Latest / Range Podcast / The AI Board That Makes Decisions Better

Transcript

The AI Board That Makes Decisions Better · Range Podcast

• Giac Soliman: Good morning everyone and welcome to a new episode of the Range Podcast. Today I'm here with Yvonne. Yvonne is an expert in strategic rewards and pay transparency with over 20 years of leadership in global organizations. As managing director of a family office, she combines strategy with value-based governance. Former Senior Director of Global Rewards at McDonald's, she led reward for over ninety thousand. Employees across 16 countries and launched the Global Pay Transparency Initiative. She speaks on the EU Pay Transparency Directive and of course AI governance and rewards. She is section editor rewards for Springer Handbook of HR Management in the Age of AI 2027, holds an MBAI, and is completing an AI board certification. I feel This list can go on forever, right? There's so much that you've been doing, Yvonne. It's so good to have you on the show. Welcome. Yvonne: Thanks for having me. Yeah, I'm a passionate and fiercely curious person. So yes, always want to do something cool. Giac Soliman: Excellent. Now moving on to AI, which is the topic of today. Well, first and foremost, this topic is something you deal with on a daily basis in different shapes and forms. And first off, I'd love to understand what excites you the most about AI. Why are you investing so much of your time in so reading, building? Researching, etcetera. Yvonne: I think AI is the new normal, right? It's going to be the operating platform. I mean, Gen AI is really, really cool, but but and to a certain degree, AI is trendy. My pivot into AI was not because AI was trendy. I saw the same pattern. Basically, leaders want speed and innovation. Employees want clarity and safety, and governance in that space arrived often too late. So this is what really excites me where it's like, how can you how can you leverage the speed and the innovation? How can you get to clarity and safety that you can trust results and How do you make that happen in an environment that is IT but not IT? How can you bring business ideas forward with speed and innovation and clarity? Governance and innovation should go hand in hand. They should not so governance should not stop innovation, but innovation should not ignore governance. I think that's that's my key point, basically. Giac Soliman: When I think about innovation and governance, it it reminds me of like of a love and hate relationship. Like they so much hate each other, but also know they need each other and sort of in that sense they love each other. It is quite a very like challenging topic. You you know, you chose, I would say this this challenge you you you know, to focus on AI governance and AI innovation for the board, I'd love to understand more of how you think that will pan out in real life because in What is your secret when it comes to bringing Yvonne: Yeah. Giac Soliman: innovation to this kind of environment that some people might argue it is actually retroactive, not proactive? Yvonne: Mm-hmm. And you're pointing out something something important here. I think it's about accountability and about doing your job right, right? One of my mottos, my my sincere belief is do the right thing, right? For people, for stakeholders, for the organization, and then do it right. So that is about governance, that is about process, that is about innovation, and you know, get things going. I believe, I mean, in boards you have the business judgment rule, and the business judgment rule forces or requires from board people, representatives, basically to use all the resources that are sensible and that are available. To inform decisions, to do research, and to me in this age, it absolutely includes AI because that is where you can actually use really yourself to very quickly research stuff, get an overview, you know. I mean, those two fifty to two hundred page board. memos that are sent out a week before the before the meeting. You feed it into AI, let it summarize stuff, let it ask, like ask it questions. Say, like, well, I don't understand that. Explain to me how does the eBIT day DI bridge for this deal that is being proposed actually work. You know, I I I don't see it. I don't like how does that work? So you can come up with the right questions to ask in a board meeting, right? If you Don't understand the material and who goes through 200 pages? I mean, it's nonsense, right? Giac Soliman: Mm-hmm. Yvonne: Can you trust it? Is the other and the better question. I mean, can you only do shortcut? No, you can't. You still need to apply your expertise, your experience, and you need to challenge the AI. Sometimes I ask my board, basically, I build a digital board. I ask, like, I don't believe this. Like, where did you get that from? Right? What's the reasoning? Explain it to me. Why are you coming to that conclusion? And then I almost have a like a challenging discussion with my AI agents to say, Well, you know, I don't quite believe this. What if we look at it from that angle? Or we look at it from that angle? Or I haven't thought about that. You know, so it's more a challenger team than just doing what I tell them to do, the other is perspective and insight. And I think there's a huge difference between the two. Giac Soliman: That conversation you're having with your digital board sounds very insightful. It also sounds like a conversation I could go on all night, right? But for everyone's Yvonne: Yes. Yeah. Yeah. Giac Soliman: listening, could we define what is a digital board to start off? Yvonne: Mm. Giac Soliman: For people listening that have very little clue perhaps of what a digital board is, including myself. Perhaps. Yvonne: Well, yeah. I build myself a team of agents and I basically build them having particular skills. So I do have a strategist. Her name is Selena. I give them names because then you know it's easier to have a proper conversation. So Selena she excels