IBM’s UK&I general manager: ‘Don’t treat AI like a science experiment’

Speaker A: My name is Leon Butler and my leadership lesson is around clarity, simplicity and consistency. Be clear in communication what you ask and what you set out, whether that's strategy, goals or targets. Make sure it's understandable and don't change priorities every five minutes.

Speaker B: Hello and welcome to Management Today's Leadership Lessons podcast. On this week's episode, MT's former staff writer Ailish Cronin talks to Leon Butler, General Manager of IBM UK and Ireland. As a former IBM vice president responsible for AI sales globally, Butler has a clear viewpoint on what businesses need to do to get the most out of the technology. When it comes to those at the top of the organisation, this comes down to the need to lean in more and actually use the technology, as this enables you to not only talk about it, but really understand it and in turn effectively implement it within your organization at the same time. Where initiatives often fall short is where they are treated like a science experiment as opposed to rooted and clearly defined objectives. Butler is an IBM veteran in the truest sense of the word, having joined as a grad and worked his way up the ranks before clinching the top UK job a year and a half ago. He talks about the leadership lessons he has acquired over this time, not least the need as a leader to really understand your business. He says, I'm in the technology industry and if you don't understand technology, how on earth will you run a company that rolls it out? This understanding underpins everything from the quality of your communication with your teams to their ability to convey the company's offer to ultimately the strength of your clients understanding of your solutions. The episode also covers how AI is changing the attributes valued in leaders, the challenge and opportunity of upskilling and the importance of psychological safety in harnessing AI's gains.

Speaker C: So Leon, you first joined IBM in the year 2000 as a graduate before leaving 18 years later and you went to work for Oracle for about four years. That's right. You then came back to IBM in 2023 to run its global AI and Data division and then in 2025 you were promoted to General Manager of UK and Ireland. So looking back to the year 2000, what first drew you to join IBM?

Speaker A: Thanks, Ailish. And look, thanks for having me on and appreciate it and you've absolutely got the history right, so thank you on that one. Without going through a load of history, I was very proud to join IBM. I didn't have most probably a, uh, traditional entry into a large corporate. Actually just taking a step back, I. And if you can just humor me, For a couple of minutes we as a family we lived outside of the UK for a little bit, came back and unfortunately for reasons we're going to, we kind of came back with nothing. And um, what that allowed me and the family to do was ah, arrive in places like bedsits. We were lucky enough to get a council house and I got the opportunity to really push myself incredibly hard. I came from a loving family who provided but I certainly learned what a work ethic was and uh, worked my way through my GCSE A level, managed to get my degree in the end and certainly MBA later down the line and it really, it really got me thinking about what I wanted to do and I really enjoyed it. I did a computer science uh, A level and IBM was very much on my radar and I applied for the graduate scheme and managed to get it and it was one of the most proudest days I had in my career. Joining company now that's over 115 years old, uh, that has been iconic for that period of time and um, yeah I was very proud to do it with it came, as I said, a work ethic. It also taught me a few other things as well as I went through. I'll be very open. When I first joined I pretended to be someone I wasn't. I learned uh, that actually authenticity is really clear. I didn't come from public school. I learned to actually become myself, be authentic, understand that people have very different backgrounds and actually you need to treat people in very different ways depending on their backgrounds. And ultimately you know, if you look at things actually things like education, things like uh, learning are real gift. And from that I took full advantage of joining IBM and really all of their training programs and the wealth of knowledge that uh, and training that IBM gave me and that allowed me to kind of progress through uh, IBM and you know I went from really technical to sales roles very quickly and worked my way into different roles. Lots of new business aspects of it, to running, getting my first management teams, to uh, leading kind of the software parts of the organization and progressing from there. I won't take you through the entire history of it, but safe to say I think I took advantage of everything that was in front of me and worked incredibly hard but enjoyed learning all the way through.

Speaker C: And what brought you back um, to IBM in 2023 because you put 18 years in um, and then you left to work for Oracle for four years and then you came back uh, to run the global AI and data division. So what drew you back at that time?

