Andy Baldwin: AI Is a Contact Sport: Leading Through the AI Shift (IBM Consulting)

Garrick JonesHello and welcome to this episode of the Curious Advantage podcast. My name is Garrick Jones, and today I'm here with my co author Paul Ashcroft.

Paul AshcroftHi there.

Garrick JonesToday we're delighted to be joined by Andy Baldwin, Senior Vice President of Offerings and Growth and Chair of the IBM UK Board. Andy leads IBM Consulting's asset-based consulting transformation, helping drive the company's AI-enabled offerings and scale its go-to-market engine. With more than 25 years of consulting experience across banking, insurance, wealth management, and technology, he's advised clients on transformation, innovation, and regulatory change throughout the UK and Europe. Before joining IBM, Andy spent more than two decades with EY, most recently serving as global managing partner. He led operations and strategy across 154 countries. He's also held leadership roles at Capgemini and worked extensively with governments, regulators, and major global organizations. Today, Andy sits at the intersection of leadership, AI, technology, transformation, and the future of work. And today we have him with us. Welcome to the Curious Advantage Podcast, Andy.

Andy BaldwinGreat, Garrick, Paul, good to see you and good to see you guys again after so many years. We obviously worked together in our when we were a lot younger and I was a lot less grey. Although I don't think I don't think Paul I don't think Paul's changed since then.

Paul AshcroftYeah, there's some very good filters going on over this side.

Garrick JonesPaul's the only one who looks younger than all of us when we started. And it let's start where we always do, which is how do you personally define curiosity? And what role has it played in your life and work over the years, do you think?

Andy BaldwinI think of curiosity as really a desire, a need, a focus on effectively constant learning, lifelong learning. You know, for me, both you know academically, I you know I did a number of different courses, college professional exams. I've always had an interest from that on the academic side, and then obviously in my career, I've done lots of different roles and different jobs, and often it's been driven by desire to do something different. And so I think that's really what's motivated me. So for me it's this point about constantly looking to learn new things and then have new experiences.

Paul AshcroftAndy, I guess at the moment, as we all are, you have the lovely opportunity to learn a lot of new things. You're leading a global consulting business at a time when both the consulting industry and the wider technology landscape are undergoing profound transformation. I think probably fair to say even greater than we've seen in the past. From where you sit, what's really changing?

Andy BaldwinYeah, well look, Paul, I think, you know, when we obviously with experience and you know, all three of us have worked and advised clients for many years, we've seen a number of different technology shifts and changes. you know, c come twenty-five years ago, I was working a lot of internet startups. So we've seen the internet wave come, we've seen blockchain wave come, and we've seen other technologies. I think the difference with AI is that I think the scale and the size and the impact. It is more transformational and it's more ubiquitous in the way it's affecting lots of things. So if I if I look at it in the way I'm looking at it in my new role at IBM, and I've been there a year, you know, we're seeing a lot of clients go through major transformation in terms of how they're thinking about AI and how AI is going to change them or disrupt their business. But the technology has a far broader application, you know, across the enterprise than I think many people have seen. At the same time, you know, in professional services, the consulting industry is going through a major change because obviously there's a lot of clients who need help and advice around how to how to think through the impact of the new technology. But also the businesses, the consulting businesses, the professional services business are also going through a major transformation of how they deliver value to clients, how they deliver services. To clients with this idea or this notion of a blend of digital and human workers coming together. And then the final area, which I think is very different, is you know, consulting for the last 30 years has been on an increasing convergence with symbiotic relationship with technology. And if you look at the changes that have taken place in the technology industry, I'm thinking here about the traditional ISVs, SAP, Oracle. Cloud providers, AWS, Microsoft, you know, Google, with the emergence now of these frontier models and AI companies now with the valuations are you know in the in the trillions for companies that only existed five or six years ago. Those three things together: impact on clients, impact on the industry, impact on the way in which you serve clients, and then the impact on the technology itself, all of those things together, I think, are having quite a profound effect, much more so than I think some of the previous technology waves that we've experienced before.

Garrick JonesAndy, there's so much in there that I'd love to unpack. The one that really stuck with me, and we I'm sure we'll get to all of them, is where you said in passing that AI tech has a broader application than just the obvious one that we're seeing at the moment, although you said it was ubiquitous. Tell me more a little bit about what that broader application is in your mind.

