[MUSIC PLAYING]
Welcome to The new shape of work podcast series. I'm Kate Bravery, Mercer's insight and advisory leader. And today I'm chatting with Martin Smit, head of global compensation solutions at Takeda.
And we'll be chatting on how AI will transform the reward function. And what does it mean for you and I in terms of the work experience? Martin, it's really great to have you on the podcast today. And, as always, it's a pleasure to be partnering. So thank you for joining.
Great. Thanks, Kate. Thanks very much for inviting me. It's a pleasure again to speak with you about artificial intelligence, a topic close to my heart.
I know that. Your passion is infectious, which is one of the reasons why I wanted you to join me on the call. And why don't we kick off with a little bit about yourself, your role, and maybe why you're on such an AI mission at Takeda.
Great, thanks. Well, so my name is Martin Smit. I'm based in Zurich, Switzerland. And I'm the head of global compensation solutions in Takeda, as you mentioned.
Takeda is a global pharmaceutical company. It was founded 1781 in Japan. Currently has about 50,000 employees in 80 countries. And Takeda's mission is to change the lives of our patients by focusing on our values-- integrity, fairness, honesty, and perseverance.
Myself, I'm very passionate about technology and especially AI. I like your words, infectious or passionate. And I think that all probably started in around 1983, '84 when I got my first computer. And, actually, I was a big fan of Star Trek. And their motto to explore strange new worlds where no one has gone before is something that I really love to do.
And, in terms of AI, 2017, I really started deep diving into AI. I took a course at MIT on how that impacts the business strategy, wrote a paper about how it could disrupt the pharma industry, and created a simple website about how organizations can embed it. And, funny enough, yesterday, when I was cleaning my archive, I found a paper that I wrote in 2017, which I called at that time human language AI code. And that paper, I wrote in there that I think you need to have AI talking in human language to people for it to be widely used.
And, at that time, I thought it was a very crazy idea that nobody would listen to that. But, now, when I look back, and we've got chat GTP, a crazy idea six years ago may not have been such a crazy idea. So I do think what we've seen with GTP is that the ability for humans to talk in such a simple way in a human language with complex solutions, I think, is making a huge difference. So I'm very passionate, and I see a lot of opportunity in this space.
Well, I know you do because you share it with a lot of the people that follow you. And, of course, I am one, and we'll talk a little bit about that later. But I didn't know that Takeda was around since 1781. That is staggering. And, also, always good to speak to a fellow Trekkie.
[LAUGHING]
It's so funny because I do feel that being soaked in science fiction as a background actually opens up our minds and helps us think what could be. I think last time we were together it was on the day that ChatGPT 4 had just brought out its translations. And you were sharing to me that it was picking up on your subtle accent and saying, should we translate into another language?
And, of course, the first thing I thought is, like, oh, wow, we're now going to have our own personal translator very much like first generation head.
Yes, well, actually, there's one of those devices-- I think it was one company that proposed these buttons. So it's-- I think science fiction is really a way for many people to imagine things in the future. And then people start building it. So, yes, I'm fully aligned.
Yeah, and sometimes it also constrains how we buy it. I've been following the AI pin as well. And, again, there's some reminiscence there. but then there's also some conversations about maybe that might not be the way it will head but very exciting.
Martin, we should get going with some of the things we wanted to talk about today because the time is going to slip away because I love speaking with you. And you know that we've been working pretty closely on global talent trends recently. And one of our headlines for the 2024 study is that, again, 99% of people are embarking on transformation this year, no surprise.
But the number one driver of that transformation is productivity. And you won't be surprised that the reason why people are driving those productivity gains very clearly sits at the feet of AI. Two in five executives believe AI will fundamentally change their business model this year. Yet only 14% of HR say they're ready for an era of human machine teaming.
And that's a topic you and I have been talking a lot about. Clear, if we're going to unlock these productivity opportunities by AI, we need everyone to be leaning into a digital-first culture but scaling that, encouraging that I think has been a lot harder than we thought. I don't mind-- do you mind coming off mute on what's been the journey at Takeda? And how have you got people embracing, particularly gen AI?