Speaker A: A couple of Things IBM? Well one, I absolutely love my time in IBM and I can tell you I have a strong integrity compass and IBM is built on trust and transparency. And um, as I said it gave me a lot, uh, the role, having the privilege to take the worldwide vice president role of data and AI in such an evolving, changing time as AI was just before it exploded, uh, was a huge attraction for me and I believed in our journey, uh, that we were going on under uh, Arvind and Rob Thomas's kind of leadership certainly here in IBM. And I knew I wanted to be part of that particular journey and be part of it. And certainly from an AI perspective, I think a number of us could see what was coming. It was before everything really, really did take off. But you could see, and I certainly could, uh, absolutely, opportunities that are in front of us on that one as well. And it has just gone from, it's just accelerated quickly and quickly and quickly. And IBM has certainly given me the uh, the ability to lean right in and be at the forefront of a number of those technology developments.

Speaker C: Mhm. And now as the UK and Ireland General manager, when, when you were brought into that role, were there any leadership lessons or something that you learned from running the AI and data division that you've brought into your new role? Have they influenced your leadership style in any way?

Speaker A: Yeah, it's um, the interesting thing about that role is um, first of all, pace, you know, things were moving and have moved at an incredible, incredible pace. I mean it's the most exciting I think times in the technology industry and I think it's about the fastest moving industry out there actually on that one as well. So being able to embrace that level of change in pace and react to it I think is something uh, that I learned and embraced what I also discovered and learned and it's been always part of my career. But you need to understand your business and it sounds an obvious statement, but if you don't understand I'm in the technology industry and if you don't understand technology, how on earth will you run a company that rolls out technology? You need to understand it not only for your teams, to better communicate to your teams and make sure they can articulate it to clients, but also that your clients understand it. So understanding what we sell as a business, our solutions, and how that actually solves client problems is really interesting, but it's essential. And if you can't communicate it with clarity, you know, you can't expect your teams to do it or your clients to understand it as well. So really Delving into that understanding of technology is something I learned, I've always learned it. But obviously having that particular role, having some of the World's uh, best SMEs uh, under me at the time and still do, made um, sure that I uh, really did embrace and learn and be curious about technology.

Speaker C: And I was reading an interview that you'd done previously where you said that you're steering IBM's UK and Ireland operations towards measurable productivity gains and practical applications. Could you perhaps share an example or some examples of this in action?

Speaker A: This really gets into the world of AI and AI transformation and when we talk around productivity gains, um, talk is really cheap. I think, uh, at the moment what I saw in that worldwide role I had, and maybe you heard me say this before, is I saw a number of pilots that didn't go anywhere kind of science experiments to really delve into uh, AI means that you need to understand not only how you apply it, but actually how it gets into the real core of the business and its processes. When you do that you start getting proper productivity gains. And I'm not talking about basic AI when it comes to extraction documents or summarization of minutes as an example. This is looking at core processes, core business functions around things like hr, procurement, supply chain, invoice to cash seller productivity, uh, which really are processes that spans across the business and across multiple applications. When you get that right, and certainly we do within IBM, do it to ourselves first of all you start seeing some great results. And one of the examples I do give is uh, our uh, agentic solution around Ask hr, uh, where I can ask queries through that particular HR agents and it will answer those particular queries as well. So 94% of our queries, queries are really answered through the Ask HR agent. I had a situation not so long back and maybe it's an example of using technology which I'm a big believer in, of being able to do a talent management day where I was able to through one command send an email out to my team, say look, you know, we want to look at the talents across the business. I'm a big believer in developing talents. You know, we're able to get uh, each, each of my direct reports were able to get the top talent from their business or their talent across the business, look at their historical performance, input into that, pull it back together, collated into one uh, particular viewpoint that we as a team could come together and actually view to be able to do that. That would have taken huge amounts of spreadsheets, people Meetings. All of that was done through just a set of simple instructions as well. And um, that becomes, makes me and the team a lot more productive and, and ultimately anything you can quantify and measure from a process perspective and you can accelerate that is really where you get the productivity gain. I mean IBM globally has got four and a half billion dollars of productivity gains as we've implemented internally, but it really is using it for yourself. And one of the big things as a leader I think is making sure that you can actually and you should utilize that technologies available.

