Why the Key to AI Success is Leadership, not Technology with Andreas Welsch

Speaker A: Welcome to Leaders of Analytics. Leaders of analytics is about data driven decision making, modern business leadership, and the use of data and artificial intelligence in business and society. Each episode investigates the strategies, tools, techniques and leadership required to succeed in a world increasingly driven by data and analytics. The show's guests share their stories and experiences in a way that helps you understand the big concepts and small details that make all the difference in today's world of business. I am your host, Jonas Christensen, and I hope you enjoy listening to this episode of Leaders of Analytics.

Speaker B: Andreas Welsh, welcome to Leaders of Analytics. It is so good to have you on the show today and I am happy to have you here for two reasons. One, you've just released a new book last week which we're going to talk about today, which is A.I. uh, Leadership Handbook. So that sounds like something that everyone needs in their hands if they want to be a business leader in the coming decade. And secondly, I was actually on your show a couple of months ago, so I'm really happy to be able to reciprocate and uh, have another conversation with you and hear all about you, your career and where you've come from. So I'll hand it over to you.

Speaker C: Yeah. Hey, thank you so much for having me on the show. I'm really excited. I know we've been in contact for quite a while and been following each other on LinkedIn for going on three years now, I think, uh, when you, uh, published your first book and your second book. And I get so much value out of the conversations that we've been having, having both in person and online. So super excited to also share some of the information with your audience.

Speaker B: Yes. And we have this important topic to discuss today, which is the contents of your book. But before we get to that, let's learn about you and who you are. So could you tell us about your background and who you are, what you do?

Speaker C: Yeah. So look, I've been living in the US on the east coast, actually, uh, near Philadelphia for the last 14 years. But as you can probably tell by my accent, that's not where I was born and where I grew up. I originally grew up in, in Germany. And you know, my wife and I wanted to come to the US for, for two years and see what it's like to, to live here and work here and make friends and like many of those expat stories go, probably yours is as well. There's another year and another year and another opportunity. So we're looking at 14 years already actually this month. And ah, I've always Been in and around the IT data, uh, AI space throughout that career for, for more than 23 years now. I used to work at SAP, one of the leading enterprise application vendors in the world and have had many different roles throughout my career there. And the most relevant ones and the most exciting ones were along the last eight years from helping Fortune 500 customers in North America understand what is machine learning and where can you use it and what can you use it for to building a center of excellence in SAP's largest development organization for products that's called S4HANA around enterprise resource planning. So many, many of the large and even small companies use these kinds of products to manage their material flow, employees, finances, you name it, anything that business needs to operate and run smoothly. As you can imagine, there are tons of opportunities to use data that people have in their systems, uh, how they interact with companies, how they interact with systems. So the goal was how can we find those nuggets, those really important use cases, the drive value both for the customer as well as for the company as a software vendor. And then lastly, last two and a half years I got to apply my experience from doing business development, from building AI applications in a more product marketing focused role. So on one hand there was working uh, very closely with sales and go to market. How should we evolve SAP's commercialization and commercial model with generative AI now being on the map, but also how do we communicate to the market in very simple and succinct ways where the value is from AI in a business and for a business. So those were the last 23 years, but now since July, uh, I'm an independent AI advisor because I do want to help many, many more businesses and leaders figure out where and how can we use AI, uh, in our business. That's what's getting me really, really excited. And also that's why I've been talking to more than 60 AI leaders and experts, um, in the prepar of creating the AI Leadership Handbook and you know, collecting the best insights there as well.

Speaker B: Nice. And uh, I'm imagining you the young Andreas starting at the SAP and it's right after the dot com burst, right? We have the dot com boom and the dot com burst and everyone's probably a bit jaded about the Internet and technology and all this stuff. It's a bit hyped. And in that time you've been there 23 years at SAP, you would have seen the whole evolution of technology and how we use data and all this stuff. And now uh, AI in organizations and in A way that is in fact not the hype but the real day to day. Because SAP is an enterprise system for so many corporates in so many industries. And it's really sort of embedded in this, in the organization because in many ways it's the thing that facilitates the organization. So it's not like working at Facebook or Google where everyone's these young 20something Stanford PhDs and so on. It's working with banks, it's working with insurance companies, it's working with uh, manufacturers, the old school industries and trying to bring them up the technology curve. So you would have seen and experienced so much in that time.

Speaker C: Yeah.

Speaker B: Could you share with us some sort of milestones or critical moments in that time that you think back on as pivotal moments in the world but also for you and the way you thought about you and your career?