Andy BaldwinYeah, well look, I think if I if I take, I mean, our own experiences from an IBM point of view, you know, I think there's a lot of firms talking about being client zero and how they're applying technology and how they're applying AI. If I look at the journey we've been on, it preceded my arrival, I should I should stress, over the last three or four years, it's really been around you know, deploying AI. Obviously in the very early stages it was machine learning AI. and more latterly it's been the agentic AI that we more associate with OpenAI and Anthropic, etc. But actually that, you know, that implication has been, you know, organizational-wide. So we've applied it into the what I say, the finance functions, you know, the finance, the HR, the procurement, the operational functions. We've applied it into individual employee productivity. We've applied that into the way we're thinking about service delivery. We've applied it into how we're building our new software products in the software side of IBM. And increasingly we're thinking about it in terms of how we go to market. So for example, you know, IBM is always associated with serving what I would call the Forbes 2000, the largest enterprises in the world. Actually, a big part of IBM's software strategy now is to go much more into what we call the select area, which are companies of a billion, half a billion plus turnover. But what we would probably call you know mid-market, mid-corporate, you can only really serve that market through a digital and an AI-based market approach and solution. And so if I look at my own company, the way in which AI is touching every facet of the organization, both how we run, how we manage, but also increasingly how we engage with our clients and sell, particularly in that middle market area. And I think when you look at know the work that we're doing as a company, you know, along with Accenture, we're one of the few companies that disclose our generative AI bookings. So if you take the data for last year, because we were going through our close at the moment, you know, last year we booked close to I think it was nine, ten billion of AI related sales as a as a as a consulting business. And again that, you know Me, that covered all industries. It covered a lot of the same functional areas and a big focus around productivity, but also around how you get the data ready to take advantage of the technology. So I draw that as a contrast, Garrick, to the work I used to do when we're, you in the internet startup period. A lot of it was new business models. It wasn't about actually, I would say, fundamentally rethinking how you run and manage and operate your existing. business and that's why I say for me this technology has more applications and more potential in how it's deployed both by employees, how it's deployed serving the clients, and also how it's deployed across the enterprise. I mean that's quite different to some of the other technologies that I think we've seen, certainly in my you know thirty odd year consulting career, where I think there was some very clear use case applications, but they weren't as widespread and as broad based as the ones that we see now emerging. And some of them are still emerging, but I think some of them are far more, you know, established, proven. And we're we're we can see the return on the investment in the technology in those cases.

Paul AshcroftI think we'd fully agree. I we're seeing the same that many organizations are all in on AI, but all in terms of trying multiple experiments at once, which is quite a large bet on your organization about how are you shifting it? How are you bringing in new technology and new ways of working? What's your experience of that? And how are you helping? companies start to see value from this? And is this connected to the show and tell culture that I know is really important in the way IBM works around AI?

Andy BaldwinYeah, look, I'd say, you know, if I look at the last two or three years around the general, and obviously it was doing some of this similar work when I was at EY before, I'd say many clients have been on this sort of journey of do a number of proof of concepts, test the technology, test the deployment of it, see how employees used it, f think about the change management. I'd say over the last 12 months, I think we've seen much more of an inflection point, Paul, where I think people are now going to scale deployment. And I think one of the proof points for that is the reason everyone's talking about token economics, the cost of tokens, token maxing as they're calling it, is because I think when you were doing proof of concepts and small scale deployments, the actual data you're using, the compute power you're using was much smaller. As you go to broader-based enterprise-wide deployment, whether that's at functional level or across the whole enterprise, you are seeing an increase in the compute piece. And I think that reflects the shift to scale. But also, I think it also reflects the maturity of clients. Because I use the analogy, you know, there's many clients using a Ferrari to go and get the morning paper or a bottle of milk. You know, they're using frontier models to do. decision support tasks or activities that actually there are simpler, cheaper solutions to use. And that's why I think you're going to go through this next phase of you know, calibrating and balancing the models, the agents to be the right agent for the right task, using the right data with the right cost of inference. And I think that will actually improve, you know, the economics or the costs of that. In terms of to your point around where clients are I think clients have moved beyond the experimentation phase. I think they're now looking at scale. And to do that, they're realizing you've got to manage a number of things. One, you've got to really think carefully about governance around the agents, because there's one thing having what I would call productivity agents for the employees, where employees build their own agents, as I'm sure you guys have done. with you know copilot or whatever you whatever tool you're using whether it's Claude Code or it's one of the products but also enterprises are looking at you know and agents to be deployed across the enterprise you know our research showed that in the average company is going to be deploying something like by the end of this year up to fourteen hundred agents by the end of this year at the enterprise level. So but to do that you need the governance, you need the telemetry, you need to know who's using them What are the agents being used for? And I think as you start to mature that, you can start to see the telemetry, the token costs, you can start to manage those costs most appropriately, you can start to choose the right model. And I think that's the journey, you know, I think clients are starting to go on now. And I think that's really gonna help. When you look at the use cases and some of the areas of application, you know, there's clearly already proven models around. know, software development lifecycle. You know, I see what we're doing, you know, we're obviously you can imagine we've got a very large software business, a very l a large number of engineers. We can see already the data, the improvement in productivity, the improvement in the performance of our engineers in the way they're using that technology. Many clients are seeing the same, and that's a very, you know, I'd say that text but it's very proven use case. We're seeing other proven use cases around you know operations call center productivity, efficiency, et cetera. So I think as clients get more comfortable with the technology, they develop a better harness for it, they understand their own data better and also how to get the right data for the right agent, you know, at the right, you know, price point from a model point of view, I think you're gonna see more benefits of that. To your to your point about the you know the show and tell point, yeah, well I'd say it's one of the big differences Having spent a large part of my career in a consulting firm, which I'd say is one of the best advisory consulting firms in the world, having gone to an organization which owns a very large consulting business. I think IBM Consulting is probably top four, top five in the world. But it's a software company. IBM is a software company, software hardware company. And in the software and hardware company, it's all about. Proof of concept, it's all about minimum viable product, all that it's all about showing the client the product. And actually, we try to take that approach in consulting. So whenever we talk about things with the client, we quickly try and show them what it is. We show them the agents or we show them the platforms, we show them. And that's that's I think quite a big difference from where I work before into where compared to where I work now. But I guess it reflects, you know, IBM as a software company, technology company. It's very focused on implementation, execution, and then running things on behalf of clients. but we try to engender that you know, show, not tell culture and approach across our 160,000 you know, consultants. And it feels different to where it was before, but you know, like any organization, the change management takes a lot of time and you've got to build it and embed it into the culture as well.