Speaker C: Yeah.

Speaker A: Hm.

Speaker C: Yeah. As you said, a lot of companies do get stuck at that sort of pilot stage of AI deployment. What systems have you put in place at IBM to help move that process along at pace?

Speaker A: Yeah, it's really bedded into it's across the business. So it's less about just general functionality and um, more about, as I said, a business function. Um, on this one as well. I will tell you there's a big cultural shift if you look at kind of embracing these new technologies. The first part of it is using IT and having the confidence to use it. And I talk a lot about this with clients. We not only have put those systems in place and we encourage people to use it, we make sure there's psychological safety, uh, to make sure that people can experiment if it goes wrong, that's okay, experiment and absolutely test yourself on those ones as well. We have globally what's called a Watson X Challenge. What's the next is our AI portfolio. And um, across the entire business, uh, we will get together across the world and uh, bring problems, business problems. How we can enhance the business together using our technology. And it is fun. Everyone gets together different locations, works together and we use WatsonX to basically look to solve problems. And then what we do is we get the best problems, the best challenges and we vote on that one as well and then we implement them as well. So we really do encourage to use it, but do it in our day to day kind of activities of it as well. And myself, uh, as a leader, you know, I do embrace it and make sure that I do use the tools that are available for me. But IBM corporately allows me to kind of embrace it and the rest of the teams by putting things like that out. Um, the challenges are great fun, we celebrate it, the rewards and then the winners get to see those solutions implemented. It's pretty cool.

Speaker C: M I was watching um, an interview that you did uh, on YouTube and you were talking about how at IBM you're focusing on smaller and more Efficient AI tools. What does that look like in real time?

Speaker A: Alish, you may have heard me talk around the analogy of the Swiss army knife, which is you've got these huge trillion parameter models, take huge amounts of GPUs. And actually what I see and what we see are uh, more smaller agile domain specific models that are fit for purpose for a particular set of tasks. You don't need all of that, that huge amount of processing power to do a lot of the tasks that we do. The benefit of having those smaller agile models is that they are quicker, uh, they have more accurate information, they're trained on specific information. One of the biggest inhibitors on AI and AI use cases actually clean data and access to data, data is siloed. It really can be really hard to get to. So it's all very well using an AI model, but my goodness me, you need to make sure you get an accurate answer from it as well. And training those models on clean accurate data is really important. And ultimately it takes a lot less power and GPUs to be able to run those. And uh, at the moment there's a huge amount of racing out there to build bigger and bigger, uh, kind of data centers and more energy. And actually this deals with a lot of that as well. So yeah, hopefully that explains that.

Speaker C: What metrics are you using to measure the outcomes of these AI projects?

Speaker A: Yeah, I mean they're industry standard when it comes to kind of rate and pace of those particular models. Obviously from a speed perspective it's going to be accuracy and it's going to be speed. As to the accuracy, the answer on it, you know, I think it's got to be the governance around it and I think some of these things are pretty obvious. You know, I'm sure there are a lot of listeners out there that have gone and put a, uh, question into whatever AI tool that they've got on it and intuitively they know that's not quite the right answer. It could be anything from a complex uh, prompt to a very simple one around a kind of a question about, I, uh, don't know, their favorite hobby. You need accurate information on it as well. And that is around accuracy. And it really kind of then gets us into the world of governance, which is if you are going to have things around models and AI, actually how do you embrace and actually embed governance across certainly from an IT perspective, business perspective, your infrastructure, your applications, your business. So one, uh, you know that those models are clean, but separately you are getting the right information from the different parts of your business. Thirdly, that you're getting the right answer and if the answers aren't right that you get alerted to it. In the AI world you hear about drift and bias and the worst thing you do is have models that kind of give you the wrong answers and you know, from a hiring decision to uh, a loan decision or anything else on that one and there's no alert to it. So having that governance and human oversight is going to be really important as we go forwards. And you know that that's very much around four pillars that we talk around really having uh, kind of a control plane to deal with those agents. There are lots of agents, everyone's got a definition for agents at the moment as well. And one of the risks is going to be agentic sprawl agents everywhere. So we have a control plane that sits above to not only control but govern those particular agents. I ah, talked around the data aspect of IT as well. Ah, there is the kind of the governance aspect and the sovereign kind of aspect, the hybrid aspect of it as well. Making sure that actually the information whatever environment you want, whether it's a private or public cloud on premise or anything else in that one as well, you can not only run those different environments accurately but you can actually control it and make sure it's sufficient and fit for purpose as well. We're in a world where it is complex. We think AI will drive most probably a billion applications in the next few years and that's going to be across multiple platforms. And uh, with that comes complexity. Ah, having the ability to be able to harness that not only just from a data perspective but be able to run it on multiple environments. What we call hybrid uh, is going to be hugely, hugely important. Not everything is going to go in a cloud, not everything is going to go in a private on premise Estate clients will be running their AI applications of businesses across multiple, multiple environments.