Yeah, it's a great question. I mean, early in the year, I actually thought about it a lot. So I was extremely excited, especially when the first version of ChatGPT was coming up and others.
But I noticed that many others didn't share my excitement. And I was thinking, so why is that? And I was thinking about, well, why would companies or people not be as excited as I am about, let's say, this technology? And what can I bring?
Although, I think, the last few months, a lot of things has happened in the media. It's amazing. Now, so I think, firstly, I think people are just busy with the jobs.
They focus on what they need to do today, next month, or even next year. And that prevents people maybe to think a bit and step back about, what else is there on the horizon on what we can do? So I think that's-- as with many things.
Now, the second reason why I think many are not necessarily there yet is that they may not have the systems, the training, the guidance. Or they worry about what they can and cannot do with AI because they're not clear. And if they're not clear or they're worrying, they just don't take action.
But, lastly, I think that AI is a relatively new type of technology. It's abstract technology. So I can imagine, for many people, it is not so easy to imagine and visualize, what can it exactly do? What is an-- exactly an example that they can use on a daily basis in helping in the job?
And what you and I have seen and some of the other meetings we've been in, the moment you start talking about practical examples of people, they say, ah, I didn't think about it. So I really think the key is basically in providing practical examples and use cases rather than keeping it very abstract. And then you're going to see more and more people embracing it and starting to connect to what they do on a daily basis.
So, yeah, I mean if your question is, what are we doing in Takeda? I mean, I'm very happy to say that, in Takeda, digital or AI is one of our top imperatives. It's really on the top of-- one of the top of the agenda of the executive leadership, as well as in our HR function.
Within Takeda, I think there's encouragement for people to explore and use AI with the right systems and guidance. And we're getting really great support from our digital function, which is I think important to have. Otherwise, you do not believe what to do.
So there's a lot of initiatives going in our company, working on digital assistants. We're introducing more AI-enabled processes and tools. But, especially, we're trying to make AI more visible, exactly what I've said before.
And, for example, next week, I will have a virtual call together with another HR colleague who is very passionate. And we will talk to a few hundred HR people in Takeda, what impact could AI have? To make it much more practical to them.
Specifically, in the reward function, there's a lot of cautiousness about using it in the reward function. But the reward function does a lot of things beyond just compensation planning. Maybe you can give me one or two examples about how you're making it pragmatic or how you're being very clear on how much time and dedication people should be spending to playing with this new tech.
Yeah, that's a great question. I mean, in terms of my own team-- so I've got a team based in Boston, Zurich, and Singapore. When ChatGPT came really out, I think it was, March or April, I've basically suggested the team to spend several hours per week to explore AI--
Hey!
--four hours or six hours. And, from my own experience, I know it's not going to always pay off directly right. And, sometimes, people feel, well, I'm wasting my time, or I'm playing around.
But I am convinced, based on my own experience and what I'm seeing others that, yes, this playing around will actually pay off for them, for us. And, sometimes, playing around as some people call it, can lead to the best and biggest innovation. So I think it's really important that people build in time.
And they're saying, we're not going to focus on the project. Or we're going to reserve Friday afternoon. So they can try to find out, what can these technologies do?
And it doesn't mean that I think they need to do a training necessarily or they need to start programming. But they need to experience, well, what can these things do? And, for example, learning that-- asking the question slightly differently is giving very different output. And that is a skill that you only get by practicing.
So if I look at-- yeah, so if I look at total rewards, I mean, obviously, we take trust, safety, and security very seriously. So, for instance, at the moment, I am not comfortable putting employee data or people data in AI and machine learning. I think that is still a no-go.
But, as you said, at least plenty of other stuff that you can still do. And, for instance, we use AI in designing frameworks, such as global job part detection. And you can use AI to help develop job descriptions or how you match them to the market data, such as Mercer's.