Garrick JonesIt's amazing, Andy. Every time there's an inflection point in the economy, you know, we get extremely excited about all the opportunities and the new worlds that we can imagine. And then followed up very closely by the realities of what that means to the you know our lives and our lived lives. And I l I love the idea about the show don't tell. And how everything's becoming so much more practical in some ways, like, you know, learn by doing and s it's become becoming the norm. But the thing about the economics I'd like to just go back to a little bit, because we have to not only you t let's say coding, for example, you talked about software development, that's seen a massive shift. You've got all the agents working on behalf of coding extremely quickly, efficiently and very accurately these days in hermetically sealed situations. And the but what's happening on the economic side? Because if there's tokenization and token economics as you talk about, does it mean that the cost, you know, you get this increasing efficiency, and people imagine there's a you know, you take out some of the development costs, but are you simply passing costs over taking, you know, c passing the costs and the fees to the anthropics of the world and so on? or is it jet do you is there a genuine sort of better economy that emerges as a result.

Andy BaldwinYeah, look, it's a great question, Garrick. I think you gotta look at I think the evolution of professional services over the last, let's say, twenty, thirty years. You know, twenty, thirty years ago, the industry was a P times Q, price times quality, rate per hour, and people bought hours that they bought people, and you may have made that fixed price. You know, I think over the last 10 years there's been an increasing focus around what I would call assetization of, you know, how you take some IP intellectual property. And you try to charge for that alongside charging for the human labor. That assetization, I think, has been turbocharged through the emergence of the frontier models and AI. So, you know, when I think of our business, the business I look I'm involved in, we talk about human and digital labor together. And you know, we talk about the agents and deploying the agents, and we talk about pricing the agents into the bill of materials, into the work for the client. And that can be agents which are standalone, fully autonomous, or it can be a forward deployed engineer supported by two or three agents. And we're we're obviously going to try and charge for those agents because it makes the individual that much more productive. Now, to be track to be very open, that that is a change management challenge on a number of dimensions. That's a change management challenge because you have to be able to explain to your own partners, your own staff, that the agents and the assets. Have an intrinsic value in their self. You also have to explain to the client that those assets and those assets and agents allow you to do things more rapidly, but they come with a cost. Obviously, you're committing and investing IP. You've obviously got the compute costs that come to that. And at the same time, I think reflecting this shift towards your human and digital, I think we're seeing a lot more clients asking for what would call outcome-based pricing. By that I mean if you look at one of the shifts that's happened as a result of the emergence of AI and the frontier models, the buyers of consulting services, specifically those that are more technology centric, has shifted dramatically towards business buyers. Yes, the CIO and the CTO are still involved in the process, but actually the line of business owner, the CRO, the CFO, the COO, the divisional leader, actually is having a much stronger input. on the role of the AI, where they want the AI deployed, because you need that domain and industry expertise to get the value out of the technology. And they'll often talk in terms of their language. So if they're a CFO, it's how are you going to reduce the cost of my finance function through this work? And that shift, you know, towards a more of an outcome based, I think, is being accompanied by this, you have the you have the normal labor market model and you have these digital assets and labor. And how you combine those and how you price for those, because in essence, you are sharing that value with the client. You are going to be sharing that value with the client in some shape or form. Because if a project costs $100 and you do it for $75, the client's going to expect some of that saving. You're going to want to preserve some of that efficiency towards the cost of the compute, the cost of the investments, the cost of the intellectual property. And for different parts of the consulting model, that dynamic is different. You know, so if you look at the operate business, the run part, it's more profound. If you look at the advisory business, you know, that a traditional big four would do or a strat firm would do, you're getting productivity gains, but not at the same extent as you would do if you were in a software business or in a run business. But I think that's a critical part at the moment of the I would call the new pricing paradigm is how do you price this? How do you show the value? And how do you get the client comfortable that they're getting value from what they're paying, they're getting an outcome they can relate to and understand? But at the same time, you know, you're you're evolving your own delivery economics as part of as part of doing that. That's why I said at the very start to Paul's question, the transformation that we're seeing through generative AI is as profound on the consulting firms themselves as it is for their clients.

