How to upskill for an AI Age: Workera CEO Kian Katanforoosh

Oftentimes, actually, when we look at metrics that we care about within an organization, we care more about how fast people are growing their skills than their actual skills.

Welcome to Meet the Leader, the podcast where top leaders share how they're tackling the world's biggest challenges. In today's episode, we talk about upskilling for the AI age, common blind spots that hold teams and leaders back, and the simple shifts that can make a real difference. Subscribe to Meet the Leader on Apple, Spotify, and wherever you get your favorite podcasts. And don't forget to rate and review us. I'm Linda Lacina from the World Economic Forum, and this is Meet the Leader.

Oftentimes, you realize the manager has never been trained at evaluating people's skills. Very few managers actually know how to evaluate people's skills well. And so without understanding the skills of their people, managers are assigning the wrong tasks. They're coaching them the wrong way. And then as a consequence, a portion of the population that you did not want to leave ends up leaving.

It is a new year. Many of us started out 2026 with resolutions, big visions for ourselves, how we can be 5% better. I'll wager that nearly all of us started out the year with a host of OKRs to hit and formal goals, goals that will shift and move as the world does, the next big tech change or economic shift. Many of the biggest goals companies have set will take serious upskilling, but improving individual skills or team skills isn't simple or straightforward. We don't always know what we need to improve, and we might have an inflated or an outdated idea for what we do well. Kian Katanforoosh understands this problem well. He is the CEO of Workera. That is a skills intelligence platform, and their technology measures skills at scale. It can help teams understand how ready they are for things like an AI era. Kian has a front row seat to what leaders and teams need to shore up. He talked to me about the blind spots he sees, the roles leaders will need to play as tech shifts quickly, and the skilling framework that uses in hiring and with his own company. He also shared with me what prompt engineering can teach leaders about things like delegation and why he sees a future where one day it's unethical for humans to judge other human skills. He'll talk about all this. But first, he'll give leaders a grade on how well they're bridging the skills gap.

They're bad. I would give them a 2 out of 10. And there's a reason behind it. People, including leaders, are extremely unaware of their own skill sets. People have a tendency to overestimate their skills by a significant margin or underestimate their skill. In our data, in fact, we show that 89% of people who take an adaptive assessment on Workera after self-rating themselves are off by 20% or more on their final score. And that's where the 2 out of 10 comes from because only about 10 to 20% of people get it right.

And why is it a problem?

When you're a leader and you are pretending to have skills that you actually don't have, your behavior cascades down your organization. And so we find that organizations with such leaders are highly overconfident, which then creates issues with a lot of unknown unknowns. You don't know what you don't know. There's a lot more opportunity to adopt AI to display different mindsets or behaviors that you're not doing and that is hurting generally the, the the company or the organization as a whole.

Correct me if I'm wrong, their staffs also individually might not really have a correct assessment of their own skills. So if you're trying to develop someone and they disagree with you on what they might need to fix or not fix, you've got sort of the blind leading the blind there. Yes, would that be correct?

That's correct. The staff also is widely off. The one caveat for a leader is that the leader has a lot of influence. And so, an overconfident leader will lead to an overconfident organization. And an underconfident leader, the same way, will lead to an underconfident organization.

And what makes this even more difficult to fix, right? We've got just sort of the mismatch of sort of awareness, right? But what else makes this extra sort of complicated sort of bridge the gap?

What makes it hard to bridge the gap is the skills are changing fast. So you might actually have certain skills and realize a year later that you're already behind by a significant margin. 40 years ago, the half-life of skill was over a decade, meaning you can have the same skill for over 10 years and you'll be safe. You can continue to use it. Today, it's a different story. Skills go outdated within a few years, 4 years on average and 2.5 years for digital skills. I let you imagine AI skills, how short the half-life is today.

Tell us a little bit about Workera and the problem that it's solving.

So Workera's purpose is to build the best measurement of human potential that we've ever built. Measuring skills is a very complex problem. It's a problem that needs to be done fairly and validly. We really need a system that can evaluate us in a precise way and give us insights out of it so that we increase our self-awareness. And most importantly, we understand the actions to take in order to change. What do I need to learn? What mentor potentially is going to help me? What projects are the best for me? What team am I going to fit into? All of these require a good measurement or a good understanding of people's skills and potential.

Is there an example of a change that Workera has helped an organization make as a sort of a before and after that you were able to give them this assessment and then they were able to sort of take that forward and have an impact?

