Meet the Leader: Why AI 'workslop' is a leadership problem

Robin Pomeroy, host, Radio Davos: Welcome to Radio Davos. This week, a great episode from our sister podcast Meet the Leader.

Alexi Robichaux, CEO of digital learning platform BetterUp, spoke to Meet the Leader on why AI can amplify both good and bad leadership standards and why leaders need to define what excellent work looks like before AI scales up a team's output.

We'll be publishing a handful of Meet the Leader episodes here on Radio Davos over the summer, but you can find more than 200 interviews with leaders of business, politics, academia and beyond by searching for Meet the Leader wherever you get podcasts or go to wef.ch/podcasts.

Radio Davos will back next week, meanwhile enjoy this episode of Meet the Leader. It's called Why AI 'workslop' is a leadership problem, here;s how to fix it. It's hosted by my colleague Kateryna Gordiychuk.

Alexi Robichaux, BetterUp We see with WorkSlop, where the human thought input is poor, that AI can only do so much.

AI is an amplifier. If you're very clear, if you're thoughtful, If you're high agency, then it can produce really crystalline work. If you are not, then that will also amplify what you see there.

But now it's on steroids. It's really not a limitation of the tool. It's actually a limitation on the human

So when we study work stop, we're really looking at systems and cultures and are there clear definitions of success, clear standards of excellence, and are managers providing that feedback to their people?

AI in many ways challenges us to be better versions of ourselves.

Kateryna Gordiychuk, Meet The Leader: Welcome to Meet the Leader. My name is Kateryna, and today in our studio, we have Alexi Robichaux, from BetterUp. He's here to talk about all things of AI and skills. Let's jump right into it. Tell us about BetterUp. What's so special about this platform?

Have you read?Workslop: Why AI performance depends on how we think, talk and leadRebuilding trust for the age of AIWhy workforce sustainability has become a board-level financial metricThis is the secret to management in a fragmented worldWhat is AI coaching and what problem does it solve?

Alexi Robichaux, BetterUp: The inspiration for BetterUp came from my own journey and the journey of my friends. We were all young professionals at the time, and we were looking for ways to upskill ourselves, to realize our potential in the workplace.

And we realized that this was something that so many workers worldwide were trying to do.

And we call today, what we provide at BetterUp, the human transformation platform, which is a combination of digital learning, coaching and AI that allows people to realize their full potential at work, by providing them with skills that range from leadership development and leadership effectiveness, to preventative mental health, to really building teams that are multifaceted and able to tackle any challenge.

So, we partner with very large employers and we're able to deploy across the workplace – providing everything from frontline hotel workers, to managers at tech companies, to servicemen and women in armed forces around the world. A personal, always-on experience that allows them to get the help and support they need.

We were really fortunate to be one of the pioneers of virtual coaching. We have the patent on a virtual executive coaching.

For so many years before BetterUp, coaching in the workplace was synonymous with executive coaching. Outside of that word ‘executive’ being in front of it, it really wasn't a known thing.

And part of our vision was to democratize access to what we knew from the research was a highly efficacious way to help people learn, grow and change their behaviours – which was why companies spent so much money at the top of the house, investing in their top executives, so they can get better. And we wanted to bring that to the entire house, to every level of the organization.

AI coaching provides the opportunity for people to get help at literally any time and any place, but that actually helps train people to the benefit of partnering with someone – whether it's an agent or a real human being – and we see that then people have more of a desire to partner with the human coach.

How can AI coaching complement work with a human coach

Kateryna Gordiychuk, Meet The Leader: I think I have a pretty good idea of what working with a human coach would be, online or in person, but when I think about AI coaching, I don't think I yet can grasp really the idea. Can you walk me through if I were to benefit from AI coaching, what would it look like?

Alexi Robichaux, from BetterUp: A lot of the models you're used to, the primary interface would be chat. That seems to be how users like to interface. But you can do it with audio. But it's a very naturalistic conversation.