And that is previously done by humans, and this may take five or 10 minutes. And an AI can do that in five seconds now. And I think if you look into large companies such as Takeda and many other big companies is that, sometimes, projects or the type of work can take months.
But AI can really accelerate it. And that doesn't mean that, let's say, it's necessarily always much faster because, sometimes, what you want to do is you want to use the remaining time that you have left to spend on other activities, such as validation, engagement, or fine tuning. So I find a lot of benefit-- I mean I can mention another example if you want about getting better insights.
Oh, I think we love all your-- all your answers because I would love to hear-- every time I speak to you, I always get a new thing that I apply back into my work. So I think we would love that. But I agree with your point here that we should be intentional about how we're using some of those productivity gains.
I think the executives that we've surveyed around the world are expecting 21% to 30% productivity gains, which is pretty sizable. But, at the same time, employees are saying, we're exhausted. We're burnt out. We don't have much capacity.
And yet they're saying, 30% to 40% of my job is mundane or repetitive. So, clearly, an answer to all of this but, unfortunately, maybe not at the same time. But I like your clarity of direction-- spend x amount of hours or dedicate this time on a Friday for it because I think that does make a difference.
But, to that point about productivity gains, are you beginning to see that? Because that advice was a good nine to 10 months ago now. Have you started to see that yielding some efficiency gains for your team? And what are the two or three examples that others should be thinking about?
Yes, I mean, don't measure it in terms of percentages or hours. I'm not a huge fan of time writing and that people find out they've saved three hours. But I think what we see and what we notice in my own role as well in my teams is that, normally, when you would ask something, it would take three or four days. And you could potentially turn it around in a day.
And that shows, to me, that, yes, you can do these things much faster. Again, if I come back to a typical example, again, the job descriptions. I mean, yes, I could have it in the evening start.
And then, in the morning, when I come back to work, let's say, it would have done 4,000 job descriptions and not necessarily all perfect. But I wouldn't be able to do that even with a team of 10 people and say, can you give me that tomorrow? So I think the efficiency is definitely there.
But, yes, I do think we're still in a phase whereby you would want to spend some more time on validating. So I'm not-- I don't think we're there yet in blindly trusting the solutions and the advice the AI comes with. So, yes, maybe in a few years when we're more confident that what AI brings us is of sufficient quality, and the outputs are qualitative. I think that is when you're going to see efficiency gains at the moment. I think I'd rather use that time to make sure the quality is validated, and we use it, for example, for socialization and engagement.
I totally agree with you. And I think we're having a similar experience ourselves within Marsh McLennan. We've got a large language model in-house called LenAI. And it's taking a longer time to take some of our data and make it in a format that the AI can understand and a longer time to validate it on the back end. But I think we can all see the efficiency gains that will come in the future as we start to move forward on this technology, which is very exciting.
Martin, it's strange the two of us just speaking one on one because you and I have had the honor of copresenting on a panel with our gen AI-powered avatar Lara. And it almost feels a little rude to kind be leaving her out of this conversation. Maybe, as we transition into a discussion more focused on reward, we can get her point of view first.
So, Lara, I'm hoping you are still there. Lara, imagine you are in a similar role to Martin. You're a head of HR briefing your total rewards team. What, in your view, are the five most likely ways that AI will change how we do compensation planning in the future?
As the head of HR, here are the five most likely ways that AI will change how we do compensation planning. First, improve data analysis. AI can help us analyze large amounts of compensation data quickly and accurately. This can help us identify trends and patterns that we may have missed otherwise and make more informed decisions about compensation planning.
Second, personalized compensation packages. AI Can help us create personalized compensation packages for employees based on their skills, experience, and performance. This can help us attract and retain top talent and ensure that employees are compensated fairly.
Third, increased transparency. AI can help us increase transparency in compensation planning by providing employees with more information about how their compensation is determined. This can help build trust and improve employee engagement.
Fourth, better compliance. AI can help us ensure that our compensation practices are compliant with relevant laws and regulations. This can help us avoid legal issues and reputational damage.