Paul AshcroftI wanted to click into that, Andy, because you mentioned a few times the sort change management of it. I'm really keen to understand if you can share what are you guys doing at IBM, but also maybe what are you encouraging clients to do in terms of managing that change with the people? And you've now got teams of agents working alongside humans. It's an adaptation to how they work. It's an adaptation to what they do. But as you say, it's a... is a change management. How do you lead your people through that and how are you leading clients through that change?

Andy BaldwinYeah, look it's a great it's a great question. So, you know, the some of the things we're thinking about it is one, we're expecting everybody, and that means senior leaders like me, as well as graduate starters, to have a level of AI literacy that if I think back even for me two or three years ago in my old role, I would never imagine we were expecting people to have to have this level of Base AI literacy and knowledge, but we are. Because our view is these tools, this technology will have a profound impact on every single thing that we do for every single person in the organization. So the first thing is you're establishing a level of base knowledge. Second thing is you're making sure that everyone has access to those tools. So in our case, it's Enterprise Advantage, which is our AI platform. That runs on a lot of the frontier models. It also runs on any of the hyperscalers from a cloud point of view. so that's deployed across all of our 168,000 people. And we expect them to be using that every day, every day in how they're working, every day in how they're delivering. We then look at then the next level of like you know, more expert training, more specialist training. We constantly look at the skills profile of our people and if there's certain skills that inevitably There were in less demand for clients, we have to re-skill those individuals. We have to give those individuals the opportunity to be re-skilled and redeployed so they'll go into maybe a Google practice or a Microsoft practice. But they may have come from one of our traditional technology practices that was more with legacy with legacy s with legacy skills. The other thing is this point about trying to and you said this at the start, trying to encourage this constant point about Curiosity around the technology. Obviously, the advantages to be candid that we have is you know, we have a massive research on. we're constantly researching, and you sort you may have seen the announcements a few weeks ago about what we've just done on quantum. We announced Lightwell, which is a clearinghouse for open source code that we're gonna go through and you know, basically patch any open source code. Every time we say this to our people, we then say, we expect you to understand. you know, the basics of quantum, and we're gonna roll out, you know, a one-hour quantum, you know, refresher for everybody. And we expect everybody to do it, and that includes me as well as, you know, a graduate. And so this to this point about the thing with technology is the speed with its changing and it's progressing. you know, the and the way we think about it from a consulting point of view is this is an incredibly tough Tough, challenging market. The people with the best trained people with access to the best technology are more likely to be successful with clients than people who can't use a technology or they're only using basically ChatGPT on their phone or they haven't found how to harness it. And I mean because you work because we work for a software company, there's a sort of an expectation, you know, that you understand this technology. You know, I always made the joke, I made this joke about So when I was at my old firm EY, loved the firm, had a great time there. I was responsible for technology at EY. And in EY, they thought I was a technology expert because basically I could turn my laptop on and not call IT support. Because the organization a lot of there's a lot of accountants in the organization. I go into IBM and I kid you not, most of the people I work at senior level, they spend most weekends playing around with Bob, which is our coding platform. They play around with Or they play around with OpenAI. Most of the board I'm part of, there's a bunch of them with PhDs, and you know, the level of knowledge in a software company and this desire and need to constantly learn and play around with the technology is pervasive. It is pervasive. I mean, I have a I have a 25 year old coach now who helps me on all the latest technology. Every week I sit down with them. go through it, we start playing around with it, and then he shows me how to use it in you knowhat my in my day job. you know but the idea is I'm trying to get everyone to that level of basically self sufficiency because we think you need to be at that level. in our market because the way consulting is changing is so profound. You know I think I think for clients I think it's for clients it's different. Different industries are different. I mean, if I was talking to somebody who was working in the advertising industry, for example, I think advertising is being as disrupted as consulting. So I would be having a similar conversation with them as I would be with my own teams. I think if I look at say manufacturing the argues has already been down a heavy automation path for the last X number of years. So it's different by industry, different by client.

Paul AshcroftYeah. I want to pick up and click into you mentioned about the companies that win, you the access to the technology, the willingness to learn and the speed in which they can do that. I wonder if you would add to that a culture of adaptability and a mindset to actually want to do this, because there's a lot of people frozen in fear, worried about their job. Indeed, a lot of people are losing their jobs, but equal. I, my experience is this is the same as yours. I've past few years I've been like a kid with a new toy. I've dusted off my computer science background and I've been building sites, building apps, accelerating the way we do things. It's fun, you can be curious, you can get a brilliant result and if the organization that you're working with... adapts that mindset of try things in a guard railed safe environment, then you get a lot of fast change. But there's got to be a lot of adaptability required in the organization to deal with the, as you said, change that's coming. Right.