as doing strategic analysis. So whenever I have something that is around strategic analysis. I feed her documents, papers, I tell her, go do research, right? Tell me what do you think about it? Why do you think about that about it? so I basically gave her skills and I gave her a knowledge base. and then she does analysis. And then and then I do have Max, and Max is an MA specialist. So whenever there is an MA transaction where I'm looking at something. Then I feed it to Max, and Max does some of the analysis, right? And says, like, you know, this EBDA multiple is outside of the usual range. You know, you need to feed it a lot of context information, and you need to do a lot of, I wouldn't really call it loop engineering, but to a certain degree it is. You know, you go in loops, you say, like, okay, this is the initial result. Don't quite believe it, go back, redo it with this angle. Or there's another another information. So I feed it in there. Or I do have Frank Finance, and Frank Finance is my CFO. So whenever there is something that I have that needs and needs analysis, and I'm like, okay, Frank, tell me, what do you think about this? Is the free cash flow enough to sustain? What about cost of goods sold? Is there a problem? You know, things like that. and I also do have a Chief People Officer, her name's Carmen, and and I ask her questions and I'm like, you know, do you really think this is sport ready? Particularly when I'm writing a concept or when I'm You know, they always then ask me back and say, like, or tell me, well, you haven't answered this question, or you haven't looked at that angle, or you know, to actually really be board ready, you need to include this or that, or you need to rewrite it, so it gets better. It really Giac Soliman: Mm-hmm. Yvonne: gets better with the different iterations, the loops that I'm building in. Something that has taken me days to prepare now takes me hours or even still a day, but it's so much better quality. Giac Soliman: What makes you trust your digital board? What makes it reliable? I mean you probably went through a journey at the very beginning he had version one, two, three, four, etc. But for someone wanting to build something comparable to your digital board and sweating it because they don't know if it's gonna be reliable or not. And this is stuff that goes eventually to the board. So you really don't want to mess up. What steps did you take to make the answers and the output coming out of your digital agents more reliable? Yvonne: I would say I have an agent that writes pro that I use to write prompts, to build agents. So basically I use AI to to to build. And I ask it what you know, what does it need? And then I refine. And I say, like, that doesn't really sound. And then I start testing. And then I'm like, I'm not sure. And then I I use a combination of pre-built skills. and building my own and also training really describing in the agent prompt and in the agent knowledge what good looks like and what not to do this is even more important I think you know I I kind of trained it because I learned give me your confidence level every time show me the resource, show me a link list, show me the resource, I verify, I cross-test. and what I also do is I cross-challenge. So I have a conversation with one agent, and then I pull in another agent, and I ask the other agent to verify what the first agent did. So it's a kind of an evolving system. and also What I did is I built I build knowledge documents and load them. So the agents have knowledge and context that I gave them and I refine that very often. When something's not quite right, and when this wasn't really what I wanted, then I go back and ask my writing agent, my HR H H business partner agent, basically. this is the MD that I have right now for this agent or this is the prompt I have. This is the problem that I came across. How can we make it better? And then I kind of refine the prompt or refine the knowledge or look at so it's an evolving system, right? It's evolving Giac Soliman: Mm-hmm. Yvonne: because it's learning. And the question always is, you know, is the handoff proper? Or, you know, because I do a lot of chat iterations potentially, I sense when it goes off on a tangent. And then I'm like, no, no, no, no, no, that's not the right way. We need to, you know, go back on track because this is the direction that I want to take. So I think it's a combination of prompting. Training, learning, skills, and I know what good looks like. And I apply my judgment to say, Well, you know, that's not that doesn't feel right. And then I go off and potentially do my own research in perplexity and then I challenge the agent and say, like, you know, that's not right. This is not true. And and then so it becomes an evolving system. Giac Soliman: Gotcha. There are many reward leaders at your level I've spoken to that have decided to delegate most of it to someone junior in their team but it seems that you decided to go hands on and you know invest a lot of your time in it. Why? Yvonne: Because I bel I mean it's a huge opportunity, right? I want to get better, I want to learn, I'm curious, I want to explore. I want to know what good life looks like. How can I lead something that I don't understand? Right? I don't have to be the most technical expert, right? Really not. I'm not, really not. But but I need to understand it at a certain level in order to be able to lead it, right? And this is what I think most leaders, there is a certain level of AI fluency and a AI literacy. What you don't understand, you can't lead. That's that's what I firmly believe in. am I the best builder? No. Probably not. Am I able to point out where a build could go wrong and then look at and then assist in solving a problem or describing it in a way to say, like, have you thought about this or have you thought about that? Or why don't you try this? Is something that helps me, I think, stay grounded in the everyday, understanding the everyday. but