Fifth, more efficient processes. AI can help us automate many of the manual processes involved in compensation planning, such as data entry and analysis. This can help us save time and increase efficiency, allowing us to focus on more strategic tasks.
It's so funny, Martin, because Lara always has a point of view, which is always, to a certain degree, retrospective-looking. And, as we know, things are moving very fast, particularly in the gen AI space. And I'm curious, from what you've been seeing over the last couple of months, which systems have impressed you? What have you been using and testing in your work processes?
Yes, it's moving incredibly fast. I mean, at warp speed, we would say, in Star Trek terms.
Absolutely. I held myself back from that comment, but yes.
[LAUGHING]
I mean, literally, every day when I-- or every evening when I check it, there's an announcement coming up with a new functionality or a new tool. In October, when we last met and we talked about it, we didn't have custom GPTs from OpenAI. Microsoft didn't have its Copilot in Azure widely distributed. And, this week, I noticed that Google and DeepMind announced Gemini, a potential competitor for ChatGPT. So, every week, there's something happening which I couldn't predict, let's say, the week before.
Now, in terms of systems, I can only speak for the systems that I'm familiar with as there are many people in Takeda and out there that work in this space. And there's many others. But I have been primarily focusing on OpenAI, ChatGPT, the front end but also the back end and, more recently, the custom GPT functionality, which is I think very impressive, as well as Microsoft Copilot and Azure OpenAI.
And coming back to those two options, I love the custom GPT functionality in ChatGPT. And, for the people that are not familiar with that, it basically avoids end users having to explain a lot to ChatGPT. So if you now go to ChatGPT-- and I used an example recently.
You would ask it, what standards are you familiar with? It would not have a clue what you're asking about. But if you would go to a GPT that is focused on financial reporting, it knows you would want to know about financial standards. And you don't need to explain it.
So I think that functionality excites me a lot because I think especially professionals, they don't want to spend a lot of time explaining to AI what they want to know. They want to talk to AI like they talk to another colleague. And the AI solution, whether it's ChatGPT or Microsoft, knows exactly, what is their context? So I think that is really, I think, very promising.
Now, the other thing which I really was impressed was what I'm seeing on the Microsoft space. So it's building workflows. And Microsoft, of course, is able to integrate it with a lot of tools that people and corporations use. So you can build AI and workflows with teams and email. So I find that impressive.
Now, the interesting thing is that the ChatGPT solutions, I think, are very intuitive and very easy to use. But I think they lack the amount of integration. On the other hand, you see Microsoft, which I think have tremendous amount of functionality integration, but I think too complicated for many people to easily use.
So I look very much forward-- and I'm curious, how will this OpenAI-Microsoft alliance will develop? And, ideally, something comes out that you're going to get systems that are very easy and intuitive to use in the front end but have a lot of flexibility and customization on the back end. And, I mean, I think that's going to be interesting.
And then, obviously, vendors. We use systems from vendors such as Mercer but other tools as well. And I see increasingly, as you've mentioned, vendors introducing AI into their technology and some rudimentary and some much more advanced. And I think that's a great development.
But I think it's important for customers such as us for users of these systems that we need to understand how they work. And that we're comfortable with it because, in the end, we need to be comfortable with the outputs, and we need to be able to explain it. So I think there's movement on that, but we need to see how that evolves that.
Yeah, absolutely, and the combination of those capabilities is mind boggling and incredibly exciting. But I agree with you, when we are able to leverage or import more domain-specific large language models, I think that's going to have an even more powerful impact. We're beginning to start to see that in HR, certainly, for our models we were feeding it a lot with our own policies and content and thought leadership to get the answers out. But I also see there's some now available on the public domain.
And if you think about that leveling opportunity that it has, it's staggering. If you put a medical large language model in the hands of a nurse compared to a doctor, you know who gets the biggest uptick. But, yes, embedding it into our workflows to really reduce the overall exhaustion of work, I think, is where the sweet spot is.
