Transcript 0:00 [upbeat music] If you don't know your numbers, you don't know your business. 0:12 I learned this the hard way with my first company, so I created NetSuite, the world's number one cloud business system. 0:20 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. 0:32 Schedule your free product tour right now. Hello everyone, and welcome to the Digital Leader show. 0:49 We're back to being live every week, Wednesdays at noon Eastern, 9:00 Pacific, talking key issues and news at the intersection of business and technology with a special focus on automation, AI, and digital transformation in hopes of making you a better digital leader. 1:05 Today we're discussing how should CIOs measure automation and AI success with industry leader Olivier Gomez, plus headlines for digital leaders. 1:14 Thank you to our sponsor, Oracle NetSuite, a unified cloud business management platform that allows organizations to operate more effectively by automating core processes and providing real-time visibility into operational and financial performance. 1:28 ERPA AI members and followers can receive special pricing and resources. Go to erpaai.com/netsuite for more information. I'm Dan Goodstein. I head up the Institute for RPA and AI and 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 worldwide all around automation, enterprise AI, and digital transformation. 1:54 My co-host today is Chris Sirdak, Senior Advisor to ERPA AI, and formerly advisor for the White House Chief Transformation Office, Partner for Digital Transformation at Gartner, and Automation AI Practice Lead at EY. 2:07 Chris, good afternoon, my friend. And how are you, sir? I'm not too bad. It's, uh, warming up in New York and starting to feel like, uh, you know, there's no spring anymore. 2:18 So it went from winter to summer, but, uh, you know, no complaining. I'm happy. Who, who knew it would be a, a good time to be an umbrella salesman in Dubai, right? There you go. Right. Right. I saw that. Right? Crazy. 2:30 Crazy. Seeding the clouds apparently works. So lots going on. 2:34 Yeah, the first thing that kind of caught my eye, we alluded to this over the last couple weeks, but, uh, it, uh, according to a couple different sources, uh, that United health hack is gonna cost over a billion dollars. 2:49 Just another example of the, you know, j- just the financial side of, of some of these, uh, cyber concerns we've been, we've been talking about. Well, and I think when you add to that the, um... 3:04 I don't know if anybody, anybody else noticed, but the increasingly aggressive, um, enforcement activities by the FTC, which arguably is, is, you know, necessary and potentially, you know, long in coming, it could get really expensive. 3:19 And the other thing I always think of, uh, in these sorts of situations is a lot of my, my colleagues in cyber are like, um, you know, by the time you re- you recognize and are willing to admit that you have attack, it's 20 times worse than what you're willing to acknowledge or even know about. 3:34 So, you know, more to come on this, I, I'm willing to bet. Yeah, we're, we're, we're tracking a lot of, uh, different kind of startups and, and solutions, uh, you know, obviously AI-centric around cybersecurity. 3:49 So, so more, more on that in, in, in coming weeks. But, uh, yeah, it's, uh, you know, it's one of those, those AI arms race, to use that overused term, right? 4:00 Of, you know, who, who, who's, who's getting better and faster? Uh, is it, is it the, the attackers or the people protecting against these attacks? But- Well, and, and, and 4:09 I, I think we have to recognize the Venn diagram of what's going on. We're sweeping the entire internet to collect data so we can train our generative AI with trillions and trillions of data points. 4:20 And in that great vacuuming up of all that information, you're likely pulling in Lord only knows how many, uh, potential backdoors. And there's already been a lot of evidence of this happening. Um, so 4:33 assume that it's happening and th- and then the question is then what? What are you gonna do about it? 4:39 And I know that you and I have had this discussion many times on this call with, with many of our guests or the show with many of our guests. 4:45 It's part of, part of being good at governance of these systems is asking yourself and being able to answer, "What do I do if this system goes down for three days? What do I do if, if, um, 4:57 uh, I can't, I can't restore it," right? 5:00 These, these are sort of those disaster recovery and business continuity questions, which were pretty common about 20 years ago, and I, I feel like in many ways organizations have forgotten to do that. Yeah. Yeah. 5:12 So interesting disruption, more, more disruption in the, in the real estate space. 5:16 There's a, according to the Wall Street Journal, there's a, a lawsuit kind of pending, uh, nationwide kind of legal battle around some software that property management companies are using, which, uh, you know, the accusation is that they're illegally kind of rent setting, uh, between, between landlords. 5:36 Uh, so, uh, and that these, these algorithms that are, are being used in this software are artificially pushing rents up. 5:44 It's an, it's an interesting, you know, other side of AI, I don't know, that we've, we've talked about beyond just the, the privacy and the security and, and some of those, uh, issues. 5:55 But, you know, you know, may- maybe there's some, some changing business models or, you know, um, uh, legalities that need to be kind of revisited based on, you know, some of the capabilities that, that, that this technology can, can bring, you know, good and bad. 6:12 Are you suggesting that the reverse auction software that I've been using to get the best price on things for the last 15 years might be used to hurt me? Ooh, I- Yeah. Right. Right... 6:22 Talk about use case dependency, right? Yeah, exactly. Exactly. 