Speaker C: Yeah, you know like I, like I was mentioning, I started out in it, which I believe gave me an excellent foundation for understanding how do different systems work, how does security, networking, compute, storage, all of these things work. Even at a time before we were talking about cloud. So when you still had data centers and you had your servers in your data center, uh, in seeing that transition over to the cloud, in large enterprise software vendors going through this enablement, this explanation with again this uh, C suite customers on the other side of the table who are questioning well why should I move my data to the cloud? Don't you have access to this? Don't um, others have access to this? Remember these early conversations? I feel in many ways we're at a similar point now with AI, especially when it comes to data sharing. But back then seeing this evolve for, for me there was an immense opportunity moving out of it into more of a business incubation role. The idea was for more regulated industries, can we provide cloud services on their premises? So have servers, have a rack in their data center with a cloud like infrastructure that uh, scales up, scales out with application management services and these kind of things, concepts, you know that sound very foreign today when pretty much everything is in the cloud and when we do have these resources and elasticity. So one thing, the next one, next big shift for me was looking into something that was called machine learning. I must tell you, uh, when I got my degree in 2010, just before that I had an elective in those studies that was called artificial intelligence. And back then in it I figured, well, I would never use AI. Where would I use AI, I.e. robots and futuristic stuff. And you know, fast forward a couple years and it's, it's here, it's real. So I get an opportunity to talk to too many leaders, help them understand, where do you use machine learning? What is that? Even at a time when we're talking about the first hype, now five, six years later, we're at that same point again with generative AI. And I think many people in the industry are saying this, they're even seeing this. We're coming off the, uh, peak of the hype cycle, which is where I believe it gets really, really exciting. Because now is the part where the rubber meets the road, where we need to show as leaders in data, in analytics, in AI, how and where we can use AI to really make a measurable impact. It's not just about chasing the shiny objects. It's not just about figuring out where we can use ChatGPT or OpenAI's models or, uh, anthropics models or who have you. It's about how can we use them to drive a meaningful and measurable business impact. And that's, by the way, where I believe we need to start these conversations with our business leaders and where them having access to these tools on one hand makes it a lot easier because they have ChatGPT on their phone, so they have experienced it. On the other hand, there are also a lot higher expectations when it comes to, you know, making this happen in the enterprise. And we know as practitioners, it's not quite that simple.

Speaker B: Yeah, so chatgpt is a kind of a help and a hindrance at the same time because it's, it's made it real for a lot of people that the AI is something that can help us in our business. But when we come with a much more mundane use case, and it's not chatgpt esque, then why not? What's wrong with this? Is that kind of what you're seeing?

Speaker C: Yeah, that's exactly it. Even a couple of months ago, I was talking to two leaders at a large pharma company and they were saying, well, leave me alone with machine learning. That's not AI. We want the real AI. And so what's the real AI? Well, generative. Well, that's fine. That's one kind of AI, but it's not the only one. So as leaders in this space, we also need to help enable and educate our stakeholders and say there are many different flavors of AI, if you will, and some of them are perfectly suited to, um, do your demand forecasting, your recommendations, your classification categorization. And they've been around for a long time, they're very robust Methods and algorithms, and then there are others where it's more about creating information and summarizing and translating and all these kind of things. So really help them understand what you use them for. Again, I think that's a big role as a data leader, analytics leader, AI leader that we can facilitate.

Speaker B: Exactly. And we've done about 60 or so episodes on this podcast, and if I had to pick one concurrent theme throughout, uh, pretty much every episode, it is exactly that, that the role of analytics leaders and people who aspire to be those leaders in their day job, they basically have a huge task of not just producing this technical output, but calibrating people's expectations and helping them move up the maturity curve that is inside their heads, their understanding of what can and cannot be done. So you're actually, uh, an educator to a large extent, uh, in your organization. Whether you like it or not, that's the job.

Speaker C: That's true, yes.

Speaker B: And I think a lot of people don't realize that straight up. They may realize too late through challenges and mistakes.

Speaker C: That's right, yeah. And see, by the way, that's where I, uh, want to help with the AI, uh, Leadership Handbook to give a more practical guide, not just around technology. I think so many of us in this space are exciting, get excited about m new technology, and that's awesome, right? It allows us to do many, many different things, new things we haven't been able to do before. But I think also unless we put this into a business context and we understand what is the impact that we are expected to make and that we are able to make, technology in itself is rather meaningless. It's a tool, but without knowing what I use it for and what the, uh, end result is that we're trying to achieve, it gets really, really challenging. So look beyond technology, look at your strategy people. How can you bring along others on their journey? How do you find those, those valuable ideas, valuable for the business, for your customer as well? That's what I think as an AI leader, you need to focus on a lot more.

Speaker B: Yes, I agree. Now, let's get back to this in a minute, because this is what your book is about, essentially. But I want to go back a little bit, Andreas, uh, because you've, you've ended up now as an author and also, also as this independent consultant, you've started your own business this year and that hasn't come about in five minutes, of course. And I, having looked at your journey from the outside, this is how I interpreted that around 2021, you decided to start posting regularly on LinkedIn. That's why when you started, uh, showing up for me as well and uh, you've kind of taken a whole different career trajectory since then and I suppose I assume it's, that's what's led you to where you are today with the reciting your own business. Can you tell us about that time? What made you start sharing your knowledge and ideas on social media and how's it impacted your life since then?