When we zoom out and look at our data set, which is hundreds of millions of coaching interactions and moments, what we find is people tend to almost intuitively know to engage with AI coaching in a more focused, and almost transactional, way. So it's very good for practice. ‘I'm about to have a tough conversation, what should I say?’ ‘Or how would these talking points work?’ ‘I'd like to role play something.’

Role playing is one of the biggest activities we see on the platform with AI coaching. And that's not to say you can't do this with a human coach, but you can just call a coach for five minutes and have a great conversation, right? You can message a coach, but there's some latency – they're human beings, they have other clients, it may take a few hours for them to get back.

And so, we find that human coaches do really well on these things related to value, identity, long-term aspiration of goals and transformation. And AI coaching is actually really good for the here-and-now tactical help, which is a lot of management and leadership. ‘How do I...?’ ‘Give me some general guidance on how to have the tough conversation I'm about to walk into.’ And so we find the humans are sorting. In fact, we see that about 34% of people only want to work with a human coach, 15% people only want to work with the AI coach, and 51% want to work with both. The vast majority of people understand intuitively that there's a place for both and I want to figure out when to use one or the other.

Kateryna Gordiychuk, Meet The Leader: That is a very interesting idea, and I don't think a lot of people would know this, that there's this place and time and specific type of coaching for different kinds of purposes. And perhaps people who are a bit sceptical about AI, or perhaps worried about other people seeing them in a negative way, could tailor their needs based on what it is they want to be coached about.

Alexi Robichaux, BetterUp: You also find an inverse thing. There's a lot of research on what's called disclosure and communication. And for many people, the less personal the form of communication, counter intuitively, the more comfortable they are being vulnerable. They might have been uncomfortable or intimidated to share some of their concerns or worries of vulnerabilities with a human coach, but they're actually very unencumbered with an AI coach.

“

If you're not open with your therapist, if you're open with your coach, it's really hard for them to help you. And it's the same with AI. The more vulnerable you are faster, the more you're going to get actual helpful feedback, and helpful guidance

”—Alexi Robichaux, CEO, BetterUp

They're more open and more honest out of the gate, which means whether it's coaching, therapy, any behavioural intervention, being open and honest is the prerequisite to actually seeing change. If you're not open with your therapist, if you're not open with your coach, it's really hard for them to help you. And it's the same with AI. The more vulnerable you are faster, the more you're going to get actual helpful feedback, and helpful guidance.

In a way, it's a really powerful tool of inclusion, because you have people who may otherwise have been hesitant, who now feel very safe and comfortable engaging in that way.

Kateryna Gordiychuk, Meet The Leader: I'm not sure you'd have an example on top of your mind, but I'm curious to know what are some cases where an AI coach really helped somebody through a tough situation, or something that they're struggling with, and they need to address – for maybe people or companies who would consider hiring an AI coach.

Alexi Robichaux, from BetterUp: Yeah, so I look at across our customers who, most of our customers… one of our advantages, we're one of the only platforms that provides all of the coaching, right? So, whether it's a human, whether it is AI, we do it in an integrated manner.

A very common one we see, we support a lot of frontline managers in our population across customers is performance review conversations. So, these are famously difficult. You know, we spend a lot time on perfecting the performance review, which has some validity and merit, but the conversation around the performance review ends up being much, much more important, and it tends to be where managers are the least developed.

And so, we find that this is often a first use case that HR departments, in partnership with us, will roll out to AI coaching, and we have custom role plays that are built as AI simulators to help managers do this, and they're one of our more popular activities.

And it's because these are very high-stakes conversations. Most managers, they care deeply about their team, which means they want to give them the feedback, but they also want to do it in a way that motivates future performance, and sometimes there's critical feedback in there, and that's a really hard balance to strike. And it is one of these things that really just takes practice, right?