Andy BaldwinYeah, and I look I think it's a great point. And I think you know, if you look at a lot of traditional technology projects over the last twenty, thirty years, most of the investment often went in the technology itself. And so when you came to the change management pit, you were often left with and I used and I lived this was a lived experience for me as a consultant for many years. You know, the amount of the budget actually went into change management was relatively small. It all went into the technology. Actually Because a lot of what AI is doing, once you get the right setup and the right architecture at the enterprise level and you make your right technology choices, a lot of it is federated and democratized. And therefore you can spend and invest more time on that change management, the skilling piece, and also, you know, activating the individual's own curiosity for using the technology. But as you rightly said, you have to put guardrails around it, you have to put protections around it, particularly with using your data and how that data's being used. So I'd say this is a big theme and a big shift I think we're seeing, which again is different with I think the with the AI experience than perhaps some of the other technologies, because it's it's having a far bigger impact, you know, on employees, employee productivity, how they think about that. And I do think we'll also see an increasing shift in focus to use of the technology in revenue side. Because I think a lot of the focus has been on productivity, if we're honest. But I think you will see the balancing out onto the onto the growth side. But at your point, I think a critical part of it is around education and motivation and providing the right environment to the work your employees and your workers. I'm looking for honest, in knowledge based industries, I think the technology is having a more profound impact on knowledge based industries, more so than perhaps manufacturing and some other retail. But I think in knowledge based industries, knowledge based professions, it's having a quite significant profound impact. But the good news is I think in the way a lot of the technology is being deployed and federated, it people are having access to it. as opposed to it being, you know, restrictive, which I think is a positive thing because if they're using it in their work life, more likely than not they're going to end up using it in their personal life. That's really one of the trends we've seen over the last 15 years on technology is people started to s experience technology very different with the launch of the iPhone 20 years ago and they expected the same experience that they had in their personal life on their personal devices as that they went into their corporate environment. And for a few years there was a big lag, as we know, but then gradually the corporate environment, probably still behind but probably the leading technology in the personal space, but it's definitely converged, it's definitely caught up and the experience now is not that dissimilar.

Garrick JonesAndy, I want to talk to you about something you also just touched on we've been talking about AI, we've been talking about tokenization, we've talking about technology and the impact on people and how important that is. But you just mentioned quantum. And that to me is like one of the big tripods of the things impacting the twenty-first century. You know, AI, tokenization, yes, quantum. What's your view on quantum computing, quantum mechanics, and how that coming through is starting to impact? I know it really doesn't impact everybody's everyday life just yet, but that may accelerate too. What do you think?

Andy BaldwinYeah, look I think, you know, we've our chairman CEO Arvind Krishna has been passionate about quantum from when he first started doing the first quantum research at IBM about fifteen, twenty years ago. and you may have seen the announcement of the US government, they basically invested a number of a number of startups in quantum, but also they invested And a new business that we're going to launch jointly with the with the US Department of State, I think it's Department of State. It'll be a quantum chip manufacturer. So I'd say what I think we're seeing with quantum. quantum is the first new paradigm in how computers operate probably for the last 50 years. So that, you know, the machines of the moment, anyone who's in London, walk down York Road, by Waterloo Station, go past the IBM office. There is a quantum computer in the window. so these things at the moment, they are not, I would say that they're not consumer ready. These devices and machines. but obviously when you look at experience the pace and change of quantum of computing generally, I think you are look you will see a time and a place where these will be a more mainstream hardware device. You know, at the moment, obviously what they it allows you to do. Is to do a level of sophistication and complexity and calculations that is way beyond what a normal traditional paradigm computer can do. So anything, if you if you think of where it's having major applications at the moment, we've done a number of projects around cancer and molecular research in the United States. Johns Hopkins, a number of other a number of other companies were working with a number of the major life companies. You know, in to accelerate drug development and research at a space and speed that would have taken years, we're now able to do it in months. So that medicine space where you do need a lot of compute power, I think that's one that's coming through, which obviously will be a broader benefit for society. I think one of the areas one of the areas that's coming very front of mind now is obviously quantum safe effectively in encryption because obviously most of our encryption around the world is based on traditional existing technology algorithms. And so I think, you know, particularly with the launch of some of the frontier models, and you've seen Mythos, you've seen Daybreak, and the ability of those models to identify vulnerabilities, you combine that with the emergence of quantum related cyber risk. So I think that will be probably the first major application, Garrick, that we think will come off the rank, which will be a broad based, you know, mainstream application. I think as the technology comes through and gets more mature and gets more usable, then there'll be other applications where you can apply that compute power to

Paul AshcroftIs it concern you Andy, on the security side, that when the quantum computers can crack our current encryption, will already the encryption models keep up or will we start to see some big problems that we rapidly have to get our hands around?