Speaker C: So at that time, around 2021, I had a very influential senior leader who was mentoring me and one of his points of feedback was, hey, Andreas, you should do a lot more thought leadership around AI and what you see in this space because I feel there's a gap and they got me thinking. And so many of us in corporate, we have marketing departments that put out content in the domain that our company is active in. And for me doing thought leadership meant I reshare the corporate blogs. Hey, look, great feature and product XYZ for supply chain and you know, great feature in the HR software with AI. And I thought that was thought leadership. Now my colleagues and peers with within SAP definitely seem to agree. My bubble grew. My SAP bubble grew. Everybody said, yeah, great, we're the best, we're the greatest. But I saw that right, there wasn't a lot of engagement from, from others. Yeah, nobody from, from the outside would, would engage with this. Um, unfortunately there were no customers. No, you know, not the audience that I was trying to reach was, was engaging with it, rather just be their own echo chamber. And so I started tailoring and intriguing my approach, more personal perspectives and opinions. I, uh, thought, well, you know, what are others seeing in the industry? Based on my own experience building and leading a center of excellence, some of the challenges that I have gone through, like you said, many of us doing this for the first time might not have the proper information or might not have all the background. So you make a lot of these learnings, these hard learnings as you're going through it. Well, I see my mission, ah, as helping other leaders who either are just starting in the role, who want to become an AI leader, maybe are even in that role today and want to expand what they're doing and how they're doing it and help them with information, you know, from a broader community, also validate a little bit what I've seen and what I think and thought is true and might be the best approach. And did I go about it the right way when I was doing it? You know, is there a reason maybe why it was hugely successful or why it immediately failed. And in many, many cases, I saw that others are going through that same journey. So maybe number one is if you think you're going through this for the first time and you're going through this alone, trust me, there are many, many others that have gone through the same journey as well, and they're asking the same questions. Why am I not getting enough support? What are the things that we should be looking at? Why is my leadership team chasing shiny objects? How can I help them understand where the real value is and think through that? How do I bring others along on their journey? So on one hand, they become multipliers for what we are doing, um, in our AI initiatives, but on the other hand, also become scouts for me and for us. They bring back information. They say, well, you know, we've tried this out in our finance department and it's working really, really well. Or it's not working so well, but we think there's something else that would deliver more value. So all of these things coming together here and sharing that through a range of different types of media. Right? Whether there was or is my live stream, what's the buzz? And people said, hey, you know, it's great. You go live every other week, but I'm in a different time zone. I would love to follow along. Do you think you can turn it into a podcast? Let me think about that. Maybe there are other people and more people who have a similar request or a requirement. It turned out to be true, so I turned it into a podcast. Another said, well, you know, I like reading more about these things, and maybe there's an opportunity for you to think about some, some broader concepts as well. So the idea of the newsletter came about. Anyways, long story short, I really want to help data, uh, it. AI leaders succeed with AI initiatives in their organizations. I think there's a huge potential now that AI and generative AI is becoming real. And that's what gets me excited. That's what drives me.

Speaker B: Yeah, brilliant. And tell us about some of the things that have happened when you've put this content out for you personally. So you're inviting a, uh, serendipity kind of, right? You're inviting a bit of randomness. You don't control what the algorithm does and who gets shown this content. And we were talking about the show before about, uh, all the random stuff that you and I get in our inboxes because we put content out. Some of it we would rather be without, but sometimes there's some really interesting stuff comes about can you talk about some of the things that might have come about that would have never happened if you hadn't put your own thoughts and ideas out in the world?

Speaker C: So there are a couple thoughts and a couple of things that come to mind. I think when you do create content, when you are consistently active on social media platforms like LinkedIn or YouTube, or when you have your podcast, you become visible. For each and every one of us, being content creators and podcast hosts and everything, there's so many people who are in similar roles who are not getting that visibility. Maybe they're not even seeking it. And that's perfectly fine. But what has been a positive effect for me is getting in touch, getting to know people through these social networks. It's not just about pitching an idea or making the next sale. Right? We all get so many cold emails, cold messages that are not even relevant, but really meaningful connections, meaningful conversations with fellow leaders in even different parts of the world. What are you seeing? How have you approached this? And everybody has a unique perspective, which makes this really, really interesting. So whether it's opportunities to collaborate, whether it's creating content together so you leverage each other's network, or it's opportunities to collaborate on projects, or even, you know, like our conversation here would not have happened had both of us not been active on social media. So I would say don't underestimate the impact and the force of community and engaging in meaningful dialogue and conversations online. Yeah.

Speaker B: Ah, I encourage everyone to have a go at it just because it'll, it'll give you a challenge and, uh, it'll also help you organize your thoughts, which is another benefit that actually helps you in so many ways. When you have to put structured content together, you get much better at communicating that structured content as your own ideas to other people. And, uh, life becomes easier as a result. Now, let's go and talk about your book, because I'm interested in hearing more about this and why you put it together, who is it for, and, uh, who can benefit from it? So let's start with firstly, uh, what inspired you to write the AI Leadership Handbook?