at the same time also applying the right sense of leadership to say, like, you know, not backward delegating, like I'm doing it and the team doesn't, that that's also not right. I think we should learn together and I want to lead by example. Giac Soliman: Love that and it's so important to your point. If you need a lead on it, you first need to understand yourself. Because when people look up on you in your team and ask for guidance, sometimes that's also part of being a manager and a leader. How are you able to provide that guidance and sort of break the ice from a deadlock if you don't understand the topic at hand? Yes, you don't need to be involved in all the details, but you need to have strategic sort of vision and direction. You need to be able to provide that. I understand that you've been working on some sort of vibe code of conduct. You might not be the best vibe coder. And who is who how do we know if we are, right? I think just by exchanging knowledge with other people there like minded and so committed to the same cause, that's where we sort of start to benchmark ourselves a little bit. But this vibe code of conduct, what are the first, let's say, two rules on that list? Yvonne: if we if we think about AI and use cases right now, right? This is there's a lot of piloti pilotitus, I call it that. You know, everybody builds in their own backyard. And in order to actually scale, I mean, you know, having worked at McDonald's, McDonald's is a an organization that is known for Scaling, it's known for consistency, it's known for all that. So that is so engraved in me that you know, how do you move from a backyard build? Which is great because you have the innovation, you have the speed, you have like actually somebody who knows what they're talking about, are building something that is supposed to do to solve a problem, right? To provide a solution to a problem. What but then you run into the issue that you need to integrate, you need to scale, you need to integrate in an organization. And that's where you run into IT problems again, right? So so and often, and very often, that then becomes a problem because the coded application is a black box to IT. They don't necessarily understand. what's going on or how to integrate or what are or they are start to ask all kinds of questions that you potentially don't necessarily have thought about or have built into the product. So that is very frustrating because you run into this roadblock. So basically from a backyard build to a scalable IT architect you need not just data and a model or a build, you need training data, definitions, you need data owners, you need quality rules, you need cons like consistency, right? You need no conflicting sources. So What I think every organization needs and and would really benefit from, is really solving that with a vibe code of conduct to say if that is needed for anything for to move out of pilotages and backyard building stage, why shouldn't an organization kind of do the bare minimum in terms of What questions do you need to answer or do you need to build into your white coding from the beginning so that it actually becomes a viable product and that you can integrate? You know, I know this is for me the right level of governance to say, okay, let's think about it already in the build, like very simple stuff. But how does access control look like? You know, where does the data come from? Who who is owning the product? Because that breaks so many builds. And I'm so frustrated by that, right? Because I see it happening all the time. Right? You build something great, then you want to take it from your desktop to somewhere else so that actually people can work with it. And then and then that breaks because it hasn't been built with that in mind. Giac Soliman: Mm-hmm. Yvonne: So that's on my mind. Like, you know, can organizations really do a an easy bike code of conduct that lays out those principles, those easy things. Everybody thinks about that when they build. And then it scales, it integrates. So I think it's not that hard. It's just a little bit of, you know, step think one step ahead than just the product itself. Giac Soliman: Asking the questions, absolutely, I think it's is probably the easiest part. Finding the right answers probably will take time, especially if we're thinking about instilling more product thinking in our domain to rewards. Absolutely, we we definitely need more of that to make sure that whatever we build actually works. what you said is making sure that whatever you build you can trust and works not just for you but for your team, for your stakeholders, and for your board. As we approach The end of the podcast, Yvonne. I have one final question for you. And I am wondering if you could wave a magic wand and fix one massive headache you had over the years when it kind when it came to board governance, corporate governance, compensation, etc. What's that one thing? you would install in every boardroom overnight to get rid of that pain. Yvonne: That is such a profound question. And I could probably talk to you. I would say actually really understanding reward philosophy and the why behind it. Like why is it the right thing for your organization? And how does it trickle down? How does it show the values? How does it how does it connect? Right? How does it all connect? Because I think reward Can be such an operating platform for the whole HR function if it's played strategically. So magic wand would be tomorrow. The board actually truly understands how reward is used as a strategic lever, helps the organization achieve its goals, drives the right behaviors, or rewards the right behaviors, right, that are associated with values with the employee value proposition, so that it actually becomes an operating system that a board uses instead of just accepts, it's there, right? We need it. But it is at the core of how do we operate this company? Well, you know, reward is one of the key levers. So that would be my wish.