We were talking earlier about people sometimes just don't have the capacity to be learning this. But, when we're being nudged and intuitively learning it along the way, which I think Microsoft is exceptional at, that can make a huge, huge, huge difference. And the employee experience is going to look very different as a result.
So why don't we get Lara to share her views. And then maybe I'll come back to you Martin and hear yours. Lara in your view, what are the five primary ways AI will enhance the overall reward experience for employees?
AI will enhance the overall rewards experience for employees in the following five primary ways. First, personalized rewards. AI can analyze individual employee data to tailor rewards packages, including compensation, benefits, bonuses, and recognition to align with each employee's unique needs and preferences.
Second, data-driven insights. AI can process extensive employee data to provide organizations with valuable insights, helping them make informed decisions about rewards programs and compensation adjustments. Third, real-time feedback and development. AI-powered systems can offer continuous feedback and suggest development opportunities, enabling employees to improve their skills and career prospects, potentially leading to greater rewards as they progress.
Fourth, benefits optimization. AI can assist employees in selecting the most suitable benefit options based on their individual circumstances, ensuring that they make informed decisions and maximize the value of their benefits. Fifth, efficiency and compliance. AI can streamline administrative tasks related to rewards, such as benefits, enrollment, and compliance checks, reducing administrative burdens and ensuring that rewards programs adhere to legal and regulatory requirements, minimizing compliance risks.
Lara touched on many of those points there that you and I have had deep conversations on, especially with regard to data-driven insights, being able to customize and personalize and, certainly, in the current economic climate, optimize to make sure that every benefit and every reward has the most maximum impact for each individual. For many years, I think we've seen that our hands have been tied by our systems as to how we can nuance and tailor packages to individuals. And I think, if I believe what Lara is saying, a lot of that will open up. So, Martin, how do you think it's not just going to reshape the experience for employees but reshape how we make compensation decisions within HR?
Yes, I definitely think that the compensation function or, actually, many other functions within the company will change in the next five or 10 years and faster than I think most of us will realize. And I think it's going to reshape across three dimensions. So firstly is, I think the easiest, decisions will be executed faster and more efficient.
You could-- I imagine, at one point, go to a chat bot, and you say, please give a reward to a colleague. And it would happen and not now having to log into the system, put in some points, and click on the button and get approvals. You can just talk to an AI. Please reward this person, and the AI will do that. So I think that is the low-hanging fruit.
The second one I think is it will help the decision makers, which could be employees or managers, making better informed decisions, which indeed will be more intuitive. And I think AI will be able to help people to make better decisions by structuring the information options better. You touched base on a great example.
So, as an example, we've got flexible benefits in various countries. It's common I think in the US, UK, maybe others whereby employees can choose between benefits or compensation elements within a specific budget. At the moment, employees or people need to think through that.
I've got these two options. What shall I take? Shall I go for the car allowance? Or shall I go for the transformation or the transportation pass and probably calculating an Excel file what the advantages are.
I think that, in a few years, you could, if you would want, ask an AI, and it will calculate, what is your commuting time to the office? What is the gasoline price may be doing over the next year? The parking cost maybe, your personal taxation situation if you choose to provide it.
And then I, as an employee, would get very simple saying, Martin, this is the net value in case you're going to go with transportation pass. And this is the net value if you're going to go with the car allowance. And then people say, ah, OK, I choose this one.
So I think the decision-making will be better and much more intuitive. But that will not be the game changer. I think the game changer will be-- I call it the third dimension, which is something that is core to many large companies.
Every large company has policies, strategies, plans, programs, guidance, or rules. And everybody calls these things different. I call them pieces of text.
And every company has pieces of text, and they aim to achieve something. And they do it by influencing the decisions and the actions of the people in an organization and the systems. Now, currently, all of these pieces of text or policies are created by people like me, from senior level to junior level.
And sometimes these policies are very effective in what they want to achieve. Sometimes not. I also think that every company has policies probably that nobody reads, 20-page documents that are hidden somewhere because they are too complicated.