I think one of the beauties of coaching in general is it is a practice mechanism. Most of our corporate lives, most of our life as leaders, we aren't like athletes. We don't get to practice and then go play a game once a week. We're always on the pitch playing the actual game. And AI provides a really safe, almost magical form, to be able to practice and get this type of feedback. And so, we see performance conversations is one of the emerging use cases as highest value for AI coaching.

“

As leaders, we aren't like athletes. We don't get to practice and then go play a game once a week. We're always on the pitch playing the actual game. And AI provides a really safe, almost magical form, to be able to practice and get this type of feedback.

”—Alexi Robichaux, CEO, BetterUp

Kateryna Gordiychuk, Meet The Leader: Alexi, I could imagine somebody playing devil's advocate and saying, for example, through these tough conversations with real people, with a manager, with somebody you might not particularly like at work, or with a challenging manager, you are learning as a professional, you're learning as person, and next time this skill is more developed. But because there isn't this personal component with AI coaching, perhaps that could be lost. What would you say to that perspective and what can you offer in this situation?

Alexi Robichaux, from BetterUp: Yeah, I mean, I think in like, as a thought experiment, that might seem plausible. In reality, if you're a manager, I don't think AI is substituting for the human interaction. You have a team, you still have to interact with them. Let's take the performance review. No amount of practicing that conversation with an AI replaces the fact that you still have to go tell your direct report how they're performing.

I think what's new and helpful here is that you actually get to practice. And I think, you know, for most of us growing up as managers, executives, we were practicing, you know, with live ammunition to use the analogy and people were getting hurt along the way.

“

For most of us as managers, executives, we were practicing, with live ammunition -- to use the analogy -- people were getting hurt along the way.

”—Alexi Robichaux, CEO, BetterUp

And so yes, we got better at giving feedback, but that direct report we may have not done a great job, we could have really discouraged them or not given them the most constructive form of that feedback. So, I actually think of it more as augmenting and providing a new gym for people to practice in. I don't think to date we've really seen it replacing from the human-to-human interaction yet.

Kateryna Gordiychuk, Meet The Leader: If you're looking into the future in five years, 10 years, where are you hoping AI coaching would lead us, both employees and managers, and maybe it being scaled across the world?

Alexi Robichaux, BetterUp: Yes, great question. What we're building to is, I think it comes back to what do we know about how we learn as humans? And what we know how we as humans learn is: we learn very well in context and we learn very well in the moment. And so the sooner you can get learning to the actual event that transpires and you can get that feedback in situ, contextualized near real time, the better we are at incorporating that learning and practicing.

We tend to forget as humans. So, when you play forward with AI, I think that's one of the biggest promises AI can provide in terms of coaching is it can be ubiquitous. It can be embedded in the flow of work. It could be always on and it doesn't forget, right? Its memory does not forget like human memory. If you think about… we are a massive data set as a human being of all these interactions but it's so easy to selectively remember this or that and it's not the full picture of who we are.

“

AI can be a really healthy looking glass to say, ‘Hey, you think you're showing up this way, but you're forgetting these seven interactions with your team where you showed up another way.’

”—Alexi Robichaux, CEO, BetterUp

AI can be a really healthy looking glass to say, ‘Hey, you think you're showing up this way, but you're forgetting these seven interactions with your team where you showed up another way.’ And so we're building to: how do you embed coaching where it's just running in the background? It's something that's always there for you to turn to if you want that benefit and help. But it's not something, it's a destination anymore, it's really integrated into the flow of work and how you learn. And we know from the research and science, like that is the ideal way that humans like to learn.

What is AI 'workslop' and how leaders can prevent it?

Kateryna Gordiychuk, Meet The Leader: Another aspect of BetterUp's work is the research that you do on productivity and AI. And I found it completely fascinating, the notion that you did a whole report on, which is called a ‘workslop’. And I'm sure a lot of people have heard of it, a lot people have received it. What is workslop? Why is it so striking? Why is that a part of our conversation today?