Speaker C: So how do you decide which AI projects to prioritize?

Speaker A: Yeah, it is around those gains eilish on that one. Anyway, so I suppose I said earlier on about at the moment you do see people kind of playing around with just basic um, kind of aspects around summarizing minutes as an example. But if you're gonna, if you're gonna look at things maybe like customer service as an example, something that gets you really good efficiencies but also from a reputational perspective. Enhanced reputation could be around call agents help, it could be call deflection, could be assisting call takers. All of those things will help you be not only better at your particular role, but more efficient. And I suppose it comes down to what's important for the business. Uh, on that one, a company that does have a lot of customer facing, call center type individuals. On that one as well, that may well be more important. You may have companies on that one as well that has multiple chaotic and archaic kind of HR systems. You know, where someone's struggling to um, raise a P45 as an example. Ah, because it's in different systems. I think it comes down to the business need and what the business wants on that one as well. And that could be through productivity gains or it could be through revenue generation. I'll also say as well that when you kind of go on these sort of transformations, uh, the people that were doing the tasks are incredibly good at actually helping you redesign those particular processes. And one of my examples is around hr that before we kind of cut over to our Ask HR system and you see it with clients as well. It's what I used to call how do I. Or how to calls. You would phone up a help desk and say, how do I do this, how do I do that? And a lot of times managers saying how do I move an employee? Or how do I raise a kind of a reference, uh, letter for my employee? Those take many, many touch points. Many applications, I think, uh, in our systems previously it was almost 18 touch points. That burns a huge amount of time. And people rattling around back end systems doing that is not a good use of time. And it takes them away from actually their core day job, being able to free them up, to actually be able to just use one command and say, I actually want to give a reference letter, uh, which kind of works out which systems to go to, produces a letter and that person can send it out, uh, is a huge gain on that one as well. So it really is looking at the individual business and what's important to them, where have they got the biggest challenges or where do they want to get the biggest gains?

Speaker C: Now you've come into this leadership role with an already pretty impressive AI background or technology experience. But there will be many CEOs um, listening to this episode who don't have that same experience. So what crucial skills will they need in order to successfully lead their organizations and their teams through this sort of new AI landscape?

Speaker A: The first thing is to embrace the technology. Embrace it, use it. There are too many people talking about it at the moment and not enough people, um, using the technology or even experimenting with it. And that could be as simple as using whatever AI models out there for Productivity gains at home. It is getting familiar with it as well. We are in a fast changing environment and part of it is then embracing it, making sure actually if for whatever reason people aren't familiar that there is the right training in place. And maybe we could get onto the whole skills piece around this as well, because I'm a big believer in this and certainly making sure from a UK perspective, um, we do have the right skills, but it's no longer about who's got the best memory or command and control. These tools are out here now to actually empower people to harness the best information across the business and make decisions really quick, quickly. And you know, once you get your head around the fact those tools are available, it's then having the courage and also, uh, the right skills and training to be able to use those, uh, to get the best kind of results from it. And you know, that really represents kind of a shift on kind of tweaking with the business to actually looking at how the business operates on that one as well. I mean, some of the things I talked about, about actually getting companies together and certainly employees and companies to kind of solve problems with certain tools I think is a great start. And certainly our what's next challenge does that as well.