6:27 And, uh, you know, in your neck of the woods, uh, you know, this, this, uh, increased minimum wage thing around fast food workers is, you know, obviously, uh, made, made a lot of news and, you know, the, the, the, the, the Politico people, uh, back and forth on whether, whether this is good or bad. 6:46 But I'll tell you one of the, one of the side effects I thought was interesting is, you know, I, I think this drives more and more technology adoption. 6:54 I don't, I don't know if that's in the equation of some of the politicians that, that think about this. 6:58 There's this, uh, uh, this article in Business Insider where they interviewed, uh, one of the larger Burger King franchisees in, in California, and he basically said, "That's fine. 7:09 We're just gonna roll out technology, more kiosks." You know, we talked about McDonald's AI ordering. You're gonna... 7:15 I, I think you're gonna see more and more, you know, of this, especially in places where, you know, at least the, the perception by the business owners is that the, the cost of labor is just getting too high. It's... 7:27 I, I, I can assure you that in my very own neighborhood, this is happening in real time as we speak. 7:33 Not only is a value meal, not that I eat fast food very frequently, but I had recently a meal of 20 bucks for a, a value meal. I don't know when that happened, but it happened. 7:44 And yeah, uh, many of the restaurants locally are closing, and others are moving full on to full automation. 7:51 It's actually, um, to the point where if you go into one of these restaurants, there, there's rarely more than one, uh, person working there- Yeah... which itself is pretty amazing. So, 8:02 um, you know, careful what you wish for, and unintended consequences eventually show themselves. So, um, you know, I wanna kind of get to our, our guest. 8:12 You know, a lot of, a lot of the conversations we have, Chris, as you, as you know you're a part of, is around, you know, how do we, how do we measure success in these things, right? 8:23 One, one use case may be, you know, "Hey, I'm trying to save costs," you know, labor costs. 8:28 That's, that's almost an easy, uh, use case, but there's plenty of other use cases out there that, um, you know, the, the, the, the goal line kind of continually moves as far as what the, what the desired outcome is. 8:40 And so, um, you know, our, our expert today is, uh, O- Olivier Gomez, or as most people know him in the industry, OG. 8:48 He's a LinkedIn top voice and expert in automation and AI, Forbes Technology Council member, an author, advisor, and speaker, having delivered over $100 million of P&L impact for his previous roles at DXC, HP, and a slew of other clients. 9:02 OG, welcome to the show. Thank you, guys. Uh, and thank you for having me, and, um, I would say hello from France. I'm located in France, so a bit of, uh, you know, delay here. 9:15 Uh, hopefully, you know, uh, uh, technology will work fine. So, hello from Europe. 9:21 [laughs] Good to see you as always, and, uh, you know, you and I met, I think, many years ago when you were, uh, in your role either at DXC or, or, or HP and you were running a pretty big, uh, automation program. 9:37 I think it may have been the biggest automation program that I was aware of at the, at the time when, you know, many of our members were, you know, really kind of pre-scale mode. 9:47 I, I'm curious just, you know, your, your sense of where things are now as t- uh, you know, as compared to the way it was maybe, maybe four or five years ago from a, from an industry adoption standpoint. Yeah. Yeah. 10:01 So back in the days, uh, that's, that's true. Uh, I was actually running a pretty large footprint of, uh, we had up to 5,000 bots running, so it was pretty large. Uh, 10:12 uh, you know, even, uh, when you look at, uh, the largest footprint today. I would say, uh, when we look at the overall, uh, intelligent automation footprint, right? 10:24 So if we combine, uh, automation and AI together, I would say that the, uh, industry is doing pretty well, right? So we have now a, a number of technology that are considered, uh, a commodity. 10:38 Uh, I would include, uh, RPA in that list, and that's a good thing. 10:42 It means that, um, people have adopted those technology, people are scaling those technology, and yes, there's still some issues left and right, and we can discuss about it, but, uh, on average, the intelligent automation footprint, the amount of technology being utilized by technology, uh, by, by companies, sorry, across all verticals, all industries, is growing. 11:06 And people are relying more and more on those technology to deliver outcome, uh, to serve their client, and, you know, sometimes to improve their bottom line. What's your, what's your, um, 11:21 what's your sense of this, this labor issue, right? I mean, when we, we first started kind of... You know, I, I... You know, for me, for me it was evangelizing a- automation. 11:30 You know, for, for you and many others, it was kind of selling automation. Um, you know, it was more about, you know, efficiency and, um, not so much about, you know, displaced labor and things. 11:42 Uh, you know, to me, at least with the most recent AI wave, I think that, that, that labor issue becomes, you know, i- i- in- inevitable. Do, do you agree? And, and, you know, what do you see... 11:54 Yeah, how do you see that kind of balance shifting? Or what are you, what are you advising your clients as far as, you know, what to do with this, you know, newfound savings or efficiency? Yeah. 12:06 So, so obviously the, the typical answer here is, uh, you know, it depends on the environment, but I would say overall, uh, a couple of, uh, good thing here. 12:15 I think there's a lot, a lot less fear In the system right now, right? 12:20 People understand, people acknowledge, and they accept, uh, that this technology is going to disrupt, and they understand the, the pro and the cons of this technology, right? 