Alexi Robichaux, BetterUp: Sure, I'll start with what it is. I mean, we've all seen that workslop is bad work produced by human AI collaboration, at its core. The hallmarks of it are often repetitive prose, long-winded rants to say things that otherwise would have been simple.

The promise of AI is it should help you be more clear, concise, cogent. But what we're learning, just like human collaboration, is that if you approach AI in an unthoughtful or unconscious way, it kind of reciprocates and it plays ball with that. And so, workslop, really when you get to the research, we find 40% of workers have received workslop. The average incident of workslop takes someone else two hours to unwind or correct. So, in that sense, it's counterproductive.

Have you read?Workslop: Why AI performance depends on how we think, talk and lead

And I wanna be clear, AI is a huge boom for productivity. We're big believers, we're an AI company ourselves. So, we're not saying that. But what we are saying when we look at it is workslop is not the problem, it's a presenting symptom of unconscious leadership and management behaviours.

When people don't have role clarity, when they don't know what excellence looks like, when they don't know what the standards are, AI is an amplifier, and it amplifies these behaviours.

So, the same person who sends you workslop, had they just done it by themselves, it still would have been slop. It still would've probably not been great work, but now it's on steroids. And so what we find is, AI is a really powerful exponentiator. If you're very clear, if you're thoughtful, if you're high agency, then it can produce really crystalline work.

If you are not, then that will also amplify what you see there. So, when we study workslop, we're really looking at systems and cultures and are there clear definitions of success, clear standards of excellence, and are managers providing that feedback to their people.

Kateryna Gordiychuk, Meet The Leader: It's very interesting because it really showcases that it's not just about how AI is structured, but our own habits that we bring to using AI that affects the outcome. Why is it important for users to also reflect on the way that we use AI and learn from it all the time?

Alexi Robichaux, BetterUp: I mean look, AI is a tool. I'm a technologist, like all technology, it has a human input at some level, and I think that's important. We're not to a point yet where AI replaces human thought. I would say it amplifies human thought and accentuates human thought.

And so we see with workslop, where the human thought input is poor, that AI can only do so much. It's really not a limitation of the tool. It's actually a limitation of the human, and so I think AI in many ways challenges us to be better versions of ourselves.

“

We're not to a point yet where AI replaces human thought. I would say it amplifies human thought and accentuates human thought

”—Alexi Robichaux, CEO, BetterUp

I grew up as a product manager in Silicon Valley and there's probably no role, in many ways, that proves this more than product management today. You used to spend so much time as a product manager writing requirements, and figuring out what to build and how to build. Now you can vibe code and just bring ideas to life. And what it's really doing, in short, it's showing which product managers have product vision, and great imagination, and great ideas because bringing them to life, at least in a demo, is easy. AI will do that.

Whereas before that could be very lost in a long process and you might not know that for quarters or even years sometimes. And I think that's just one example where so many parts of knowledge work with AI, it’s really exemplified or highlighting what we see in the research. It's about people's core mindsets, their creativity, their agency, their optimism. These things are shaping the outputs that AI is amplifying.

Kateryna Gordiychuk, Meet The Leader: You use the term vibe coding, which I don't know if everybody knows. I mean, we're here just decrypting all the keywords today. What is vibe coding? How has it changed also your job as a leader, as a manager? You know, it's something that entered reality so swiftly. And how do you expect it to change in the years to come?

Alexi Robichaux, BetterUp: Yeah, I mean, there's probably competing definitions of vibe coding, but in essence, it's being able to programme with non-programming inputs. So being able to, even if you're not a programmer, take something like Claude Code and in natural language say, build me a website, or build me an application that does this, this, this, and the tools are phenomenal today.

I mean I grew up programming and it was a skill set and it gave you an advantage if you could make a computer do what you wanted to do. Most people can't do that prior to vibe coding. Now, really almost anyone who has the curiosity, and wants to spend a few hours learning it, can become rather functional in building websites and basic mobile applications.