Speaker C: So I want to go back and talk a little bit more about governance because about an hour or two before our call today, I got a copy of a new study from the IBM Institute for Business Value. And it had a couple of interesting stats in that it said that the accelerating pace of AI is pushing CEOs to redesign how C suite roles are structured. For example, 76% of organizations now have a chief AI officer, whereas in 2025 only 26% had a chief AI officer. So I was wondering, since coming into the role of general manager, has the C suite changed or is it currently changing to accommodate this new kind of AI landscape?

Speaker A: Governance is really, really important right now. I mean, AI has to be governed responsibly to earn trust, whether it's leaders, businesses, governments. All of us play a critical part in it as well. In fact, we work with governments, uh, worldwide to really advance smart, um, regulation, smart AI regulation that protects people. And that could be on focusing on the riskiest real world, real world kind of cases around AI, but not necessarily the underlying algorithms as well. You know, the analogy we always give is that, you know, we don't regulate the wheel, but it's rather its use on kind of things like cars and trains and planes. There's an analogy on that one as well. And it's the same approach for AI. I mean I mentioned earlier on if you don't have governance in place, you do have the risk of you know, making some pretty poor decisions or at least having some form of human oversight over those particular decisions. So you can actually make sure that if something is going wrong that you've got some human oversight on it. And uh, you know, I mentioned around uh, governance and what sort x governance is. Governance is our ah, particular part there. Growing up in a regulated industries that really does look at all of AI regulations across the world and allows our clients as they kind of embed AI to make sure that actually decisions are taken in the right way, that prompts are not manipulated, ah, that the information comes out and any kind of guardrails or params you to put in place, you're alerted to it as well. I will say as we kind of get deeper and deeper into AI, governance has to be across the entire business. You've obviously got kind of corporate and risk governance, but from an IT system in particular, you need to make sure that governance is all the way through. I mentioned agents as an example. We're seeing a plethora, every application's got an agent. But actually how do you make sure that you govern those particular agents and outside of IBM's capabilities? And I talked about that control plane that is able to kind of look and actually orchestrate across those different applications. How do you make sure those agents are governed and made the right decisions but also how do you build that governance all the way down into your application layer and your infrastructure layer as well? And you know, we start getting into a world of sovereignty to make sure that actually the right information is accessed by the right people and stays in the right place. And we'll see governance be more and more important as we go forwards because ultimately yes, it's about making the right decision, but it's also around making sure that data, citizen data remains in the right place and protected.

Speaker C: The study also found that 85% of respondents say all functional leaders must become technology experts in their domain. And between 2026 and 2028 respondents expect 29% of employees to require reskilling for a different role and 53% to need upskilling to perform their current role more effectively. So how are you and the wider IBM senior leadership team m approaching reskilling?

Speaker A: Thanks, thanks. You've taken all my stats. It was a good thing and I really appreciate you kind of looking at that before beforehand as well. That is great and has made my day and that skills are really important. I think I'm fairly visible um, around AI skills and I'm a big champion of it as we go through. And if I just take a step back at UK and Ireland we are at a critical turning point because the barrier isn't really the technology, it is having a workforce that's equipped with the skills to use, as you just said, use that particular technology and use it in the right way. As I said, not just playing with it, but actually being able to apply it to real business challenges and processes. And I think in the UK and Ireland to really unlock the potential of AI, uh, we do need to invest in people and the skills around it. I mean the UK government's AI Opportunities Action Plan, it's hugely exciting and it says that you can grow the economy by an additional 400 billion by 2030, which is really, really exciting. But to be able to do that we need to make sure that people are skilled and take advantage of it as well. We are one of the biggest AI economies in the world and with that we need to make sure that everyone has access to that. And um, we're pretty proud of the fact that uh, we have worked with the UK government on their recent announcement, it was announcement last year around training 10 million UK workers, um, around AI skills actually. And we were one of the uh, founding partners around this particular initiative, um, which is really around preparing the UK workforce uh, for AI jobs in the future. And it's called the AI Skills Hub. And uh, on the AI Skills Hub we have a uh, program called Skills Build. And there are particular learning paths and learning journeys that people can go on and access those to actually understand on that journey and how to use AI and how to use them for use cases. We talked around pilots earlier on and one of the most important things is making sure when you're using AI you've got an outcome in mind. So it's not, again it's not the science experiment. You're actually doing something to drive an outcome because that's what people want. They don't want to play with products. It's about getting to an outcome. And these learning journeys help you actually to go on that journey there. And I was incredibly proud to be one of those founding partners on it. And part of uh, Prime Minister's announcements as we go through, and I will tell you the Skills Hub is there as a centralized part that allows people to access it. Because actually I want AI skills to be across the entire country, not just London based or anyone else, but anyone who fancies training, reskilling, uh, whichever part of UK and Ireland you're in, uh, you're able to do that. And to be able to go to that centralized part or an AI skills hub and then utilize, you know, whether it's our training or anyone else's, I think is really important. You know, globally we've committed to skilling 30 million people by 2030. But if I just put my lens on UK and Ireland and working with the government, I think the government's made a great step in the right direction and we're very proud to be part of it.