Andy BaldwinNo, look, I think on a serious point, I think the pace and speed of which technology is developing and evolving, there is always the risk that the technology falls into the hands of the bad of bad actors. I and if we're candid, you know, a lot of cyber security, cyber breach is state sponsored. So I think inevitably, I think when you look at quantum, it is a concern for people, it is a concern for governments, it is a concern for countries. I think though I think this is probably why you've seen so much research go into this at the national security level in a number of countries around the world. and I think If I think of the work we've been doing or what we have been doing and we're doing for a number of governments around quantum. And we're working with the US government, working with the U we've done some stuff with the UK government, we've done some stuff in other governments around the world. I think the research will be at a pace and a maturity that I think it will it will provide protection. But then you've got to get that deployed and get that, you know, made available. And one of the things we have to accept and recognize is The future of cybersecurity, candid is agent versus agent, it's not human anymore. I mean the technology is so powerful that you your defense, your cyber defense is largely gonna be it's almost entirely gonna be agent-based, agentic defense against bad actors using AI or quantum in some shape or form to try and crack your defenses. So you're gonna see this whole I think raising of the a raising of the bar on both. suddenly the state actors and bad actors and criminals will get more sophisticated. Everyone will have to respond by being more sophisticated in return.

Garrick JonesIt's amazing. It's it's like it's complete science fiction, isn't it? It's now we get to a point in the evolution of humanity where we've developed enough technology that the technology fights with each other. But whoever's got the strongest technology, whoever's got the most resources in the tech space or the cyberspace, that's who wins. It's like it's it's right back to something you out of some of the sci-fi novels. It d Do you I know you can't go into too much of these kind of details, but you guys are now ecosystem partners with people like OpenAI and Anthropic, and as you said, th you know, these are billion dollars, trillion dollar organizations which never existed five years ago. What's that what's that like and what's the role of IBM with the new players?

Andy BaldwinNo, well look, I think IBM obviously we're one of the largest software companies in the world. We produce our own software. we're we're very strong in that automation integration AI space. As the consulting business, we clearly and obviously work very closely with our own technology arm, but we obviously are an open ecosystem as IBM Consulting. We work, you know, with all the major ISVs and software companies. So we work with AWS, Microsoft, Google, we just announced a Google partnership. We work with SAP, Oracle, and we do work with OpenAI and we work with Anthropic. And I should say we work with OpenAI and Anthropic on our software side. more with Anthropic, probably on the software side. And then on the consulting side, we have, you know, we have relationships with Anthropic. you know, we've been working closely with OpenAI as well. I think the you know, where do we see you know, the value for us in working with those organizations. So one I think Many of most of the major enterprises, if you look at what's happened with cloud Garrick over the last 10, 15 years, most organizations have become multi-cloud. They'll do some activities on the cloud, they'll have a couple of cloud providers, they'll do some activities, stay on-prem in some shape or form, and they want the ability to manage data seamlessly across a multi-cloud, multi-estate. We believe AI is going to go the same way. We believe clients will be multi-model. They'll use different models for different tasks. They'll need to manage the data. They'll need to manage the model providers. They'll need to manage the code developing tools. They'll need to manage the way in which agents are deployed into their workflows. And they'll need to integrate those agents into their complex legacy environments in some cases. For us, in our positioning, we want to be the cloud integrator of choice. We acquired Red Hat. To basically give us that positioning eight to ten years ago. We want to be the trusted integration partner. We're obviously committed to open source as a as a platform for our clients. And we work with, we recognize we need to work with all the model providers because if our clients want to work with the with all the model providers, then we need to work with those model providers. in the same way we work with all the cloud companies, you know, we believe we need to work with those. those model companies and you know we have because we work with them we have a good view about which industries, which applications some of the models are particularly strong for. And equally in some cases where some of them maybe there is a better or cheaper alternative or choice. So that's the way. So as a as part of our you know strategy to be committed to do it to an open o an open ecosystem. We that's why we announced Google a few weeks ago, GCP partnership, and we'll look at the other frontier model providers as they emerge. But bear in mind a lot of clients will use the frontier model providers, but they've also produced their own models. You know, we're doing some interesting work with L'Oreal, where we're producing for them, it's their data. We just provided the chassis and the AI engineering skills. We're producing the world's largest AI model on sustainable cosmetics and perfumes in the world. That's unique to them. That's their IP, it's their data. We've set the we've created the environment because they want to protect that data. They want to hold that data because they believe and recognize that's their basis of competitive advantage. So they're not going to make that available. Look I think I think what you'll s I think what we're gonna see is goes back to Paul's earlier point, as the clients go on this maturity journey, I think they're increasingly recognized that sovereignty. and sovereignty you know does not mean I'm only gonna use technology in one from one country in one place. It means I have oversight and ownership of data and the technology that I'm going to be using and I know where that data is what how it's being used, I know where that data's going, I know where it's being stored, and that's consistent with my risk appetite, my governance and my standards. And I think that's going to become increasingly important. I think for knowledge-based industries, inevitably your domain expertise When you strip everything away is often what really differentiates you. And so I think people protecting that data and understanding that data and then thinking about do they build their own model with that data, or if they do provide it to a large language model provider, how do they ring fence that data from a sovereignty point of view and avoid it being shared? I think for sure, is key. I mean, I always remember this story when we were When I was at EY and we were during the period of Everest when we were looking at separating the firm. And we were working with Microsoft. And Microsoft looked at a product EY had called I think it was EYI, which was basically the tax repository. And this was the largest single tax repository in the world. In the world. EY has the largest single because they're a globally integrated business unlike some of the other big four. They pull that information together, they have all that data. It was the largest share one of the largest SharePoint sites Microsoft said they've ever seen in the world. but it was all tax data. That is incredibly valuable to anybody, but it's especially valuable to the world's second largest tax consulting firm in the world, which is what EY was at the time. And so I think that's the point about the data. I think the data is gonna be I think people have real have increasingly realize that protecting the data, understanding the data, but putting sovereignty boundaries around the data, who can access it when they access it, the permissibility rules, etc., is a critical part. But people still use large language models. We use large language models in IBM, but we're very clear what we use them for. We have some of our own models we use for certain tasks and activities. This is the point about having the right harness to manage the models in the most cost effective way and managing access to the models and ensuring the models are being used for the right task. Back to what I said before, you're not using a Ferrari to go and collect the morning paper.

