we, or digital agents, and I'm using those words, unfortunately I'm using those interchangeably, um, and some may not. 13:27 But in any case, uh, there, there are those things that, uh, are assistants that support us. 13:32 But I think if you, at the end of the day when you wanna define in a Merriam-Webster kind of dictionary way what, what these things are, it's AI-driven software agents, right? We've been talking about this for a while. 13:43 So to your point, Dan, what do they do? How do you define? Is it assistant or is it platform or is it actually, you know, a colleague? That's yet to be determined. 13:52 Uh, but I would say it is important that we think about this, uh, digital workers, digital agents as the extension, the next piece in [laughs] a big way of our, of what we've been talking about from digital workforce perspective for years, right? 14:04 And this is kind of the dream come true for a lot of us that we're actually able to, you know, kind of drive it out. 14:08 And think, just to, to give an example, and I'm sure you guys have a ton of them, but the simple example that really drives the point home, digital worker versus what we've been doing. 14:17 Over the past, you know, 12 years or whatever, we've been thinking about, uh, automating tasks, thinking we're automating processes, but generally we're, we're doing sub-processes and that sort of thing. 14:26 Um, and now you have to change the, the paradigm and start to think about digitizing roles, right? 14:32 So as opposed to, uh, you know, automating parts of tasks within an order-to-cash, uh, system, we're, what we're talking about now is automating an AR clerk, right? Having somebody become a colleague. 14:44 I think one, is, as usual, it's typical that we find a lot of the stuff starts in the call centers, right? 14:50 So instead of having a coach or a mentor, which are great things in a call center, now we're talking about having level one, level two agents, right? 14:58 That are, you know, providing the support to people directly in an empathetic, human-like way. So that's, I think, where we are, guys. Chris, I, I see you kind of chomping at the bit. 15:12 Well, well, it, it brings to mind a, um, I, I wrote briefly about the, the study that came out of University of Minnesota Law School back in, I think, January, which showed w- one of the effects that we're seeing here with AI is that if you are below average in ability or skill or experience, AI immediately brings you up to average capability. 15:33 Yeah. If you are at the high end of skill, experience, whatever, it can make you somewhat faster and more efficient at what you do. But we cannot, and this just makes sense from the math, we cannot replicate 15:48 higher order, high performance o- of human abilities. 15:52 And so we're, we're sort of doing the RPA thing but, again, much more emphatically, where we can, we can do, you know, 98% of the work that's required in the task, but that means that the 2% of human contribution is even more important than it ever was before. 16:07 Is that kind of consistent with what you're seeing, Mi- Michael? Yeah. 16:10 Yeah, and you know, Chris, we were talking before that we hadn't, uh, seen each other in a while, but one thing that's a constant is this discussion, right? Mm-hmm. 16:18 And, and there's this whole idea that y- that you brought to bear, and I use it all the time, and it's, uh, it's basically what about the, the, the things that aren't automated over the past 30 years, maybe they shouldn't be automated. 16:28 Yeah. It was a mantra of yours, and I agree with you wholeheartedly, so I, you're not gonna find a debate with me there. Um, but I do think that we're, we're further... 16:36 Like, when we first started talking about that, you know, this kind of the 20% is still gonna be left for humans, and I think that's shrinking. It's shrinking, but it's becoming ever more important. 16:44 I, I, I think that organizations are gonna find that the remaining, the remaining, the remainder of our humanness becomes even more differentiated, more valuable, and more important in, in what decisions are being made. 16:58 Yeah. I agree. So it's the concentrate now. Yeah. [laughs] So Mike, where, where, where are the, where are the, the opportunities, you know, within... 17:06 You know, and, and I, and I love this conversation because, you know, so much of the hype, especially over the last year, has been around, we talked about it at the head of the show, top of the show, like, you know, the individuals, you know, uh, using, you know, ChatGPT for tasks sitting on their, you know, local machines or at home if they're working from home. 17:26 Um, but if you think about this more of in, in an enterprise context, right- Yeah... 17:29 where I think organizations now are, and, you know, I spoke to one yesterday in addition to the one that Chris mentioned, where- You know, they, they see the benefits, but they, they also are concerned about the risks, and it's kind of paralyzing them, right? 17:42 Because, you know, they haven't been able to come up with an equation where they feel that the, the, the benefits outweigh, uh, the risks. So how, how, how should people- Yeah... 17:52 think of this in, in, in the context of, you know, enterprise adoption and just, you know, automation, their existing automation programs? Right. 