12:32 In some areas, it's actually very, very good news, uh, because, uh, let's say shared services, call centers, ten, fifteen, twenty, thirty percent attrition rate, it's good news to start using technology, uh, because this was certainly not something designed, uh, for human to do. 12:52 Uh, in other area, it might be a bit more negative where a, a number of, uh, you know, jobs might be, uh, put at risk. 12:59 Uh, what we know from history is that typically when there is, uh, a disruption that is that large, uh, you know, there's a, a little bit of, uh, um, bad news at the beginning, but then things falls into place. 13:13 Uh, new roles will be created and, uh, human that are adopting the technology, you know, we, we say human plus AI is always going to be greater than human without AI, right? 13:24 So if the adoption is happening, if the reskilling is happening, uh, I'm not too concerned. 13:30 You know, most of the people will be able to find a place, uh, in the future, uh, and a number of, you know, pretty awful jobs, uh, going away. Awful because they are super boring to do and hence the attrition rate. 13:44 You know, I think overall, uh, again, I, I, I think all of that is good news. Some people will be, um, uh, you know, kind of collateral damage, uh, but it's not as bad as what people think, I believe. Chris, over to you. 14:02 Yeah, I, I, I tend to agree, and then given that we want to have a discussion about performance metrics and so forth, I wonder if, um, you could touch upon the principle that I've used a lot in automation, and that is the, the difference between effective-- efficiency and effectiveness. 14:18 Where efficiency is sort of driving the same outcome but at lower cost, and effectiveness is either giving different outcomes or more or, or better outcomes even with the same resources or slightly different resources. 14:31 Is that among the things that you find organizations are starting to recognize? Yeah. I would say, um, organization are s- are still looking for both, right? 14:41 So if you look at, um, expectation, typical expectation, uh, obviously there's some bottom line conversation, uh, people are looking at, you know, trying to optimize a process. 14:54 But there's a lot more happening around how can I help my top line? How can I help my revenue? How can I increase, um, the satisfaction both internally with employee and externally with, with, with customers, right? 15:10 So I would say that the augmentation that technology is bringing to the table is allowing to, uh, do obviously some, uh, savings left and right, but also to do things better. 15:26 You know, the quality is increasing, the speed is increasing. Um, there's some, you know, pretty interesting predictive capability, trying to predict future, trying to foresee what will happen tomorrow. 15:38 All of those are, to me, uh, augmentation that, uh, are coming, uh, above and beyond, you know, simple bottom line, simple savings expectation. So it's a mix of all of that. 15:51 And, uh, for sure what we see now is that outcome, uh, and impact are expected, right? So we are m- you know, moved, we have moved away from the science experiment or the, you know, multiple POCs. 16:08 A- and, and I would say even GenAI is reaching a point where, uh, there's some expectation for, for, uh, tangible results, right? 16:17 And those results are, as, as I said, spanning, uh, from typical savings expectation all the way up to the, you know, predicting the future type of expectation. It's, uh, so, so you're giving me an idea for an article. 16:31 It's Schrödinger's automation. If you're, if you're anticipating efficiency and cost reduction, that's what you'll get. If you're anticipating improvements in, in output or different value, then that's what you get. 16:43 It, it really is, uh, an observational bias. Exactly. Exactly. All the tools are there. The ecosystem is, uh, growing on a daily basis, and th- there's tens of thousands of tools, tens of thousands of algorithm. 16:59 Uh, so I'm totally with you, right? 17:01 And I'm, I'm always pushing for a, a, a tech-agnostic approach to the conversations that we start from the end, we understand what is expected, we understand what impact is expected, and then we reverse the engineer towards solutioning, right? 17:17 Th- this is, this is really important, uh, to do. And one thing I can add is, uh, there's still quite-- especially with AI, right? 17:26 Less with automation, but with AI, there's still, uh, a large amount of leaders that are a bit lost, right? Uh, they, they know they need AI, but they don't really know where, they don't really know how, right? 17:40 So we are-- Th- there's a large amount of, uh, uh, demand around, you know, help me assess my environment so I underst-- I know I need AI, but I'm not sure where. 17:50 You know, is it in my front-end activity to boost my marketing and sales to, uh, uh, improve my, uh, uh, customer support? 18:00 Is it in my back end so that I can, uh, optimize and maximize my HR, my IT, my, my finance department? 18:09 Is it in my product and, and, and services, so I can augment what I'm selling, so I can basically lead, uh, the market? 18:17 Uh, people are a little bit lost because there's so many use cases out there that people don't really know what to inject in their environment. 18:28 I, I mean, for me, that leads to an entirely different question, which I'm sure you work with clients to figure out, and that is if you're trying to answer that question, how do you, how do you find expertise that is trustworthy and that you can go to for that sort of information? 18:43 And one, and one comes to mind is I've seen job postings where people are looking for AI leadership, and they say, you know, "Must have 10 years of experience with ChatGPT," which is like a seven-year-old technology, which kind of tells you out, out of the gate that you're, you're probably fishing in the wrong pond, so to speak. 