And so, it's just a huge potential creativity explosion for humanity. If you think about the rate limiter to going from an idea to an app on your phone was developers, historically, and that rate limiter is no longer the case. Anyone can just develop, anyone can easily deploy to the App Store now. So, you think about there's so many good ideas in the world from people who couldn't get access to engineering talent or developers, that are quickly being abstracted – which is exciting in that we should have more great ideas but it's also showing in knowledge work that production is not the hardest part anymore. In a lot of us, it's actually the thinking that's the hardest, and it's raising the stakes on the quality of inputs and thoughts.

How coaching challenges a leader's assumptions and blindspots

Kateryna Gordiychuk, Meet The Leader: Alexi, with technology and AI changing so fast and these concepts emerging all the time, I mean it's hard to keep pace. Have you ever hit a wall as a leader? And if so, how were you able to get out of that mindset and continue on?

Alexi Robichaux, BetterUp: Yeah, I think, I think every leader does. I'm fortunate I have a lot of coaching at my disposal, and I think that's, for me, often what coaching is. Coaching is helping to remove a writer's block, or a wall, or just when you feel like there's a dead end. You're just like, I feel like I've exhausted all my options here. What might I be missing?

And I think coaching can be really powerful on… you might not be missing any facts, you might have read the situation correctly, but you may be making assumptions about what those facts mean. And a lot of what coaching does is challenging those assumptions to say, hey, with the same facts, is there a different interpretation or different story?

Have you read?Rebuilding trust for the age of AI

The reality is, you know, whether it's coaching, or CBT [cognitive behavioural therapy] and therapy would be a great example of this. At their core, they are about story. What we've learned from 100 years of psychology research is that people operate based on the stories they tell themselves. And when you change someone's story, you can change their behaviour. Over time, you can change who they are and how they show up in the world.

And so often we hit a dead end because we don't know what's the next chapter in the story. And coaching is a really powerful device to help you discover, help me discover, ‘Hey, there's a chapter I'm not envisioning here that I can actually imagine and make happen.’ And that's one way I think about getting past that writer's block.

What soft skills can help leaders thrive in an AI age?

Kateryna Gordiychuk, Meet The Leader: I was thinking about this human-AI collaboration and how we will all have to be kind of on top of our game, really, if we want to catch up with AI. What are some human skills that you think everybody should really zone in on and make sure that we perfect, maybe with the help of coaching, AI coaching, that you're thinking the next 10 years we cannot really live without?

Alexi Robichaux, BetterUp: Yeah, I think what our data is showing already is, you're seeing this premium emerging on what traditionally would be thought of as like the softest of the soft skills. And again, it's because so much of the work we're doing, and the collaboration we're doing, the computer is starting to take over and really be very good at.

So, when we look at who's really thriving in an age of AI, it is people with measurably higher levels of optimism. Optimism is a skill. When we talk about getting through roadblocks and setbacks, having an optimistic mindset, it's not a personality trait, it's a skill that can be learned. A resilience, the ability to bounce back from setbacks. Again, not a personality trait, a skill can be learned and taught.

But these feel really soft. Agency, not personality trait. Something that can be taught and trained. These things are emerging in the research as really, really important. And the reason why is because where there's a will, there's way. And that's always true, but that's 10 times more true, exponentially more true with AI, and so what you find is the knowledge workers, the workers in any sector who are taking advantage of AI, they're more curious, they have an internal locus of control, they're higher agency, they're high optimism, they are more resilient, and that enables them to continue to iterate and learn faster than everyone else.

“

And it really is a race right now, it's a speed of learning race. If you're not learning, you're already behind.

”—Alexi Robichaux, CEO, BetterUp

And it really is a race right now, it's a speed of learning race. No-one knows what the answer is, all this stuff is happening so fast, no-one can keep up with it. But if you're not learning, you're already behind as these cycles continue to compound. If you think about compounding interest rate, if you're compounding your learning along, then you are progressively getting more and more behind.