18:01 Well, I would argue first off that whether you're serving the individual or you're serving the en- enterprise, the risk remains, right? Mm-hmm. 18:08 Um, so but from an enter- enterprise perspective, that's a lot of what we're focused on right now. We're working with a group, and I hope it's okay to plug them. 18:15 It's a group called For Humanity, and these guys are all about kind of the ethical and responsible use of AI generally, right? 18:23 And so one of the things they- they've done is they've created this, uh, audit framework, right? So it's no longer about just creating policy. 18:30 Chris, I believe policy is super important as well, and you and I can help the clients do that, but at the end of the day, it's about enforcement, right? So it's actually managing those controls and looking at it. 18:39 So I bring this up, guys, because, uh, recently we had these guys, uh, as, like, right in the middle of a, of an event, uh, where we had, uh, NVIDIA, and Meta, and Google, and all these guys talking about the, the, the future and getting people excited. 18:55 But what I knew was that there was a couple CFOs sitting in the audience thinking, "Oh my, man, this is, this is scary," right? 19:01 [laughs] Um, so we brought the, uh, we brought the For Humanity guys in next to say, "Hey, listen. It is scary. There is a lot of risk. 19:07 But here's how you can mitigate it, by having the right program, by having the right policies and so forth." So I think that's, that's important. 19:13 Um, and I like to say it over and over again, and I hope you guys do, too, right? Because it's, there is a way to, uh, to mitigate the risk. Not eliminate, but mitigate the risk and help people move forward. 19:25 Yeah, I love what you say about, you know, it's kinda, it's kinda inevitable, right? I mean, whether, whether you're doing it at an enterprise level or not, people are doing it, right? 19:32 We've, we've, we've seen these- Yeah... examples where, you know, there's proprietary data being shared out to [laughs] a large language model because, you know, somebody, somebody's doing it. 19:41 And I mentioned, you know, uh, before you joined about just, you know, you know, how, how AWS I think really adopt- got adopted, which was, you know, rogue or shadow IT groups- [laughs]... 19:51 you know, setting this stuff up. So, um, is, is that, do, do you see that, do you agree that that's happening now at, at- Yeah... scale? And, and then how do you, how do you govern that as a, as an enterprise? Yeah. 20:05 Well, I think there are a couple of things that, that are important here. I'd start with the, the easy one, which is the individuals. It, it, this has always been the case, right? 20:11 It, it, me, you, the 27-year-old person that's the average age for my per- previous firm, they, they use it at home, right? They depend on it, they rely on it. 20:21 When they're gonna write a letter to their girlfriend or whatever it is, you're gonna be using ChatGPT to do it, right? 20:26 So your expectation is the same when you arrive at work, that you're gonna have these tools to support you, right? 20:32 And so if we don't give it to them, if we don't put it in a way that's governable, that's manageable, monitorable, that sort of thing, they're gonna do it anyway, right? 20:40 Um, and I, I would argue that if you don't, you're stupid, right? [laughs] Yeah. Yeah. Right? I mean, how many times have you wrote a docu- created a document and struggled with the, uh, with the table of contents? 20:51 So [laughs] why bother? But from an enter- enterprise perspective, then I think it's a little harder, um, you know, assuming you're in, in a larger enterprise. 21:00 And then, uh, so I, and I'm not really seeing that run rampant. Chris, you agree? You've, you've seen, you've seen about the same, right? 21:10 Yeah, and I'm glad, I'm glad, Michael, that you mentioned policy, um, as opposed to rules. 21:15 Uh, I, I, I, one of the Stoicists, I think it was Marcus Aurelius, I saw a quote from him, and I'm gonna misquote him, but it was that in effect, the, the more laws a country has, the more lawless it becomes. Hmm. 21:26 You kind of see this, like, with our tax code. There's 45,000 pages, which means that there's effectively no tax code. 21:33 And, and I think that many organizations, particularly with RPA, because RPA is all rules-based, we operate on this notion of you need more rules as opposed to following principles. Mm-hmm. 21:45 And, and I, and I think what we're gonna find with AI to make it effective, we need operating principles. Here's, here's the right way of doing things. Here's what right looks like, but not rules-based. 21:56 Because by its very nature, the more rules we pile on, the more constrained we make ourselves, and the harder it is to actually get anything done and remain compliant. 22:05 So I, I think it's gonna be the ethics behind AI are actually gonna be somewhat different than what we've seen in the past. 22:11 And Chris, I, I wonder, you know, as, as you were saying that, I'm thinking about, you know, infrastructure versus deployment, right? 