19:00 So, so if you're an executive trying to get answers to this sort of stuff, how do you find, um, trustworthy expertise? Yeah. 19:07 So I'm, I'm a little bit biased here because I'm, I'm doing some of those, uh, you know, uh, assessment, uh, myself, but I would say, um, some, some companies are relying on, on internal resources. 19:21 So they try to, uh, upskill, uh, some of their internal resources that have, um, uh, I w- I would say, um, experience in, um, in improving environment, and they try to augment them with AI knowledge. 19:38 Uh, but very often, uh, companies are looking at, uh, external help. Why? 19:45 Because what's interesting when you are, uh, looking for external help is to, to answer y- your question, to find people who, who have done it before elsewhere, to find people who have, uh, who are more than just consultant, to find people who have this, uh, exposure to real project and real outcome, right? 20:06 And I would say this is probably the, the, the key requirement here. 20:10 Um, what you don't want is, um, you know, a bunch of PowerPoint that will, uh, maybe state the obvious, but what you want is people who have, uh, a lot of scars, who have suffered, uh, you know, uh, uh, implementing project and failing. 20:29 So I would say the, the experience is absolutely critical and, and to be honest, in the AI space, um, there's not a lot of people with, uh, you know, many years of experience because of the, of the technology, right? 20:45 It's, it's so new. I think ChatGPT now is what? 18 months old, roughly. 20:51 So it, it's a little bit difficult to find trusted people that, you know, can help you implement something, uh, and have enough experience given, you know, how young this technology or some of those technologies are. 21:05 So it, it is a challenge, um, um, for sure. OG, how about from the, from the CIO standpoint? 21:12 You know, I think we talked on the show a couple, couple weeks ago with someone who, I forget who it was, who said, you know, so- something to the effect of, you know, CIOs have an unfair job because, you know, they're, they're typically not as responsible for innovation, but they're kind of the ones that have their, uh, you know, their butts on the hot seat if, if something goes, goes wrong. 21:34 Um, uh, what, what's the kind of biggest, you know, as you think about, you know, automation but, but especially AI, what, what, what's the biggest challenge, you know, facing a CIO right now? Yeah. 21:45 That's a, that's a, that's a very good question. So I would say, first of all, uh, they, they are back, they are back in, in the middle of the conversation, right? 21:54 So a while back when we were talking about RPA, there was a lot of, uh, uh, conversation, uh, pushing this idea that you don't really need IT, you can do, uh, automation without IT. 22:06 I would say with AI, this is, this is all gone, right? The, the CIO is in the middle of, uh, the, the, the plan, right? 22:16 And, and while there's, there's a need for legal support, there's a need for HR support to handle a, a number of other AI aspect, the actual solutioning and implementation, uh, sits with the CIO. 22:28 And I would say, uh, there's a, a, a bunch of challenges, uh, in front of them, right? So first one is, uh, everything related to, um, finding the right people. 22:39 You know, back to the previous question, making sure that the right people are available because it's not easy to find, right? 22:45 So a lot of CIOs are struggling to build a good team that can both internal, external resources to build a good team that, that can get that done. Challenge number two is AI is an engine, you need to... 22:59 The, the fuel is data. Uh, making sure that data are, are, you know, available and clean. You know, the amount of, uh, uh, work that is required, uh, to prepare the data for AI to be, uh, actually effective is huge. 23:14 And a lot of companies, because of all the legacy systems, all the complexity, are actually having, uh, you know, yet more work to do to, to prepare the data so it can be, you know, properly fed inside the, those AI algorithms, right? 23:31 And, and probably the last point is, um, having to balance this innovation expectation with, uh, the side effect of the innovation, which is operational cost, you know, security concern and all of that, right? 23:49 So this, what I call the 80/20 rule, right? Most of the CIOs are spending 80% of their budget maintaining legacy stuff and 20%, and, and frankly very often, uh, less than that, to drive innovation. 24:03 So that's also becoming an issue in the AI and automation space. 24:07 Once you start having a, a significant footprint of AI and automation technology, you have to support that, you have to maintain that, you have to secure that, and that's a challenge, right? 24:17 So people, data, and I would say overall, uh, operational impact of those, uh, solution, uh, are clearly, uh, you know, uh, for me the top challenges that I see CIO having in the automation and AI space. 24:55 I think you're on mute Yeah, Dan, I think you're muted. Sorry, Chris, I was just saying do you, you, you agree with, uh, with, with OG? Anything you'd add? 25:05 No, I very much so, and, and I think along with that, um, I'm, I'm glad you bring up the data stuff because it's, it's becoming more and more evident, I think. 25:14 Um, one of my rules of thumb in trying to help organizations, uh, figure out if they're on the path to success is look at your budget. 25:23 If, like, you least-- If you're not spending at least half of your budget in one of these automation initiatives in procuring data, shaping the data, engineering the data, and, and ensuring that you have the right data, you're probably gonna be in the ninety-five percent of organizations who don't really succeed with it. 25:39 Um, and also you bring up people. The, the people factor, you keep accelerating technology change, uh, technology keeps getting better and better, the thing that stays the same through all of it is us. 25:53 So, and with you, I, I would agree, um, people factors are among the, the, the greatest determinants of success. 