Speaker C: So in my conversations with fellow leaders, fellow AI leaders, IT leaders, I saw that there is a gap. People either have a very strong IT or data or project management background, and they're catapulted into these roles. Uh, hey, our senior leadership, uh, team said, what's our strategy? Holy crap, we don't have one. But Bob, you've always worked on data or you've always worked on analytics or it. Can you figure this out? Can you become the face of that mission? And then I see that people are struggling. Yes, they have a strong background in a certain domain, but most of the time, if they come from a technology background, that's the lens through which they approach it. And they might miss the connection with business teams. Understanding what are the business priorities? How does technology enable us to achieve these business goals? Whether it's at a company level or a unit level, or even a team level, we all have individual goals that are broken down. So really seeking that connection. And through more than 60 conversations so far on my live stream and podcast with AI leaders and analytics leaders like yourselves, I also saw that it gets a little cumbersome and tedious if you want to get this information across. A playlist of 60 videos take some time to go through. So I thought, hey, why not condense it and distill it in a book that you can read over a couple of days. It has less than 200 pages. The paperback put that in a nice format where it's engaging, very structured the information. How would I go through this again and approach it if I were a new AI leader and really make it accessible? And almost as a 360 degree view of what do you need to know when you come in this role? So the, uh, different aspects in addition to technology, like strategy, leadership, collaboration with different business teams, forming a community around these topics so people can exchange ideas, so they can share their learnings. How do you do that? How do you set up your organization? If you're building a center of excellence, especially if you're in a larger organization, how should you think about this from an organizational design, from a broad skill set point of view? We also have been talking a lot about ethical AI, responsible AI over the last couple of years and I'm glad that we have. In this industry, there have been so many examples early on where people have not paid as much attention to it and the negative side effects. So now more people are talking about responsible AI. But to me that goes beyond just ethics. Sustainability is quickly moving in as a new dimension with large language models consuming a lot more, more resources, a lot more energy. But then also on the flip side, if we use them for the right purposes, they can also help us save energy and ideally save more energy and emissions than they actually create. So thinking through those things, data, uh, data privacy, security, all those aspects I think are absolutely critical and relevant if you're in an AI leadership role. Not just the technology, not just what's the latest model that uh, OpenAI put out? The OpenAI01 uh, as the latest example, it's important to know about these things, but in business my experience has been there's a lot more change management, working with people around this to really succeed in adopting AI.

Speaker B: Mhm. Yeah. One of the things I often say to people in this space of analytics, data science, AI, uh, is you must realize you've probably chosen one of the hardest career paths for yourself. And that's because one, it's new stuff, there's no trodden path. But if you want to become the CFO, there's uh, 150 years of history of other people doing that. And you got to learn accounting, you got to get a cpa, you got to be the head of financial planning and analysis for a while and so on. This, there's a structure to it and everyone knows what those roles are and what they do in an organization typically this is all new. You got to invent it as you go. Uh, this is, you're actually kind of an entrepreneur in the organization. And the second bit is this, you got to bring everyone on the journey, education and collaboration, which is so hard. And you touch on some of these things in your book as well, which I think is great to have that included. And the challenge with that is people, often introverted, uh, technologists that are very detail oriented. So it doesn't necessarily come natural to be, be the salesman in the organization. So uh, what you have to do is also maybe your kryptonite at the same time. So this is a real challenge for a lot of people I think. And we talk about in most organizations AI typically falling to the CIO or the CTO to implement and overseas that they are also potentially a little bit in the same boat, very technology focused. In your book you talk about this. In many cases would be a new role of Chief AI Officer, which would be the person doing these things with their team. Can you tell us what that role does and why we need it?