Speaker C: M Job security, I think is a huge concern for many employees. And I think the influx of AI technologies has perhaps increased some of that fear. And you know, looking again at the study, it said by 2023, CEOs expect 48% of operational decisions where consistency and guardrails can be codified will be made by AI without human intervention, which is compared to 25% today. So with this in mind, how do you as a leader ensure that those jobs are secure?

Speaker A: Thank you. I think the first thing is making sure that we do embrace AI as well. I think the danger is anyone who doesn't embrace AI, um, is going to be left behind. I mean, there is no doubt about that. So embracing AI and using it I think is important because ultimately, yes, it helps absolutely deal with efficiencies and repetitive tasks, but it also allows people to get to higher value tasks as well, which I think is really important and actually is more exciting. I think the value increasingly sits in human capabilities around things like judgment, creativity, problem solving and taking things like ownership as well. You know, going back to the HR example I had on that one as well as we made things more efficient, actually what we had a number of our HR professionals become data scientists and actually being able to use HR language and speech to actually enhance and design those systems. And we had situations where actually their levels increased on that one as well. So, you know, it will drive efficiency. If you embrace it, it opens up a huge amount of opportunities. And I think what we'll see is a change of dynamic around the shifting roles, but really bringing people to kind of that higher value aspect of it as well. And uh, I do think that becomes a lot more enjoyable as part of the roles, uh, versus kind of repetitious tasks.

Speaker C: So if we look ahead, what should business leaders be doing now to prepare for the next wave of AI innovation?

Speaker A: I think as a leader it is embracing that technology. It is using it for the right reasons, whether it's you know, as I said, I think I said earlier on, the technology is there for me to be able to make some decisions very quickly. Quickly, but data, uh, bad decisions, uh, on that one as well. But we do need as leaders to lean in a little bit more and actually one, as I said, use those technologies so we can actually not only just talk about it, but actually understand it and demonstrate it. Because when you understand it, you can really implement it. I'm a big believer in that. I think the separate piece, as I said, is that I think our leaders need to look at the bigger challenges they're facing across the business and pick those and then say, right, that's how I'm going to tackle it. And it doesn't have to be, hey, wholesale change across the entire company, but again, we talked around those verticals, around things like hr, procurement, supply chain, and actually tackling one of those challenges that maybe a CEO or someone knows is not going well, that does have a number of complaints or feels it's not efficient. Actually, those are the things to lean into as well. Because as soon as we absolutely deal with those processes and that underlying business, then things change. You start getting to a point where not only you more efficient, not only do you get more information faster and more accurate, but actually you can start redesigning your business and actually then getting people into the right roles and right mindset to make some really good decisions. And I think that's the difference between again, playing with AI and actually implementing it where it is actual true business change at the heart of the business that addresses a business problem and makes ultimately that business successful through either productivity gains or revenue gains.

Speaker B: Thanks for listening to Management Today's Leadership Lessons podcast. If you enjoyed it, please subscribe. You can find us on Apple, Spotify or wherever you get your podcasts.