And so you have to ask who, what skills make for the best learners, and those are really the psychological skills that we're seeing. They matter more than hard skills, they matter more then, you know, traditional communication skills, things like this. Underneath it, it's those skills that matter the most.

Kateryna Gordiychuk, Meet The Leader: And they make us humans, right?

Alexi Robichaux, BetterUp: They make us human, that's right. And I do think that's the thing with AI is like, most of us are not great at being human. I mean that's like the sad truth. Like our whole business is, I started BetterUp because I was like, I have some huge deficits in these human skills. And so I think what AI is doing is saying, there's a really powerful role for humans. But we have to get good at what makes us uniquely human.

Kateryna Gordiychuk, Meet The Leader: It is such an interesting way of putting it, because people are afraid of AI, they're afraid that this humanness will be, this humanity will be taken by AI, but in fact it maybe is showing us what's unique about us, right?

Alexi Robichaux, BetterUp: That's how we view it. That's what I think.

Kateryna Gordiychuk, Meet The Leader: Is there a book that you would recommend to our listeners and viewers?

Alexi Robichaux, BetterUp: Oh, so many. I get to work with a lot of great researchers and authors. I think one that just came out, which is phenomenal, is by Brené Brown. She runs the Center for Daring Leadership at BetterUp, and it's called Strong Ground.

And it is a book really on organizational and leadership transformation and how the two intersect. And the central theme is this theme of grounded confidence. That's really something, it's a skill that we can learn to be centered, to be grounded, to be poised, in all these situations is when the swirl is going around us and how do we lead others in that way? And it starts with how we lead ourselves.

What leaders should prioritize now

Kateryna Gordiychuk, Meet The Leader: And finally, looking into the future and being here at the Annual Meeting now with so many leaders here, what should leaders prioritize, do you think, in 2026 and in many years to come?

Alexi Robichaux, BetterUp: I think what leadership prioritizes is people right now. I think it's hard to find a company that isn't prioritizing the technology part of AI transformation. But I think most companies are still willfully under-invested in is the talent part of human transformation, or AI transformation, the human transformation part.

And what we're finding, and what you're seeing, in these early transformations is the adoption's not where we want it to be. The productivity gains are certainly there, but they're not where we all thought they would be. And when you really dig in, it seems to be through multiple studies, the rate limiter is humans.

It's these human skills. Humans are resistant, or humans aren't taking fully advantage of the new technologies. And I do think that people are at risk of feeling left behind, and in some sectors and some jobs, truly being left behind by technology.

And so, I think one thing we should do as leaders is really say, ‘Hey, as much discipline, rigour, investment, ROI modelling as we put into the technology build-out of these AI transformations, which is enormous, do we put any even 10% of that into the talent build out? And the answer for most companies is probably not.

And so, if we can shift that mix, and it's like, hey, for every $5 we're spending on technology, are we spending a dollar on upskilling our people so they can take advantage of it? Are we spending $1 on org design to figure out what the new roles and responsibilities are? Or are we just kind of dumping this on the old org chart and expecting the same people to operate in new ways?

The reality is, there's never been a technology transformation that's pronounced that anyone alive has seen, and even in prior technology transformations, there was a lot of training to teach people how to use electricity, how to use the telephone, and we've kind of forgotten that there were huge societal efforts to move... farmers and people outside of urban centres along in some of these shifts where they didn't have technology and we haven't yet seen that level of societal push, around the AI. It's mainly been early adopters self-selecting it, and then companies investing in it.

Kateryna Gordiychuk, Meet The Leader: A lot of food for thought about how to do it more structurally. Thank you so much, Alexi. That's all the time we have today.

For more video podcasts, please go to World Economic Forum's YouTube page. And for more podcasts, go to wef.ch/podcasts. Thank you, so much.