Paul AshcroftIt's fascinating Andy, it's like a ball of string or those old iPhone headphones that we used to have that we didn't always tangle themselves up no matter what you do with them, isn't it? That there are so many threads to pull and questions to ask, you what are people going to be for? How to use that technology? How far should we go with the technology we're using? I read in the news recently that philosophers are now more in demand as graduates than computer scientists, which says something about the state of the and what we're having to think about. I wanted to ask you a personal question, if I may. You've been in a position where you've led very large organizations, hundreds of thousands of people, many, countries all around the world. But what strikes me speaking with you and we've known you for long time is that the thing you always seem to keep alive is your own personal curiosity. The way you talk about the challenges we're facing and what you're doing at the moment, I get a vibe of curiosity from you. And I wonder how, when you're leading these large organizations, how do you and do you try to bring that into the organization that you are leading as opposed to just a very compliance-driven... organization, how do you as a leader orchestrate the people that you are leading?

Andy BaldwinYeah, look, I think it's this point of and I've been very fortunate. you know, I worked for an insurance company many years ago, great company, but I've really worked for EY and IBM twice each. So I've only worked for two companies in thirty odd years. I just happened to work for them both twice. shows you I'm not particularly original. but I would say both organizations have a have a point about the I think this desire for learning in the organization. You know, because it's it's a professional services consulting business where people progress through the you know through the ranks and as part of that progression you need to demonstrate your knowledge, you need to demonstrate your expertise, you need to demonstrate your ability to work with clients. And so I think I've always taken a very strong notion of stewardship of the businesses I've worked in and led, which was wanting to leave them better at the end than it was when I s when I started. think on balance in most of the k most of my time I've done that. and I think a critical part of that is developing people and bringing people through and helping them. You know, it gives me immense pride if I look at EY. There's lots of people who work for me or graduates who work for me are now running parts of EY in the UK or globally. And I think that point about constantly learning, encouraging people to constantly learn is a critical part of it. But I've been very fortunate in that EY and IBM, although they're different industries, actually very similar in some ways, culture. there's this real strong desire around learning and education and professionalism, and that I think made it very easy to encourage people. you know, you would be amazed when we release these learning programs, these one hour, these podcast learning, you'd be staggered how quickly they're deployed, how quickly they're downloaded and consumed by our people. Mind blowing.

Garrick JonesAndy we could keep talking to you for so long. I wanna ask you, Andy, what are you currently curious about outside of outside of work? well my

Andy Baldwinwhat am I? I actually I've always had an interest in politics. And maybe it's because my dad was a many years ago in Liverpool. My dad was a politician, Labour politic he was a union leader and a Labour politician for many years. I grew up in the when I grew up in Liverpool. And actually I just find it's funny, I've been fortunate to serve on a number of major company boards and you always thought talked about geopolitics, but actually The impact of geopolitics now is far more profound, particularly when you combine that with the impact of AI. And so, you know, I've always had an interest in it, but I've been reading, maybe it's because I work for an American company, I've been reading a history of American politics by Simon Jenkins, which is quite an interesting, quite an interesting book on just the history of America, actually. If we forget, I mean, there are schools in the United Kingdom that are 400 years older than the United States of America. 350 years older than the United States of America. So it's a very young country. So I've been reading about that. and also I have been reading just on AI, just generally about I you know I we touched on some of these points, you know, as a father of four children, I do think a lot about the impact this technology is gonna have on the s on the world and the environment that my children are gonna come into or and join and be part of. Going back to one of the things you just said there at the beginning, Paul, about AI. Interestingly, my daughter, obviously incredibly proud of her, and she's super bright, but she's an English graduate. But she's in the AI division now of one of my competitors. And that's sort of like I look at it and go, you what? Actually, there's a lot of change, but there'll still be the opportunity for people who are bright, want to learn, motivated, think about how to use the technology. And you know, they'll just

Paul AshcroftReally amazing.