Absolutely. I'll mention a few. A lot of organizations have bought AI softwares. They spent a lot of money in it and As McKinsey used to say, with every $1 invested in technology, you need a dollar invested in people. And they then went over it and changed it into, for $1 in technology, you need $3 in people. I think the truth is even more because today we're seeing a lot of our clients struggle to get adoption on their AI tools. And they ask us to come in, they ask us to run our AI readiness assessments that are customized for the organization and understand where do we sit. In general. Who are our champions? Who are the people that are going to amplify adoption of AI around themselves? And who are the detractors or who are the people that don't have the skills absolutely and we need to help and handheld in order to adopt these tools? We are able to answer this question and provide very guided or personalized roadmap for people to get to the level of skills where they are effectively adopting these AI tools. That's very common. In other areas, companies really want to make talent decisions with fairness and more meritocratically. They oftentimes call these efforts skills-based organization. So think about an organization that moves people within in the right projects, in the right teams, gives the right opportunity to people who are deserving of it. We are often the backbone of this, meaning we evaluate people, we give them feedback, we allow them to have badges, we allow them to become mentors, we accelerate projects, all of that starts with a good skill measurement typically.

There's always been a gap even without the sort of the influx of AI, but there's a gap between what talent wants, what employers are looking for, but now also sort of what the market actually needs.

We actually know from our surveys that when you ask leaders if they're ready for the future, they will tend to be way more optimistic than when you ask actually their workforce, where the workforce will usually say our leaders keep talking about it, but we are not being supported to the level we need. So there's there's definitely a gap between leadership and their employees on these topics of AI or future capabilities. How do we bridge that? I think when it comes to an assessment, understanding where we stand is very important, not only at the level of the workforce but also the leadership. So the best behaviors I've seen is when a leader actually shares their skills profile and says, look, we're all in this together, I don't have every skill. In fact, I'm lacking skills in critical thinking and problem solving, and I need to work on my resilience. And they show that to their workforce in order to get buy-in, create psychological safety, and make it clear that it's a priority for everyone, not just the workers.

When I have these interviews and people say, "Hey, you know, you gotta be humble and you've gotta really show people that you're improving and you're really working on these things just like everybody else," right? But in the everyday, that's hard. It's really, really hard. And there's a lot of reasons why maybe somebody might wanna be a little bit more close to the vest or they might be hesitant to be vulnerable like that. What is needed to sort of convince leaders, even if they think that they are being vulnerable, to sort of take that extra step and say, look, you know, here's where I am, here's my actual development opportunities. How can we make sure that people feel comfortable, leaders especially, saying they've got stuff to work on?

So first, we have to make sure that the measurement is fair and accurate so that there's trust and agreement that, you know, if we're showing a certain score in a certain area, that it's meaningful. Second thing, Humility is part of a leadership skill, and I think you need to be able to be humble when you need it, inspiring at the same time when you need it. I also want to make it clear that skills is only one component of performance, and in fact, not having certain skills is not correlated or equates being a poor performer. Some of the best performers in organization have critical skills gaps in many areas. Oftentimes, actually, when we look at metrics that we care about within an organization. We care more about how fast people are growing their skills than their actual skills. Meaning that at the end of a quarter, we might look at your learning velocity. Where did you start versus where you are today? And that information turns out over time is way more valuable than where the person is today. Understanding that projecting their learning velocity a year from now, here is a worker or a leader that is adaptable, that is going to be way more skilled in a year from now than they are today.

Do you think that the skills conversation sometimes goes a little sideways where people think, well, I just get the skill, you know, just like I get a merit badge and then I'm done? What do you think?

Yeah, I think skills in the last decade have been misrepresented, misused. They've been very generic. And so the first thing I recommend is making sure that whatever skill taxonomy or ontology you are looking at is directly connected to the business outcomes. So a company starts with, "What do we care about?" Goes one level down, "What are the skills needed in order to do it?" And then, "Let's measure those skills." And based on the results of those measurements, let's determine what are the course of actions to take in order to close the gaps. If you follow this process, the skills that are represented, measured, are the skills you care about, and there will be outcome tied to them. Develop the skill, you'll get reward. Also, I always recommend multiplying the carrots for the workforce, meaning answering the question, "What's in it for me? Why should I develop that skill?" You should develop that skill because you'll get rewarded for it. Some companies reward in cash, other companies reward in a brown bag lunch with the VP, others create other types of incentives like becoming a project leader or a mentor to other people, and those are great career developmental opportunities that the workforce is hungry for.

You mentioned this taxonomy and sort of identifying what's needed. How do you apply that same structure in that skill development at Workera?