Speaker C: We've seen similar roles be created before. It's the Chief Data Officer, chief Digital Officer, chief Analytics Officer or Chief Data and Analytics Officer. I think in many, many cases also the Chief AI Officer is really this new role to spearhead this effort. And chief AI officers in many, many cases focus on acquiring knowledge, staying on top of things. What is happening in the industry, what are the relevant developments? Yes, the new models, the new techniques that are coming out that we need to know and that we can use in our organization, that's definitely the strongest foundation, um, having that knowledge. But the second component is also helping your business organization come up to speed, learn about what is AI. Again, I firmly believe with generative AI it's never been easier to communicate that and to make it tangible for your business audience. Everybody can use ChatGPT, at least as a consumer and play around with it and you can give it a prompt and ask it a question and it does something for you and you get a result. You don't have to be a trained statistician or mathematician to use AI or to create something with AI. I think that's immensely valuable. And as a chief AI officer you need to contextualize this. What does it mean in our business? For example, just this morning I was giving a training for a large multinational company on how do we use Microsoft Copilot in our own isolated instance of OpenAI's models, almost like a ChatGPT and how do you prompt and what are different techniques? And um, I'm so excited and so encouraged by this and seeing many other companies do similar things and follow that path here to really also drive that enablement, make it tangible. What tools are available? How do you use them? The third part is around governance. Think about again larger organizations. There are many, many different stakeholders. For example in different business units. Maybe your HR department, your legal department, your engineering, your it, your sales, your supply chain, Everybody you know, you need to get them and rally them around the table and get them behind this idea that hey, AI is something that moves the business forward. But how do you do that in a structured way? How do you set up a program within your organization to drive AI on one hand, really top down as a strategic priority, but also get feedback bottom up. Where are the opportunities for us in a specific process? Then the last one is around vision. I think that's one of the most important ones to show where can this lead us? What can we do with it in our business? Also how can we evolve our business model? Many of these things sound pretty superficial and high level. But the goal and the objective for this chief AI officer is to make it tangible. If we have this technology at our disposal, thinking through maybe with your, with your service department, what would that mean? You know, we know we want to improve service quality or our net promoter score. Well, maybe if we can give our customer service agents capabilities like that at their disposal to first of all get a better history or a faster history of what has uh, a customer already reached out to us about? What is this new inquiry about? Is it related, you know, spin this further to drafting of drafting a response. Right. We start Cutting down the time to respond to an individual customer while still maintaining at least similar, if not even better levels of quality. So these are the kinds of impacts and things that chief AI officers can make. But in order to do that, they really need to be at the right level of the hierarchy. You alluded to that AI usually falls to the CIO CTO role. In my experience as well, there's hardly one department that owns AI, um, or hardly one role. It could be the head of AI, again, could be enterprise, um, applications leader. So making sure that the Chief AI Officer really communicates at, uh, eye level with other C suite members is really, really key. Because only then or especially then can they work at that strategic level and engage and go down into the finance organization, go down into the procurement, Procurement organization with the mandate that both the CEO or president and the other C suite members have given them.

Speaker B: Yeah, and this is one of the harder elements of, of this role, I think, which is you only succeed when others take advantage of the solutions that you provide. So you're implementing these, uh, AI products, data products in the organization, which they will be if they're built right. But the success is not to do that itself, but to get the usage and the results from that usage for the organization.

Speaker C: Yeah.

Speaker B: So how does it compete with the priorities of all the other chief roles in an organization? And how does it get a remit that is, I suppose, big enough or the influence that's big enough when it's a role more of negotiation and inspiration as opposed to what I imagine is, uh, for instance, a CIO can say by decree, this is how we use our it. The Chief Financial Officer can say, this is how we do some of our accounting practices. Etc, etc. This role is more of a negotiator than a, uh, determiner of how we do things.

Speaker C: So the way I see it, you need to have excellent people skills. It's about negotiation, like you said. It's also about influencing. It's about creating excitement. A lot of times we talk about creating a little bit of fomo. Fear of missing out. You start with one part of the business, you help them, you get some early wins, and you take this to the other part and say, hey, look what we were able to do with the finance department. You're in sales. Wouldn't you be interested in doing this? We're able to help them save a certain amount of time or dollars or help them. On the other hand, if you, uh, start with sales, sell so much more. So we can help you do similar things in your Business. So you get them excited and you say, well why are you doing that with this department? I want you to work with my department as well. I think also that's the best way to see the effectiveness. If you turn this, this push, can I work with your teams? And you convert it into a poll and they say, hey, I heard you work with finance. I'm in sales. Can you work with my team as well? That's when you see you're creating this momentum. But in order to do that, like you said, having this mandate is absolutely critical because even if you are in a CIO CTO organization, you're most likely constrained by the goals that these organizations have. Cio, uh, return on investment, uh, system uptime, maybe modernization, things like that. You're really more of a cost center cto. I'm painting with very, very broad brushes here, more technology focused. Where can we use technology? Probably not as close to the business. So your business stakeholders might see you more as a technologist or more as the IT leader than somebody that spearheads this for the entire company. So it's that mandate to have that chief AI officer or similar role at that C suite level because the CEO and the board recognize that AI is such an important topic that it cannot live anywhere else. It uh, affects ah, um, influences the entire company and it's a way to elevate your business strategy. I think if you think about it in that way from a company point of view, then it makes sense that AI has to be at the same level as finance, as procurement, as hr, at least for a certain amount of time until all these different departments are up level, are upskilled, have adopted AI to a degree where they are comfortable running these things independently, assessing where there are opportunities for AI. Um, and that's why I also believe the Chief AI Officer is a transformational and transitional role. We'll see if we have chief AI officers 10 years from now, or maybe even five years from now. Probably more five than 10. But at least at the moment it needs that central role that UM has and aggregates and updates the knowledge that does define the governance that creates the vision in these things to really move the organization forward.