22:18 So, um, you know, put a lot of rules around what can come into your environment, and from, and I'll use that as infrastructure, um, versus how you deploy. 22:26 And I think, you know, what I'm discussing really is more about deploy, right? You know, I cannot, you know, ethically try to use m- my app Digital Worker to, to, uh, encourage kids to smoke, right? [laughs] Yeah. 22:39 I mean, things like that, that are actually important. Those are, to me, those are much more monitorable and auditable. Um- Yeah... so. And we'll follow them. I mean, 22:49 if, I'll never forget going to law school, and everyone, the, the hardest class in law school is evidence, and that's because there's the rule of evidence, and then there's 73 exceptions to the rule. 23:00 And it's like- Yeah... any rule with 73 exceptions isn't a rule. It's barely a suggestion. Yeah. [laughs] Right? 23:06 And, and, and so, so you, you, you reach a level of ridiculousness, I think, with some rules that make, effectively mean that it's, it's unenforceable. So- Yeah... 23:16 having a principle instead, I think, is what's gonna be key to effectively adopting this stuff. Right. 23:21 No, I, I understand, you know, particularly from a policy perspective, controls perspective, you know, that can get onerous, and nobody's gonna follow it if it's some, whatever you said, 70 pages, right? 23:30 But what if we were to, uh, actually automate the automation, right, in terms of- Mm-hmm... 23:34 what's, what's, you know, and we, we've done this in a lab where we're looking at automated controls that are adaptable- Right, based on whatever policies and, and procedures we put in place. 23:44 They'll adapt and, and they'll go back through and, and make sure that those controls are in place and that they're being followed. It... What do you think? I mean, is that, is that... 23:52 I mean, we, we were talking about digital twin be- you guys were talking about digital twin before. So if I have a digital twin in my process, for example, couldn't I, you know, con- consider how these changes w- uh, 24:03 would impact- The-... the regulation and so forth? That sort of simulation, I think, is gonna be absolutely critical because 24:14 p- part of what I think you find with the scale of, of certainly LLMs like ChatGPT is by the time you realize something's gone wrong, it's far too late to do anything about it. 24:23 Um, and so the, the need for that sort of simulation to get in front of it, I think, is gonna be critical. Yeah. I agree. And we, we're playing enough Whac-A-Mole, right, [laughs] in technology. 24:34 We don't need to add to it. But fair enough. You know, Mike, you, you, you sat in your previous role in a, I think, I think an inter- interesting intersection of, you know, managing innovation versus risk. Um- Yeah... 24:47 you know, I, I'm curious as to, you know, y- y- when you see, when you see the government chiming in, like we talked about before, you know, do, do you, do you feel that they get that balancing act? 25:03 Yeah, what we've seen so far, I believe so. Right? 25:06 So I mean, if we look at the stuff in the EU, particularly because it's, uh, you know, it's better defined, it's been thought out, there's a lot more people that have thrown darts at it, you know, had time to throw darts at it. 25:14 Um, and, and I, I do think it is well thought out, right? That they didn't do it... It's not people that don't understand this technology that are writing this, that are creating these policies and regulation. 25:25 Um, of course, we don't have regulation here yet, just yet, is w- the way I look at it. Uh, so I think they're, they're being supported by us. 25:32 I think we should continue to support the government in their efforts because ultimately everybody's going to have to adhere to that at the very minimum. 25:40 Yeah, I just, um, you know, like I said, some of, some of this, this newer, some of these newer technologies and ideas and things that, that- Yeah... 25:48 are coming out, you know, seem, you know, ma- make me, make me feel optimistic. But, you know, you, you contrast that to some of the congressional hearings over the years of- Yeah... 25:58 you know, uh, congressmen interviewing Mark Zuckerberg or even the, the one that's, that's burned into my head is when they, they, uh, interrogated the CEO of Ticketmaster after [laughs] like the- Yeah... 26:10 Taylor Swift meltdown. Uh, and it's like, you know, these guys have, have, have no i- no idea even what questions to ask, never mind- Right... you know, how to en- en- enforce policy, right? 26:22 But may- but maybe, maybe that's just- Wrong guys... politics and there's maybe there's people behind them that do. [laughs] Hopefully. Yeah. I, I would, I would argue they are. 26:28 Those, those folks that you're talking about aren't the ones creating the policy, right? 26:32 Um, and, and there's kind of two sides to this, and Chris, you know this way better than I do, but there's the kind of the business of government, right, which is a bit easier to control. 26:40 And then there's the business of, uh, taking care of humanity and, uh, managing your defense and all of those things that are completely different than, than just the business of government. 