26:00 Do you also agree, Olivier, that that's actually becoming even more so now that the technology is getting even more advanced? Yeah, hundred percent. 26:09 And, and I'm, I'm totally with you on the fifty-fifty from a cost standpoint. Uh, uh, preparing the environment and the data and making sure that, you know, it is, it is ready to be fed inside an AI engine- Mm-hmm... 26:23 whatever that engine is, is very often overlooked by, by companies, and they don't really understand that cleaning and preparing data and structuring data is, is potentially, you know, depending on the baseline, a, a huge headache, right? 26:38 So it, it requires a significant amount of, of budget to be, uh, you know, targeted at that exercise. Otherwise, uh, everything happening, uh, uh, after, uh, will be of poor quality. 26:51 Uh, and, and the, the outcome of the AI algorithm will probably be weak. 26:57 I-- If you, if you heard the, um, our initial discussion about UnitedHealthcare and the billion-dollar loss because of security, uh, penetration, I wouldn't be surprised if someone did an investigation and found that, uh, some procurement committee spent about two million dollars talking themselves out of a five million dollar investment in some, uh, security software a couple years ago. 27:18 It's something, right? Yeah. Yeah. 27:21 And, and then I think this is a very typical mistake to focus all the energy and most of the money on the shiny object, uh, which could be, you know, yet another very cool solution, uh, out there, uh, and people tend to forget everything else, right? 27:40 So the, the what's happening before solutioning, so making sure that you are running in the right direction, you know what you are trying to achieve, everything happening after solutioning, including support of your solution and the data, right? 27:53 So it's very typical. People are, you know, uh, attracted by cool technology. It's, uh, it's-- I think it's, uh, one of the human flaw, right? So OG, what's your- I'm sorry, build the feature and above that. Yeah. 28:12 Uh, uh, OG, I was gonna ask you, you know, what's, what's your advice when you, when you speak to, uh, uh, a, a CIO about that, that measurement thing? I mean, I, I, I speak to so many members who, 28:24 you know, tell me that they're struggling with, you know, the ROI of automation. And I, and I, I wonder if AI will just be the, you know, like it is in so many other ways, the, the acceleration of the same. 28:39 Um, uh, but, but, but, but, you know, a lot of them also, I, I think it's, it's an unfair, uh, ask, you know, what the ROI is because I, I don't know that they, you know, spent the time to really figure out what they should be measuring or whether they're measuring the, the, the right thing. 28:57 So, you know, what, what's your advice when you speak to a CIO on, you know, how they should be me-- you know, what are the metrics and, and how, how should they be measuring success? Yeah. 29:08 And I would say I would probably extend that to more than just the CIOs. 29:11 But the, the CIOs are typically providing services to the rest of the organization, uh, and they are seen as the, uh, central point for solutioning, uh, those innovation. 29:27 When we look at the overall ROI, I would say both the CIO wants to have an ROI, and the business unit leader wants to have an ROI as well, right? 29:37 One as the internal client, if, if, you know, that's a bad word, but the internal client, and the other one as the internal service provider, right? 29:47 So to answer y-your question, I would say there's, there's a number of, uh, typical mistake that I see happening over and over again, and, and what we've seen in previous wave of technology we now see with GenAI and any other AI capabilities, right? 30:06 So the, the first one is to make sure that you very carefully select the technology based on what you are trying to achieve, right? 30:15 There's always a tendency to buy something that is too big, something that is too expensive, and if you don't apply that to an opportunity th-that is going to be, um, matching that size, then you're gonna have a poor ROI, right? 30:34 So helping understanding what are those opportunities, if you look at, uh, the full landscape inside a, a company, inside a department, uh, inside a, um, a business unit, it is really important to understand the size of the opportunity. 30:52 And based on that size, you will right-size the tooling that you're gonna be using, right? 30:58 So if you have a, a ten million dollar opportunity And you buy a half a million dollar tool to, uh, tackle that, that makes sense. 31:05 If you have a 50K opportunity and you buy a half a million dollar tool, uh, you're gonna be bleeding money, right? 31:11 And that is a very typical mistake, not to rightsize the tooling, not to rightsize the project to the opportunity. Why? 31:22 Because people don't spend enough time understanding the opportunity, baselining the opportunity, measuring, uh, you know, the, the overall impact, uh, that will, uh, that, that will happen once you deploy a, a technology project, right? 31:39 So this is something that we do all the time is, yes, it's exciting. Yes, you wanna start running towards that new shiny object. 31:46 But take your time to understand your environment, to make sure your compass is working, and to make sure you have identified areas that are correctly understood in terms of size so that you can correctly solution, right? 32:00 Again, there's, uh, tools out there. I'm going to take a practical example. Uh, GenAI chatbots. You have GenAI chatbots that are from $10 a month to $50,000 a month. Which one do you use? 32:15 Do you need this back-end API integration to your ERP system that justify a more expensive tool or you don't, right? So understanding what you are targeting in terms of opportunity will drive all the rest. 32:29 So that's, that's probably the biggest, uh, um, uh, um, mistake that I see over and over again. Uh, and, and while this is now under control in the RPA space, I can tell you this is not under control in the AI space. 