Speaker B: Yeah, so it's really about ushering in AI and the use of uh, data and analytics in a much more advanced way than it typically is in an organization. And that's the transformation that we need to make. And uh, I think uh, the starting point is kind of a well known one for most organizations. So these are sort of commonly touted statistics of more than 80% of AI data science projects, they fail to get adopted by the business. So that's kind of the failure rate, which, uh, is not very good or impressive or something that the business should aim for. But necessarily that's kind of what still happening out there. And also in the episode I did with Randy Bean, we talked about, uh, some of his research where the average tenure of, uh, a chief data officer, chief analytics officer, chief officer. I'm kind of painting them with the same brush here. Their average tenure is sort of a year and a half or so. So they either get frustrated themselves or the organization can't figure out exactly how to place them, how to get value out of, of this role. They can see the need, but there's, there's a difference between desire and ability for the organization to actually do the things that it wants to do. So in your opinion, Andreas, where are the typical gaps in most organizations, uh, around AI leadership and what should sort of be the first things that someone in this sort of role or with this sort of remit, what are the first things they should establish to be successful?

Speaker C: I think many organizations already lack that central leadership role to begin with. Many leaders, business leaders, are being asked by their board, by their CEO, what is our AI strategy? What are we doing with AI? Uh, it sounds like our competition is doing something over here. We must make sure that we don't fall behind. What's our AI strategy? And on one hand that sense of urgency is really, really good and important because it's that opportunity to create that central role. But unless you know that there are many, many different facets and aspects involved in AI leadership in a company, it's challenging. So now even if you assume that there is a person, a leader that is responsible for creating this AI program, you get to the next stage of again, how do we set up the organization? What type of people do we need? Typically it's a mix of business analysts, data scientists, product managers, change managers as well, because there is a good amount of change involved if you think to introducing AI and people are used to working a certain way. I've always run through this process like this in the following sequence. I've been doing this for the last 12 years. Now you introduce AI. What does that thing? Know it better than I do or why do you tell me it knows it better than I do? I don't trust it. So it's about change management, bringing people along on their journey as well. Great. So you have your center of excellence in place. Now what do they do? That's the next critical phase. So you certainly don't want to have them sit idle and you don't want to have them work on just science projects. So it's a matter of finding good ideas, valuable ideas that make a measurable impact. It's not just creating happier employees or reducing the number of clicks. Maybe that's a byproduct and certainly everybody wants to have happier employees. But it's really, really hard to put a price tag on this in uh, a return. So what are your business KPIs that you're able to influence where are good opportunities? So build a pipeline, evaluate those ideas and then move them from ideation to validation like your pilots or proof of concepts implementation, really building that software capability to all the way operation and maintenance. Even. So thinking about this funnel, this process in a sense as well. And as you're moving from ideation through validation, in my experience, data, uh, again coming back to the age old topic of data, uh, is a key issue. People thought now with generative AI, we don't need data. These models know everything they do about language and what's the next likely word in a sentence. But they don't know what's specific to your business, to your context, to your industry, to the products that you make, to the services that you offer. So again you need data. And thinking even further ahead, AI agents, agentic AI, the big buzzwords of the last few months. If we get into a future and into a situation where there are software components that will work based on goals, no longer just on you know, structured uh, information and instructions. But you say, you know, maybe book this trip, this business trip for me to Canberra. Well, what are the airlines that go there, what are the hotels? Maybe the hotel chains you typically stay at. Loyalty program, you have the rental car, the company, right. All the things that maybe your assistant or business assistant would do for you. If you start delegating these things to that agent, all of a sudden it needs data. Good data, fresh data, accurate data. And so we're talking about these things again. So I think challenges are manifold along the way. And the last one just to end on is also the most important one. You can build the best product and the most technically sophisticated product in the world. If people in your business are not going to use it for whatever reason, all that investment up front was unnecessary. It's a money pit. So that's the last one bringing your people along.

Speaker B: Nice. The last thing. And the first thing I'd say start with that. Don't build anything until you've Done that. The graveyard of data, uh, science projects is very big for that reason, namely that we got excited about building something that no one ever used or was going to use. Hi, dear listener, just a quick message from me. Are you ready to take your data career to the next level? Then I've got just the thing for you. The Leaders of Analytics newsletter. Every week I send you actionable tips to master the art and science of analytics leadership and to help you grow your career as a data professional. Every issue comes packed with information designed to help you grow your leadership and, uh, influencing skills in the world of analytics. No big theories, just practical, real world strategies that you can start implementing right away. So head over to leadersofanalytics.comnewsletter to subscribe and start making a bigger impact tomorrow. Subscribe now and your future self will thank you for it. Now you describe in your book the concept of multiplier communities. Could you explain what they are and why we need such communities in an organization?