But I think, with AI, these things could be something of the past. So I think AI will be able to support these people in the design policies, in making much better policies or strategies because it will simply understand much more about your organization. It will understand and be able to measure effectiveness.
It will understand what kind of people are in your organization, what typical decisions they'd like to take, how they'd like to absorb information, if they like details or summaries. And I think, with an AI being able to support people to make better policies and frameworks, that should lead to better decisions. And, by better decisions, we should be getting better outcomes.
So I think it will be a game changer. And, honestly, in October when we met, I thought that will be years out. And then I saw what's coming out over the last month and say, ooh, I think it can be much faster than what we realize.
Wow, that's fantastic. That gets me even more excited. And I agree with you.
I also think that the AI will have a chance to reduce the length of some of that content as well. You're saying it's all content. I agree it's all content.
And we actually have a partner that says, every single policy that you write for us has to be basically a third shorter than what you would have it. And we spend a lot of time reducing it to reduce the load on new joiners, et cetera. Obviously, AI can do that for us moving forward, which I think is really exciting, particularly when people are feeling exhausted, not having enough capacity to learn new things. So I think it's part of the solution as well as the solution itself.
And then your comment there about benefits and car versus transport, when open enrollment or benefit selection here in the UK rolls around, I just do what whatever I did last year, not because of any good reason, just because the policies of benefits are pretty long to kind of plow through and read. And to have a partner that maybe knows a bit more about my family history and can combine it with some of that information to make a better decision just sounds wonderful.
The only thing that worried me about what you said, Martin, is you know that I recently wrote a book about working difference. And one of the chapters in the book says, it doesn't pay to stay. And we did a manual balance sheet.
If I stay with the organization over 5 to 10 years, where will my career be? Where will my health be? Where will my pay be? And then if I left, what are the opportunities based on how hot the market is, how my skills are trending?
And now I worry that the tech can probably do that for other people. So if you haven't got your employee value proposition really tuned in to what people matter, it's going to be a whole lot more transparent in the future.
Yeah, no, you're right. And I think AI will help. But, as you, I think also said in the previous meeting, people still make a lot of decisions based on emotions.
And emotions are difficult to quantify, at least by AI. So it's the same-- I remember I bought a car a few years ago. And I came up with an Excel file.
And one car brand was in Excel, let's say, less costly than the other. But I really wanted the other one.
So yeah, you change some assumptions maybe. Maybe I can drive a year longer in that car. And then, hey, suddenly, that car. So I think emotions-- people are emotional people. And I think that will be part of the decision-making process.
Well, the good thing is we still have a role to play. And I think it's actually good to recognize when it is an emotional decision because we're tweaking the facts or changing the criteria. And maybe that'll help us learn a little bit more about our skills as humans.
You, obviously, spend a lot of time playing with these different applications. And I love the word play. So I'm using it intentionally there.
But what about your team? How are they getting familiar? And how are they sharing some of their learning or maybe what isn't working for them?
Yeah, I think the challenge is that the systems are changing so fast. I am-- I'm careful in suggesting people to get familiar with the specific system because, the moment they learn system A, system B will be out there and will be better. So what, for me, is really key with the team is that, again, they experiment, and they find out, what can these different type of systems do in a safe environment?
And, for me, it's not really about-- it's about a mindset change so that they understand how this type of technology can help me with things I'm doing now or could be things in the future and providing a different view. So let me give an example for myself. So now I know what AI can do-- or at least I think I know what it can do.
When I see an activity, whether it's at work or when I walk down the street, it goes through my head saying, oh, that actually could be done by a robot. But I would only be able to have that thought if I would have some hands-on experience, even if it's playing around. And I think, for me, it's really key that we get that mindset.
And what I try to avoid, at least the team that I-- a great team that I'm working with at the moment, is spending too much building the solutions because I'd rather have the team being ready in being able to use the solutions when they come available than start building them. And if I make an analogy, compare it to a century ago. Humans learn that we would have the capacity to build a car.
And we would roughly know what the car would do. You could sit in it, and it would transport you somewhere. And it would be different.