32:46 Uh, Olivier, if I may, I, I can't agree with you more, more strongly, or, uh, you can't overemphasize this enough, and this is a point I made in my book on RPA. 32:57 For the first half century of the IT revolution, the cost of acquiring the technology was much higher than the cost of operating, right? 33:05 So I'd, I'd, I'd spend a quarter of a million or a quarter of a billion dollars buying SAP to implement it, and it would cost me maybe $10 million years around. 33:14 And so for half a century, our entire paternal of this technology was aligned with really high risk of making the wrong acquisition decision, and the, and the wrong decision wasn't that important. 33:26 When we put everything in the cloud, and we saw this, uh, very clearly with RPA, we totally changed that dynamic, where the cost of acquisition dropped almost nothing, but the potential cost of operating is dramatically higher than it ev- than the cost of acquisition. 33:40 And it, and it sounds like you kind of see that as well. Are organizations seeing that huge change in how you need to value the technology and how you even do an ROI calculation? Yeah. 33:53 So that's, that's a, a very, very good point, right? And, and I would say, 33:58 uh, again, there's a number of, uh, objective that organization have, a number of dreams that organization have that are sometimes, uh, uh, pushing them in a direction that might not be the right one, right? 34:12 Um, so to your point, um, a lot of those, uh, uh, solution, uh, out there, a lot of those products out there, a lot of those platforms out there, uh, require a fairly large amount of activity in order to tailor them for the environment, in order to, um, configure them, in order to, uh, implement inside, uh, the right use cases so you get, uh, the expected impact, right? 34:42 So the overall effort is being, not just the tool, but everything else, uh, in all the way down to the support, uh, nee- needs to be really well understood when a company is picking a tool, right? 34:55 And I, I'm almost, um, um, um, you know, a big, a, a big fan of starting, uh, maybe with a, a tool that is slightly too small, slightly too weak, slightly, uh... 35:11 And, and, and, and go up rather than trying to do the, the other way around, right? 35:15 So if you start with the, I don't know, the Ferrari, you know, uh, and you're stuck with it, you will not be able to go down to buy, uh, you know, a, a regular car, right? 35:27 So that's really, really important for people to understand. Um, nowadays, uh, people think, "Okay, I'm, I'm, I have a subscription, I can switch. I can switch from this one to this one." 35:38 Well, that's not that simple, right? 35:39 Once you go and pick a, a platform, even if it's subscription-based, and, and you start putting a lot of energy, a lot of integration, a lot of content creation, you, you, you, you know, you, you might have, uh, some difficulty to switch. 35:53 So this is a really, really important step. And in order to do that, in order to pick the right technology inside your, your solution, you need to do all the prior steps. 36:04 You need to be really clear on what you are looking for and what you are expecting in terms of impact. I, I repeat, right? 36:11 A 50K impact and a $5 million impact is going to drive totally different, uh, technology decision. OG, what do you, what do you think is gonna happen next? 36:25 You know, a lot's been talked about, you know, uh, some, some folks in the industry are throwing around this term RPA is dead. And, you know, certainly the majority of the conversations I have are, are more around 36:38 AI or at least, you know, in more intelligent automation. Um, w- how do you see the future of automation, uh, as it relates to, you know- Yeah... RPA versus GenAI or plus GenAI? Yeah. So this is... 36:53 So first of all, this is happening all the time. Uh, every new wave is killing the old one. So that's, that's, that's a fair game. 37:01 Um, I would say, and I said that at the beginning, RPA is, uh, dead in a sense that it is a commodity Meaning that this hype around it is kind of going away, not because it's the, the, the technology is going away, but because the technology is becoming, uh, embedded everywhere. 37:22 And, you know, no one talks about, uh, Excel or Teams or, I don't know, Slack, right? I mean, the tool you use on a daily basis. 37:30 So it's that in the sense that it's not hype anymore, it's part of the tool set and it's, uh, it's, it's old. It's not new. That, that's why people consider it's dead. 37:41 But to, you know, to your point, I would say automation, whether it is in the form of an RPA tool or any other form of automation, workflow automation and all of that, is actually being merged with AI. 37:58 And what we see is actually RPA, if we stick to RPA, RPA company moving towards Gen AI and Gen AI company moving towards automation. So what... 38:08 And, and what I use as a, as, as, as a mechanism for people to understand is, you know, you have the brain, which is the AI, and then you have the body that can do stuff, take action, and those two are being merged. 38:22 The automation and the AI are being merged, and what we are going to see is, of course, the hype is now all about Gen AI, all about what- whatever it is tomorrow, GPT 5.0 or whatever it will be, large action model, you know, whatever we wanna talk about. 38:38 But bottom line, the automation and AI are going to be combined to give way more power in terms of what you can do. 38:47 You have more, uh, cognitive capability, decision-making capability, potentially predictive capability, and you need to have action. 