Speaker C: One of the biggest realizations for leaders in these AI roles who've gone through this, for example, during the machine learning hype and maybe even big data, is that you cannot go at this alone. You cannot do this out of your ivory tower. And not even, you know, with a team of 10 data scientists or multiple of these teams of 10, you really need to understand what is the business problem that we're trying to solve. And from my experience and that experience that many other AI leaders have shared on the show, it's starting with the question of how can we get closer to the business? How can we get multipliers? How can we get advocates, champions, catalysts, whatever you want to name them that are experts in their business domain, say finance, for example, but who have an affinity for technology, who are usually at the cutting edge of things, who are embracing change. We all know some people who kind of fit that profile in our realm. So if we give them more information about what is that AI thing, how does it work, when should you use it, when should you rather not use it, what are some great opportunities that it will bring that will help us unlock, and what are some of the challenges we need to be aware of as we use it? We give them some hands on training. Here are some tools that are available to you that you can use already in your business. And again, they don't have to be super, super technical. They don't have to be data scientists. Actually it's even better if they're process experts or if they're, uh, process owners are doing these processes day in day out because they will know where the issues are. If your job is to match incoming payments to open invoices, you know what's working well. And, um, where the exceptions are, especially the ones that take you a long time to figure out why there is an exception. So, um, these subject matter experts are an excellent source then also for information back to your center of excellence and to your teams to say, well, now that I understand what this AI technology can do for me, and by the way, also machine learning, let's not neglect that. Where are the opportunities in my line of work? Do you think we could do something here? You want to have these multipliers on one hand to advocate for you. Look, here's AI. This is how we can use it, this is how it can help us. But on the other way, also create that feedback loop to your center of excellence to say, hey, and here are new opportunities. I think when that works, it's super exciting to see because you also notice your multiplier starting to have conversations among themselves even when you're not in the room, even when you're not facilitating that. So they know that they can call each other and they can say, well, I heard you mentioned you went through this problem, for example, of data, uh, acquisition. And he said it wasn't that easy getting the data, but after, uh, six months you were able to get it. So if you were to do it again, what would you do differently? All of a sudden they start to have conversations and help each other and help each other figure this out.

Speaker B: Yeah, nice. Again, it's about creating that boss, isn't it? And mainly about communication. You got to have the technical skills to show up with solutions, but the job is at least 50% communication, if not more. When, uh, it comes to actually implementing these technologies. Really interesting stuff. And we're almost at the end. What I want to ask you now is to look beyond where we are now, of course, and to where we're heading with AI in the next, say, five to 10 years. What do you see as the opportunities, risks and disruptions that are coming our way?

Speaker C: Five to 10 years, that's a very, very long time in AI. You know, even five quarters is a long time these days.

Speaker B: Two years ago, we didn't have a chat GPT in the format we do now, so.

Speaker C: Right.

Speaker B: It's a hard forecast to produce.

Speaker C: Yeah. You know, if you asked three years ago, where is machine learning going to be? There was on a downward trend and people realized it's not that simple. I think very few people would have Said, hey, there's uh, a rejuvenation of AI and interest in this. So who knows in that sense. But you're also not asking for precise information or investment, uh, advice. So look, in five years I think AI will be even a much stronger part of the products of the services that we use. It will become even more invisible in that sense, more natural to use it. And maybe we don't even call it AI. Maybe the best thing I think that can happen is it's a new capability, it's a new feature, it's a new product. Kind of like when voice assistants started uh, to emerge. I'm confident there, there are many different patterns and usage patterns. Think of things like multimodality all of a sudden have audio input to create an image or to create a video. Before we went live for this episode, we talked a little bit about the art of the possible and more futuristic ideas. The thing that uh, I would love to see is when it comes to entertainment we have this renaissance of virtual reality glasses with the Apple devices, uh, with Microsoft's devices and others and I'm sure they will get even smaller. So you have these devices with plenty sensors that can see your surroundings, but they can probably also analyze your emotions. Think about watching a show today, it's pretty linear. I'm convinced that in the near future there it's going to be hyper personalized. If your virtual reality glasses are sensing or the sensors are sensing, you're maybe dozing off or you're not quite as engaged. You keep looking at your phone every other minute. Maybe the plot changes and there's generative AI that can tweak it to what you usually like maybe the hero drives off the cliff and all of a sudden your adrenaline level rises and uh, you're pulled back into the story and so generating this information um, of how that story goes. We're seeing this with image generation, video generation M. I'm thinking about this and uh, getting excited about seeing this as part of the future. Being able to use these technologies really to accelerate many of those things and again personalize them. Now entertainment is one area. If you talk about business it's the same thing. How can we personalize experiences? I remember talking to a cto, ah, of a large theme park in the US and he said well with generative AI emerging we're thinking about personalizing the customer journey. So you know, if you go to uh, the Minions park, we you know, craft emails and communication that's, that's more despicable me. Like if you go to the Harry Potter park, we use a totally different tone and language and personalize this for you. And while you're there, you're getting messages and information, again, personalized to you and to your package and to the things that might interest you. So I think there's a whole new opportunity for companies to have a customer experience right. At another level of what we're used to, and even simple things like customer service. I truly hope that they will get better with AI and because of AI, because there's a lot of potential as well. Again, making it more personalized.