So, instead of having the team or the organization build the car, I think it's better to have other people build the car because they're much better equipped in doing it. And we can then think, well, once we can buy that car or rent that car, actually, what would we need to do to make effective use of it? And can we already learn to drive it.
And can we already think about what we do with it? And I see it the same with AI, saying, OK. Thinking ahead is when these solutions come in, like with Copilot. How can we build solutions that we can utilize?
I love that. I love that. I think it is all about mindset and skill set as you were saying earlier.
We learn from the machine what prompts work. And this machine is learning for us. I love that symbiotic relationship.
And I think you're right. It is about, at this point in time, having sort of a fungible mind where-- rather than fixed systems because a lot of that is going to change. Hey, how about we go across to Lara. Lara, maybe I can ask you a similar question. What advice might you have for a new compensation expert that wants to work with AI?
If you are a new compensation expert who wants to work with AI, here are some pieces of advice that can help you get started. Learn the basics of AI. Before you start working with AI, it's important to have a basic understanding of what is, how it works, and what it can do. This will help you identify the tasks that can be automated and those that require human intervention.
Identify the compensation tasks that can be automated. Once you have a basic understanding of AI, identify the compensation tasks that can be automated. This can include tasks such as data collection, analysis, and reporting.
Choose the right tool. There are many tools available in the market. Choose the one that best suits your needs and budget.
Look for tools that are specifically designed for compensation tasks. Train the AI tool. Once you have chosen the AI tool, train it to perform the tasks that you have identified.
This will help the tool learn and improve over time. Monitor the AI tool. Keep an eye on the AI tool to ensure that it is performing the tasks correctly.
Make adjustments as necessary to optimize its performance. Stay up to date with the latest developments in AI. AI is a rapidly evolving field, and new tools and techniques are being developed all the time. By following these pieces of advice, you can start working with AI to improve your compensation processes and increase your overall efficiency.
All right, Martin, I knew it was going to happen. We are coming close to time. So final question to you, if we zoom into 2030-- and I know as a fellow Star Trek fan, you know exactly what that could feel like. And, therefore, human-machine integration is yesterday's news.
It's the way we wake up and breathe. What do you think will be fundamentally different for HR? And maybe what skills will command a price premium?
Yeah, 2030 feels a bit like that message that you sometimes see in your side view car mirrors-- things may appear further away than they seem to be. I'm saying that because I still remember 2017, looking at AI, and now it's suddenly 2024. But if I would need to make a guess about 2030, which is far and close at the same time, I think, first of all, many corporate jobs will be different.
We've talked about it, such as HR total rewards. And I think AI will just be an integral part of the daily operations, the same like we use email and we use Word. So I think that is one thing that will definitely be there.
And I think what is not going to change is for roles and the skill to understand what the business and your customers need. You still need people that partner with the business or partner with customers and think these are the solutions that may be beneficial. And I think that will stay.
Now I think where things will change in terms of skills and roles, I think there will be an increased demand and requirement for what I call domain experts. And these roles are not only needed to implement and run these systems but I think, also, to determine if these outcomes are actually helpful. And I see it as a human expertise strengthening and validating digital knowledge.
And, as a compensation professional, I can say that if I would ask one of these AI tools to make a proposal for an international move, it comes very convincingly with all kinds of argumentation. But I know, as a compensation professional, it's not correct. Or it's not the right way to do it.
And you only know that if you have expertise in that. So I think the more we're going to get these AI solutions, I think the more important organizations will need to have experts. Now, that brings me to the other side of the coin, which is junior and analytical roles.
I think a lot what analytical and junior roles do today could be done by AI in a few years and definitely by 2030. But if people then start worrying that there will no be analytical roles anymore, I don't think they need to worry. I still think you're going to have these roles, but they will be different in skill sets what they're required to do.
And I think they will be less analytical and maybe more AI-lytical. And what I mean to say with that is you're going to have junior AI-lytical. Roles that are very quickly able to identify which solutions could help them, what data would they need to put in there, and testing it. And would it work?