38:56 A brain that does, does not take action is never going to be as useful as a brain that can take action, right? So those two will be combined and, uh, there will be new terms, new 39:09 hype, but automation and AI are here, are here to stay, for sure, together. If I, if I may, Olivier, I love how you point out the, the need for action. 39:19 Actually, the worst case ROI you can generate is getting to an insight from analysis and not acting upon it. 39:27 Because you would build all the cost of setting up the technology, getting the data, doing the analysis, coming up with the answer. If you don't then act upon it, it, it is by definition a negative ROI. 39:37 Wouldn't you agree? Fully aligned. Fully aligned. So we used to have, 39:43 uh, many, many years ago, and I've nothing wrong, um, um, you know, with consultant, but we, we used to say, right, the consultant are basically telling you what to do, but they are not doing it, and therefore, you know, no action. 39:56 Uh, we, we have now with AI and some great platform, amazing insights, and someone needs to action that, and that someone is going to be automation. 40:08 That someone is going to be people who can basically take the insight, solution it into something that will do something, right? 40:15 So again, combining the, the AI findings, analysis, mining, uh, all that insight that can be found and injecting it inside an automation platform that can take action is really going to be the recipe for success. 40:33 OG, what's, uh, what's got you most excited for the, for the next couple years? I mean, uh, you know, like, like you said, the, the, the use cases seem kinda endless. 40:43 I see new ones every day, which, you know, I can't say the same for with RPA and automation. Um, but, um, a- any- anything kinda make you, you know, most excited over, uh, capabilities over the next year or so? 40:57 Yeah, there's quite a few, a few cool things happening, and I would probably segment here the, uh, the enterprise world from the consumer world, right? Yeah. Um, 41:07 I, I would say this, this AI enablement of product and services that we see in company is going to be amazing. 41:15 Because what we are expecting from this is an amount of customization, an amount of, um, um, quality. Th- Think about conceptually, right? 41:31 Um, N- Netflix, uh, times 1,000, meaning, you know, Netflix understand what you like, and it basically just an algorithm that can suggest the right thing. 41:42 If you apply this to everything you do, everything you buy, and if you combine that with edge AI, meaning if that capability now sits with your fridge- Yeah... 41:53 with your TV, with your car, with your, you know, everything, uh, that you are using in the kitchen, uh, this is going to be amazing, uh, innovation in terms of product and services are extremely augmented by the AI capability, right? 42:10 Now, on the, on the consumer side, uh, I would say the, the biggest revolution here, and I don't know who's going to win that, win that race, is this, 42:20 uh, AI assistant, that this digital twin that people are dreaming of having that- Yeah... you know, extra brain, external brain. 42:31 Think about, you know, AI being able to understand all your environment and remind you and offer you to automate tasks that, uh, you are doing on your PC. Uh, uh, you know, this is going to be a game changer. 42:46 So we are going to see a lot of exciting innovation on amazing product and services coming from companies, AI enabled, and AI assistant that will make our life, even our work life, way better. I, I love the AI assistant. 43:02 Well, I, I love kind of getting beyond and, you know, all my, all my old buddies are, you know, shared services guys, so, you know, not, not to knock the, the back office and the, and the finance, but yeah, I agree with you, uh, OG. 43:13 There's so many, you know, new use cases that, that Unilever story, uh, about them, you know, using AI modeling to create new, you know, products, but also personalization and variations of, of products. 43:27 Um, s- you know, there's a lot in the customer experience space to elevate customer experience and, you know, the digital twin thing is exciting. I know people... 43:35 you know, kind of makes, makes people nervous but, uh, uh, I, I'm all for it. I'd, I'd love to not be doing the many of the things I've gotta do every day, right? 43:43 Uh, especially as, uh, you know, someone running a- an association or a business. Chris, any others that, uh, have you excited, that you've been tracking? Well, I, I, I love what, what you're mentioning now. 43:55 I, I refer to it as, uh, attentionomics. We'll be monetizing actual human eyeball minutes, and whoever can capture the most actual human eyeball minutes is going to be the, the owner of attentionomics. 44:08 Um, but I, I, I think the other thing to keep in mind, and, and I know you touched upon it as well, is, is this notion of the automation, at least intelligent automation, is really helping uplift some of the lower half of people doing whatever the task happens to be, 44:26 and it can, it, it can enhance the speed with which top performers perform. So I, I'm curious if you're also anticipating that sort of bifurcation of the labor market where you're gonna automate away a lot of the lower, 44:42 I guess, I guess lower value added, less complicated tasks because they're, they're gonna be fully commoditized. 44:48 But then we're gonna see an, an even deeper enhancement and then deep, a deeper valuation of human creativity, human input, thought work, and so on. Do you see that sort of bifurcation continuing? 44:59 We saw a lot of that, I think, in RPO as well. Yeah. So, so there's, there's a few, uh, very important aspect here. So, so first of all, um, uh, we know that most of the job 45:15 will be impacted by AI, not, not just the lower level part of the job. So for example, doctors, 45:22 um, will be, um, impacted and there's a number of things that doctor are doing very well, but there's a number of things that doctor are, are doing, uh, less well than machines like, uh, skin cancer, for example, right? 