Speaker B: Yeah, it's going to be a wild ride. Our, uh, world's going to look very different in 10 years. You don't have to forecast exactly what happens. But yeah, with all these things, especially entertainment, I think going to be so different. And the consumers that are growing up are also going to be expecting more of that. So that drives it. I'm sitting here thinking of a couple of friends that I have that read bedtime stories to their young kids at night. And, uh, their method of that is they ask their child of three, uh, things that happen in the day. Then they put those into ChatGPT, and ChatGPT writes a story about them as the character in the story with some of these elements of the day. And that's highly engaging and it's a basic version of your entertainment. But you can see that kid, they get exposed to very personalized content so early, and that's going to become not a wild thing for them because that's what dad and mom did as well when I was a kid, uh, when I grew up. So why not on tv?

Speaker A: Yeah.

Speaker C: Very fascinating. Yeah. Right. And in that example, you can also easily change the story. Well, no, I didn't like that part, or this isn't quite what happened. So tweak the prompt, tweak the story, and it goes in a different direction, you know, so it doesn't have to be all the tales and fairy tales and things that we were taught as kids.

Speaker B: Yeah, that's right. 1001 Nights, all that stuff. We can rewrite it all in five minutes.

Speaker C: Exactly.

Speaker B: And there's last question about your book. If there was one key takeaway that you'd want the reader to take from it, what would that be and why?

Speaker C: As leaders, and especially as technology leaders, we get so excited about new technologies and what we can do with them, what we can enable with them. It's an important starting point, but it's really just that, in my opinion, a starting point, there are many other dimensions of successful AI leadership that you need to consider, especially in business environments. We've talked about creating communities of multipliers to help you advocate. Finding the right opportunities to engage in and to pursue to begin with is another key one responsibility. With more AI, with great power comes great responsibility, and we need to be aware of that. And it's not just the ethical sides and the fairness and reducing bias, but also now more on the sustainability, environmental side, um, societal as well. What are the impacts of that? Releasing this algorithm, this model into the world will eventually have in many, many more dimensions. So one key takeaway for me is it's not just about technology. Technology is an enabler, but you need to have a 360 degree view of all the other aspects to make AI, uh, useful and successful in business and to turn the technology hype into business outcomes.

Speaker B: Nice. Thank you for that. I think that sums up the content of the book quite well. So I do encourage all listeners to go and check out the book. It's very affordable on Amazon. Go and have a look at it. You won't be disappointed with what you read. And there's some real gold nuggets in there that'll that'll help you just firm up exactly how you can roll out AI in your organization. Now, last two questions, Andres. Firstly, I ask every guest to pay it forward and the way you do that is to recommend the next guest on Leaders of Analytics. So can you tell us who you think should be the next guest on the show and why?

Speaker C: So when I was in Germany a couple of weeks ago and we recorded our episode, it was a perfect time zone to work with and record episodes with folks in Australia. I made one, um, excellent acquaintance, uh, by way of introduction, also through LinkedIn, who might not have been on your show yet. Marek Kovalki, who's a professor in Australia focusing on management information systems. I think he would be a great guest on your show and uh, he writes about and talks about the economy, uh, of algorithms when AI agents become your digital minions. So we had a very engaging conversation and I think also within uh, your times of probably audience would be a fantastic guest to have on your show.

Speaker B: Uh, look, I'm always happy when I don't have to get up at 5am to do podcasts, which most of them are recorded around that time. So that helps me in the first place. But he also sounds like exactly the guest we want to hear from. So thank you for that. Uh, lastly, uh, where can people get in contact with you if they want to have a conversation with you or get a hold of your content.

Speaker C: So the best way to connect is definitely through LinkedIn. If you look, uh, up my name on Andreas Welch, you'll find me there. Also, like you mentioned, I just released the AI Leadership Handbook as a way to help new leaders learn about all the different ways they need to consider when implementing AI in their business. You know, um, I'm energized by helping many more leaders go through this journey successfully and get these good outcomes, get their business and their stakeholders behind them. So please feel free to reach out if you have questions, if there are opportunities to collaborate. Always open for these sorts of things.

Speaker B: Great. And Thais, thank you so much for being on the show today and congratulations on your new book that's been published. And thank you for all this contribution that you give and through books, podcasts, video shows, etc. And for sharing your knowledge on here of course. And uh, yeah, all the best with that and your new business venture.

Speaker C: Thank you so much for the opportunity. It was great being on the show.

Speaker A: Hi dear listener, Just a quick note from me before you go. If you enjoyed this show, then please don't forget to subscribe to future episodes

Speaker B: via your favorite podcast app. I have loads more great stuff coming your way.

Speaker A: Also, I'd love some feedback from you on this show. So please, please leave a review on

Speaker B: Apple Podcasts, uh, itunes, Spotify or wherever you listen to podcasts.

Speaker A: Thanks for listening and catch you.

Speaker C: Sam.