And these roles we need because you're going to have the domain experts that will retire at one point. And you need to have career paths and entry points for people in your organization to develop that. So I think, in a nutshell, I think you're going to keep your business and customer understanding.
That will be critical. You're going to need domain expertise. And you're going to have more AI-focused solutions that focus on work.
Now, that brings me to, I think, another thing that I just thought about is that I think there's a new skill that I think we need. And I couldn't find it actually in LinkedIn. I'm just going to add it later, which I call imagination.
I think you really need, and you can use with AI and skill, that you think about things that you could not do before. I think what you see with digital technologies such as AI is that, once a person anywhere in the world, at any level, imagine something new and is then able to build it and it will be easier and easier, the impact is exponential. You can build something in a day. And, a second later or a minute later, millions of people can use it.
So I think having people in the workforce, even outside the workforce, to imagine what can they do that they cannot do at the moment and what, I think, will be extremely valuable. So I look with a lot of optimism towards 2030. And I'm happy to jump on a podcast around that time. And I'm sure if we cannot meet by that time, we both have digital twins that can do the interview with each other together.
That sounds fantastic. I love that. I love your thoughts on moving to hyperdrive.
And, when you were talking there about domain specific, it was reminding me, maybe, we'll start to wear red, yellow, and blue jumpers to denote our areas of expertise and knowledge. Fantastic, and I love your heartening comments there about, yes there will be some dislocation in what we mean by an analyst role. But it will move up the value chain, not out of the organization. And I love that comment around AI-lytics. I hadn't heard that before, fantastic.
And, Martin, we could talk forever as always. But thank you so much for joining me today, sharing a lot of your insights and experiments. And, as I mentioned earlier, I know you share a lot of those experiments on LinkedIn.
So, listeners, if you've not connected to Martin Smit on his LinkedIn, please do Because he's very generous in sharing what he's learning about new technology and experiments in the reward space, which is fantastic. So, for me, my big takeaways today is it's all about mindset and inspiration. And you can get that by spending more time watching sci-fi and playing. I think that was part of your advice.
[LAUGHING]
For us, who might be leaders, we need to make sure that we're really clear about what we're asking about-- asking of people as they begin to explore. We need to move away from the theoretical to the practical. If executives digital and HR are all singing from the hymn sheet about the imperative of that, it really makes a difference.
And we've just ended up here talking about, we need to build the right culture where people do have those curiosities and fungible mindsets that are going to set up for success in the future. Very exciting, lots to digest, and I'm excited about that future that you paint for us when our avatars will be getting together. Just, Martin, before you do go, you also read and listen a lot. Is there any podcast that you would recommend to our listeners?
Yes, well, of course, I follow you. I learn a lot from your postings as well. So it's definitely I think a source.
And I think there's another podcast. It's called AI-volution.
AI-volution yes. Jason Averbook, yes. That's a really good one.
That's a great podcast. Economist sometimes has great podcasts. But I think what I would also recommend people is I'm sure that, in every organization and every company, there's somewhere, somebody, could be very junior, that is very passionate about AI.
And find those people, ask them to join a call or join a leadership team meeting, and give them a demo because that is where you'll see a lot of passion and a lot of ideas that will not bubble up. So, yes, LinkedIn, your podcast, my profile, but try to learn and connect with other people in the organization that have these ideas, I think. It was a pleasure talking to you, too.
It was great chatting to you, Martin, I agree. I think the inspiration and energy of the younger people that are growing up with this in the way that they work is going to be phenomenal. And, certainly, for me, that's where I get a lot of my inspiration and my ideas from.
We do have to close today. Martin, thank you so much for joining me. It's been wonderful.
Listeners, if you are interested in some of the topics we've been discussing today or hearing from our experts and leaders, please do visit mercer.com, and subscribe to the podcast wherever you get your podcasts. Martin, thank you again. Listeners, have a great rest of day.
Thank you. It was a pleasure talking to you today.
Great fun. Thank you, Martin.
[MUSIC PLAYING]