45:37 Understanding from a picture whether, uh, you know, there's a risk or not. We know machines are doing that, that m- so much better. Uh, it's so much more accurate. 45:48 Uh, so we are, we're gonna have pretty much everyone impacted. 45:52 Um, we don't really know exactly to what extent, you know, some jobs might disappear, some new jobs might be created, and to what extent, you know, we will free up time for human to do some, uh, cool stuff. 46:05 But I, I, I think that the key message here is no matter what, what you are doing today, if you don't adopt AI, so if you don't accept to be augmented by AI capability... 46:20 Uh, I, I, I give you an example, uh, software developers, right? Some of them are embracing AI to write code. Some o- some others are fighting against that because they think that, uh, they know better, right? 46:33 And that's going to be, for me, how we are going to split the world in the future, right? We are going to split the world by people who are adopting and augmenting their capability. 46:44 Uh, we are saying up to 10X the current productivity that a human has, uh, will be achieved if you fully, uh, uh, adopt AI. 46:53 And you're gonna have other people that will reject AI because they consider that either they are afraid or they consider that it's not, uh, yet good enough, and those people might be left o- left over, right? 47:06 So it's, I don't think it's really related to the nature of the job, it's more related to the mindset of the human, whether they are willing to embrace and reskill and augment their capability. 47:17 And to those who are doing that, um, I think that, you know, we are really going to see, uh, uh, um, uh, really, really good news in a sense that, you know, career, uh, you know, the, the, the, the projection here is, is going to look really good. 47:34 Those who don't want to adopt, those who are fighting, those who are missing the train or the boat, whatever you wanna call it, they will be impacted. 47:44 I, I guess if I can add to that, I'm, I'm, uh, sort of working with this one startup company that wants to use AI to get improved healthcare outcomes, and the example they gave is for this particular ailment that they're dealing with, in a particular African country, for example, there are, I think, the, the number of specialist doctors in this area is measured in dozens. 48:05 It's under 100. 48:06 And the number of people in the population that have this ailment currently is measured in the hundreds of thousands, and then there's o- only half the people in the country that have this ailment are aware that they have the ailment. 48:19 So right there, simply by noting that other half, uh, the doubling of, of population people who have the ailment, are you gonna double the number of specialists? And that's in the dozens of people. 48:29 AI might be the only way of appropriately dealing with those people now. And, and, and that's a real life use case that I was just working with last week, so. Yes, 100%. So this is, th- for me, this is all exciting. 48:43 Once, once we are past the, you know, the sci-fi movies and all the fear and all the, the bad news stuff, the Terminator or whatever, you know, you wanna... 48:52 I, I think, I, I think, you know, and there's, there's always a risk with everything we do. Uh, there's always a risk when you cross the street that you're gonna be hit by a car, right? 49:01 I mean, my point here is there's risk everywhere. It's for us to manage. 49:06 And, and, and governance At different levels, you know, at the, at the individual level, at the enterprise level, at the country level, at the world level, are going to be really critical. 49:17 But the point here is this is good news. It's a good news conversation. I love ChatGPT. 49:22 I, I use it every day, and it's, while there's a number of, uh, weaknesses, uh, there's things that are amazing me, you know, on a daily basis. OG, thanks so much for joining us. It's good to see you as always, buddy. 49:36 And, uh, hopefully we can catch up with you in a, in a couple weeks and, uh, see what else is new. It changes so quickly, so, uh, a year from now is way too long. One thing new. I'm using two minutes here. Uh, 49:49 I, I actually wrote a book on gen AI that will be released on May 1st, right? 49:54 So I will be putting a bit of, uh, noise out there, and for those who are interested, so we wrote a, a guide to articulate how to best leverage gen AI in, in a business environment, 50:06 uh, with, with my, uh, co-writer, uh, uh, Kieran Gilmurray. So again, a bit of, uh, a bit of noise here. If you guys are interested, the book will be released in May 1st. Great. Send us that link. 50:18 We'll send it out to, uh, to our members and definitely have you back on the, the show. Kieran I think's gonna be on the show in a few weeks as well, so, uh- Cool... 50:25 we'll make sure to kinda talk through that, uh, that book. But OG, great seeing you. Thanks again. Thanks for your time. Chris- Thank you... good seeing you as always, buddy. Hopefully we can get- Yes... 50:35 some AI to fix your audio issues. Uh, I, I sincerely doubt it. It's, it's that, you know, through Sigma out, uh, problem of it. I'm gonna start a GoFundMe page for, for Chris to get a microphone. 50:46 [laughs] But, uh, good to see you, buddy. Thanks everybody for joining us. 50:49 If you need help with this, if you're looking for, for help with an automation or AI, just an assessment of where you are or a roadmap of where you should go, uh, we've got a team of people doing them, folks like Chris and OG. 51:02 Uh, let us know. Go to erpaai.com/assessment. Uh, we're off next week, but we'll be back on May 1st, and then we're back to live every we- uh, every Wednesday, uh, at noon Eastern, 9:00 Pacific. 51:15 Check out erpaai.com in the meantime. Have a great week, everybody. Thanks for joining. 51:19 [outro music]