26:53 Would you agree with that, Chris? I mean, you know, it's this kind of- Yeah... the things that happen day to day and, but there's also the protective side, the, uh, defense and everything else. 27:02 There's a tremendous amount of theater that goes on and, and behind the screen, you know, when you pull away, wh- when Toto pulls away the curtain and you see, you know, the, the Wizard of Oz actually has a control booth there- [laughs]... 27:15 and you kinda know what he's doing, that's very much, I think, uh, a, a adequate representation of what's going on. Yeah. Yeah. 27:23 Mike, I'm curious, you know, I'll, I'll stop beating up on government and, and, uh- [laughs]... but I'd like to, uh- I'm sorry I didn't go along with you there. [laughs]... 27:30 some, some other, uh, you know, industries, right? So, you know, a lot, you know, this, this technology is inherently industry agnostic and, uh, we tend to talk a lot about it in, in, in generalizations. 27:40 But I'm curious, a- a- are there any, any particular industries that you're most excited about, where you see the, the biggest or quickest, uh, impact, uh, as it relates to automation plus digital workers? Yeah. 27:52 I think anybody that cares about their customer, right? 27:55 And maybe that's just because we've been focusing on it a lot with our, with our teams recently, but this whole idea of customer satisfaction is becoming more important and the, you know, kind of the, the consumer experience. 28:05 And w- and not just because of competition, but because you wanna please them, right? Um, so this, this sense of that... 28:12 Something we could, we, we talked about this, Chris, for years too, with this idea of personalization as somebody in Peppers. I remember this back in '99. 28:21 Um, then it became important that we knew who you were when you called in and so forth. 28:25 But now it's important that I know everything about you, and no matter what channel you kind of contact me through, I still have this string attached to our discussion. I know you, I know your likes and dislikes. 28:35 Um, and so I can, I can ta- tailor my responses to you, right? Maybe it's even in from a language perspective, maybe I give you a Southern drawl if I know that's what you like. 28:45 Um, and, uh, so I think this, this is gonna be a big deal, Dan, across all in, uh, all industries, whether it's insurance, it's retail, uh, call centers, the whole thing. 28:55 To me, that's really exciting s- sense of hyper-personalization and the impact it has on the business. I agree. 29:01 I mean, I, I think, um, you know, as a, as a consumer, having lived through, you know, several years of communicating almost entirely with organizations through call centers and things, right, in the middle of the pandemic, what have you- Yeah... 29:17 um, and in the post-pandemic world, um, you know, I'm, I'm a big believer, I've said it before, I'm a big believer that a- a- and as things become more commoditized and more AI-centric, you know, I think experience becomes the differentiator. 29:31 I think that's what, that's what sets brands apart, uh, is, is a, as a, as a differentiator. 29:36 At the end of the day, um, you know, for the most part, you know, if you need a checking account It doesn't really matter what bank you use, as long as they're not on the, uh, you know, Silicon Valley Bank or one of those [laughs] Yeah. 29:50 [laughs] Um, yeah, that, that, that actually destroyed my argument. I think we had a conversation about that, and a week later, that bank failed. But o- other than that, um- [laughs] It happens... 29:58 you know, a lot of these, uh, a lot of these organizations kinda, you know, uh, deliver something that's becoming more and more commoditized, right? So at the end of the day- Right... 30:07 the, the experience is the, the differentiator. What I'm not sure about is how do you... So that, so there's that. 30:16 But then you also have the, the AI thing happening, which, you know, you could argue actually eliminates a lot of differentiation. [laughs] Mm. So how do you find that right inter- uh, intersection? 30:29 Well, I think that's what Chris was re- re- relating to earlier. It's, uh, there's a lot that has to go on on the front end. 30:35 Um, I'm gonna give you an example, one that I really like as it relates to this, something we're working on. It's a... The company is a startup, high-tech startup, and they do concierge work. 30:44 This is concierge for high net worth, di- diplomats or, you know, rock stars, that kind of thing, and they've done this with, through a combination of technology and human beings forever, but now they can't, they just can't scale to people fast enough or, or cheaply enough to, to make this re- a reasonable thing to do. 31:02 Well, they've created this kind of, uh, capability with, particularly reliant on gen AI, where the people don't know whether they're talking to or, or communicating with a person or communicating with, with a bot in this case or an agent. 