Andy Baldwingo down a different path perhaps than what someone would have done five, ten years ago. They may have been a mathematician going software. Actually the technology allows an English graduate, a very smart English graduate I should say, but an English graduate to go and become an AI an AI consultant.

Garrick JonesIt's amazing. The thing about our kids, it really was our kids that prompted Paul and I and Simon to write the first book, The Curious Advantage. Because we were thinking, What the hell do we teach the next generation? And how do we keep things from being, you know, so narrow in the digital domain? and that kind of was the spur that got us, you know, we landed after doing all our research on curiosity. That's how we got into curiosity, and now we've got the curious imperative and so on coming on. But Andy, I'm I'm gonna ask you to just for one final thought. About what it is you want to leave our listeners with and our viewers. But before we do that, and it'll give you some time to think about it, because I'm gonna just cover some of the things we've talked about in this fascinating conversation. you started by talking about constant learning and curiosity and as a the link between curiosity and constant learning, and talked about what's changing, you know, the tech stuff, and how things have become shifted from just when we were in the internet age to just l into an age where we're looking at scale and transformation. And then how tech has AI tech has a much broader application. And the volumes are huge and we're moving to a world which is around a blend of digital and human and that has impacts on both consulting and tech and also everybody's everyday lives. We've covered the growth of tech and AI, we've covered gen AI, agentic, and how the shift is to organizational wide implications and how it touches everywhere. and the shift from new business models and shifting into completely new fundamentals around how business is done and how it's organized as a result of all this. And we had a conversation about how do we deploy it, where does the in value come out of it? The inflection point that's related to scaling. a conversation about token economics, which I'm very interested in, and the shift from you know, showing well the shift towards showing, not telling, and how Token economics and assetization has turbocharged things, human and digital working together and how do we account for that and so on. the impact on consulting, and the how are we moving to an outcomes-based shift? How do you manage the change? Another part of our conversation. You talked about giving people access to the tools and expecting them to use it. reskilling, retraining and encouraging play, with the tech. And how to focus on the people, which is what it all boils down to activating their curiosity. A few more things. We had great conversation about AI and tokenization and quantum and how quantum is the first new paradigm that we've experienced in computing and how it's going to just touch everything. And you talked about some of the applications we're already seeing for things like drug development and encryption. And one of the things I was doing while we were getting this all done was just some of the words that that you were using and some of the new words that really kind of define a whole bunch of stuff. And my list includes things like agentic, cybersecurity, cloud, ecosystems, open source, multi-models, and sovereignty of data. I mean there's a book to be written right there. Andy, thank you so much. it's a brilliant conversation with you as always, straight out of your experience. What's the one thing y you'd like to leave our listeners with?

Andy BaldwinAI, look, AI is a contact sport. You have to play with it, you have to use it. whether you've got access to tools at work or whether you've got access to them at home, most people have got access to Copilot at home. have a go at building a couple of agents just for yourself, for your personal use of think things that you need to do at home and have a go at building the agents because I learn more trying to build some of these things myself. without having someone do it and then showing me. but AI, it's a contact sport. You've got to basically learn to use the technology and it doesn't matter what age you are, what profession you're in, what level you are, it's a contact sport. You've got to learn to use it.

Garrick JonesAndy, AI is a contact sport. Thank you so much for the conversation, we really appreciate your time.

Andy BaldwinNo worries. Good to see you guys again. Cheers.

Garrick JonesGreat. You've been listening to the Curious Advantage Podcast. Get ready for our new book, The Curious Imperative: Why Questions Matter More Than Answers in the Age of AI. AI can now write, analyze, design, and decide at extraordinary speed and scale, as we've been talking about. And as the pace of change begins to outrun our ability to adapt, curiosity is no longer simply an advantage, it's becoming essential to how we navigate the future. And we draw on neuroscience, economics, organizational research, art, language, and more than a hundred conversations with people at the forefront of human capability. So the Curious Imperative explores how we can recover the curiosity many of us lose between childhood and adulthood. This is not a book about AI, it's a book about people in the age of AI and whether we are curious enough to navigate what comes next. We're curious to hear from you. If you found something useful or valuable in this conversation, which I'm sure you did, please leave a review on your preferred podcast platform and tell us what resonated with you or what you learned. We always appreciate hearing our listeners' thoughts and continuing that conversation, which is vast. Join us at #CuriousAdvantage on all the socials. The Curious Advantage is available on Amazon and all other good bookstores worldwide. Subscribe today and keep exploring curiously. See you next time.