31:16 Um, we built in e- empathy to it, right? So if somebody says, "Hey, I missed my flight," it's not just, "What flight do you want?" It's, it's, you know, "I'm sorry to hear that. 31:25 You know, let's look through some options." Right? And then, "Oh, by the way, I see you're going to San Francisco, and when you go to San Francisco, a lot of times you like to go to a ball game. 31:33 Should I get you tickets?" Yeah. So it, it's just a whole different, uh, e- experience. 31:40 But the reason it's like that is not just because the tool exists, it's because the team spent an awful lot of time and an awful lot of experience that they've had over the past 10 years of being in the concierge business to be able to train this capability, right? 31:52 So that's the differentiator. Yeah, I agree. 31:55 I think we're, you know, beyond the conversation we used to have, especially I remember, like, some early on conversations around, uh, you know, IT service desk where people didn't want a chat bot or then an AI to, you know, speak to a customer. 32:10 Now, I think most customers prefer it because- Yeah... it's a transaction that is easier for a machine to do than, than, than a person. But, um, you're right. 32:20 I think that, you know, back to, you know, the, the experience being the differentiator, I think that empathy becomes the, the differentiator- Right... um, and, uh, makes a, make, makes, makes the difference. 32:32 Uh, I, I'm curious though, you know, as you think about that, the impact, and, you know, you can't talk about this without talking about the, the, the workforce, right? 32:41 I mean, h- how do you think this plays out in the, in the workforce? Because, you know, with, with automation, it was largely, hey, you're, you know, you, you... 32:51 It's, it's people doing jobs that they don't like doing, and so they're gonna go to a, a better job. 32:56 And, and I think we said that early on as it relates to, to AI as well, but it, you know, the numbers are starting to show that that may not be the case. Yeah. 33:05 Well, look, I think it, it's another thing that goes to what Chris was talking about in terms of the concentration of the good work, right, the fun stuff to do. 33:11 Let's say, you know, in, in my old business we had a lot of accountants, and we had a lot of accountants that did this. They stared at one thing and compared it with another thing, and these are professionals, right? 33:21 And so what did we do? We used AI to remove the stare and compare piece, and what that left these professionals, these highly trained accountants to do was accounting, right? 33:31 To find out wh- wh- why this anomaly occurred and, and kind of the fun work, or maybe it was the 2% that was left, it was more than that. But in any case, uh, it, it changed their lives, right? 33:40 It made it just much more interesting. They could bring on more, more work, um, and just really concentrate on the fun stuff. 33:45 So I think that to the extent that we can do that and, and just basically make these, uh, digital workers, digital agents into colleagues and, you know, as opposed to they're, they're not taking your job, they're taking the crap part of your job, uh, but they're doing it wholly, entirely. 34:02 Well, um, Matt, I really resonate. The, those that view AI as a way of, um, supplanting people as opposed to supporting them are, are getting this really wrong, I think. 34:15 Um, um, w- we are gonna continue to automate more and more, you know, more, uh, decimal places or more, more significant digits after the decimal place. 34:24 But that means that that ever-shrinking proportion of human contribution becomes ever more valuable. Right. 34:31 And so you're getting tremendous scalability and efficiency from, from these automation technologies, but you're actually emphasizing the importance of the human contribution, uh, in, in that tiny landing area at the end, you know? 34:45 Well, you know, be- before pandemic, [laughs] uh, there was this, the talk, we were talking a lot about kind of micro gigs. The sense- Yeah, gig economy... exactly what you're talking about. Yeah. Yeah. 34:54 So yeah, so from taking the gig economy down to even, like, micro gig, because I am the best at, uh, moving a, an insurance claim through this particular part of the, uh, of the process, right? I'm the best at it. 35:05 So every time it comes up, you send it to me and pay me a nickel. Uh, and, and so I think that what you're talking about is we, we could even move faster to that sort of thing- Yeah... 35:14 if, if we're to, you know, take this seriously. I, you know, and I have an example. 35:19 So I'm, I'm advising a, a healthcare startup that is working on a chronic condition in the human population that's exploding all over the place, and the challenge that you have is the healthcare specialists that, that specialize in this particular area, there's, like, 1,000 on the planet versus potentially a billion people that are gonna have this, this problem going forward. 35:42 And so the only way you can possibly manage- Scalability issues like that is with, with this sort of automation. Right. 35:50 So, um, yeah, I, I think we're gonna be able to address use cases that weren't addressable previously, um, and there's gonna be a big, uh, criticality to doing that effectively. Yeah. 36:01 Well, we ran into this in the RPA craze in the beginning as well. We, we actually made some false claims [laughs] intentionally, right? No way. 36:09 You're gonna be able to reduce your workforce by X, Y, Z, whatever, right, which was, you know, didn't prove out. Uh, but I think what it did do, over time that is, is it led us to different, different ways of measuring, 36:22 right? So when we're thinking about... All we were talking about is, is revenue and profit, right? Stuff that you can see, uh, in the financials. But we were missing really the KPIs that matter. 36:33 You know, taking this back to kind of customer satisfaction matters. It matters to both the, you know, the revenue and profit. Uh, so 36:42 and it's just change of focus I think is going to be important here so we can, a- a- again, coming back to kind of how the people are related to it. 36:49 If I'm responsible for the actual outcome, but not all the work that happens underneath, I think I'm pretty happy about that. 36:57 Mike, one, one last question, you know, o- other, other than the workforce ramifications and the, you know, the risk and compliance things that we've talked about, what are, what are some of the other challenges that, that you see organizations facing in implementing digital workers, and, and what's, what's your advice? 37:17 Well, there's a couple of things. Look, right now we're tracking... I, I just looked. It's 82 different, uh, companies that are in this space. 37:24 They're either providing, you know, the, the bots themselves, right, or the agents themselves, or they're re- they're creating platforms to be able to build this out. 37:32 And the reason I bring that up is because the biggest things right now in terms of barriers to entry are around cost, right, complexity, 37:41 and, and, and kind of speed, and then the last thing, just what we were just talking about, fear. Right? Uh, so I think, you know, we have to address each one of those things in order. 37:50 You guys, before I jumped on, you were talking about some of the things that are happening with, with the chipsets to be able to increase speed, with companies like Nvidia, uh, Groq, and all these guys that are addressing this stuff. 38:01 It's not just about speed, it's about cost, right? I mean, right now, if you really want to scale something up to be equivalent of thousands of p- of workers, i- it's gonna be cost-prohibitive. 38:11 You're not gonna get what you need out of it. Uh, from a complexity perspective, uh, it's all in the integration. It's in, in the data, right? 38:18 Uh, so many of the companies, as soon as you get out of, uh, Fortune 500, I mean, a lot of them are in the Fortune 500, people don't have a single instance of SAP and everything is there. 38:27 I'm sorry, of an enterprise, uh, s- platform and everything is there. Data's everywhere, uh, which makes it mu- more complex and takes more time. So I think that's, that's really what it is. It's, it's not unusual. 38:38 Cost, complexity, speed, and fear. I think those are the biggest things. And, and Michael, I'd, I'd just like to chime in because, um, 38:47 th- the reason I used the RPA, 'cause everyone's like, "Well, you know, we're, we're doing this new thing called automation." You know, we've been automating for 50 years. That's what IT is. Yeah. 38:56 Um, and if something, if, if by 2015 something hadn't already been automated, there's probably a reason it wasn't already automated. Right. 39:05 And we, we need the same terminology that I always use, that cost, complexity, compliance. 39:10 If something has not been automated already, it's because of cost, complexity, and compliance, and we shouldn't under- we shouldn't underestimate those barriers to further automation. 39:22 So AI may be able to do it, but those barriers are probably still there. Yeah. Well look, there's, there's more to come on this, right? 39:30 I mean, I think if you s- start to think about what's the future that's now, you know, now we're starting to talk about the, the, the agents that collaborate with each other and learn from each other, Chris, right? 39:39 Mm-hmm. And so, I mean, and kind of closing that gap between them I think is gonna be a big deal. Uh, and then the other part of it is, of course, you know, looking at more autonomy across the enterprise. 39:48 So now I'm not reliant, you know, as a human being on somebody else's mistake, 'cause that mistake didn't occur, right? 39:54 So I think as, as things become more and more au- autonomous across front, middle, and back office, some of that it will start to, to at least close, the gap will start to close a little bit. 40:04 Mike, great to see you, buddy. Thanks so much for joining us, and, uh, hope we can have you back, uh, maybe in a few weeks, see what's going on- All right... in your world. Terrific. Great to be with you guys. 40:14 Thank you very much. Thanks, Chris. Have a great week, buddy. Yep. Thanks for joining us, everybody. We're back Wednesday at noon Eastern, 9:00 Pacific. 40:23 In the meantime, go to erpaai.com to join our free membership or schedule a complimentary Art of the Possible coaching session or industry use case briefing. I'm Dan Goodstein. 40:34 Thanks for joining us, and have a great week, everybody. Bye now. 40:36 [outro music]