All our workforce is certified with our own AI assessments. Even beyond AI, each function within the organization has a set of custom assessments that the employees take. They know where they stand today, they know their target score by the end of the year, and they're able to work through it at their pace. On top of that, In our interview process, we leverage skill assessments. We also ask a lot of questions that I've started asking way more often than before, such as, what's your learning strategy? How do you keep up with progress? And what I'm looking for is someone who's going to join us today but grow very drastically with the company in the coming years. I sometimes ask people to teach me what they learned over the last 3 weeks, you know, something they learned. Make it easy for me because it tells me that they've learned something. And on top of that, they've learned it well enough that they can teach it.

Given the assessments that you guys have, what's a skill that you are looking to improve on?

I'm working on a variety of skills. Delegation is a skill that I'm working on, understanding who to go to for what, and again, the way I communicate when I delegate a certain task. I've also worked a lot on leadership skills in general, including creating trust within leadership teams. I think trust is often one of the most important aspects in order to build a high-performance team. Do people trust each other? Are people able to focus on their own tasks and trust that someone else is going to take care of the rest? And all of it coming together to create a high-performance environment. That's been something that I've worked on extensively recently.

And it's something that, of course, it's a moving target for all leaders. When it comes to building trust, what's something that you are sort of learning to do or putting into your repertoire that maybe wouldn't have occurred to you before you founded Workera? Like, what's something that you do now that is really helpful in building trust?

The most helpful thing might be that I think before speaking. In the early days of Workera, I used to be very real-time, meaning I would constantly be messaging, I would constantly be speaking with people. And it has benefits because people know what's on your mind. Pretty much at any point in time. But over the course of the last few years, I realized as the company grew that what I say has a different weight to certain people. And so I might have a thought that took me just 20 seconds and I thought it was interesting but not super important, and I just mention it. And suddenly, because of the weight it had on people, they might be distracted for hours at a time or for days at a time thinking about this problem that I posed in front of them, even if they had more important things to do. And so I've started to be more thoughtful about what I share and how do I say it in a way that makes it clear. Is this an important and urgent problem for me? Is it unimportant but urgent? Or is it neither of those?

And how did your assessments surface up the opportunity to develop more on trust?

Yeah, so for example, when I took our effective communication assessments, I had a gap in the understanding the Eisenhower Matrix, which talks about something being important urgent versus urgent but not important versus important but not urgent. And that framework, I learned it through an assessment. I actually didn't know I didn't know it.

And what was your reaction to that?

My reaction was, wow, it's a great framework. I wish someone had taught me that framework earlier and that I didn't have to learn it so late in my career.

You also mentioned delegation, which is, you know, something that of course every leader has to sort of constantly work on, especially as sort of needs and priorities change. What is a way that you have also sort of changed how you are either approaching delegation or how you, um, actually accomplish it?

Part of what changed is now when you delegate a task, you need to think, is this a task you wanna delegate to an AI agent or is it a task that you delegate to a human? And oftentimes the answer might be a human. Other times it might be a system that you're aware of and you know is gonna get the work done. We're still in the early days of agents being able to do enterprise-grade tasks end-to-end, but I anticipate that it's going to happen sooner or later. So when it comes to delegation, I think the words that you use matter a lot. Like when you delegate a task, you want to make it very clear what you're delegating, when you need it by, as well as what's the reason or the motivator behind it in your head. And I think earlier, a few years ago, I didn't really think about those things. I would just give a task and realize that, wow, I didn't define it well. I didn't explain the rationale to the person. And as a consequence, it went sideways. Today I do it better, and it turns out it's also a skill that's used in prompt engineering extensively. Maybe not the time you need it by, but when you actually configure an agent to perform a task, you wanna be very clear with what you need it to do. Otherwise, it's just not gonna do it the way you want.

You mentioned that some of the skills in, in prompt engineering are, you know, very useful in and delegating to humans or to AI, right? There, those, some of those skills are useful in both situations. In your opinion, will managing AI agents help people be better at managing in general?

So prompt engineering is a trend. A lot of people think about it as a role or a job, and I think that's the wrong way to think about it. Prompting is a skill that any person needs to have, just like you need to know how to search on Google, you need to know how to prompt. And how to get the AI to help you complete a task. Prompting is gonna change over the next few years, and you're not gonna need certainly the craftiness that you need today to get the AI to do certain things. I was actually at a dinner with a friend last week where we played this game where you take a picture of your friend and you ask the AI to guess their spirit animal. And it turns out that he tried and he took a picture of me and he couldn't get it to work. I took his phone, I took a picture of him, and I knew how to prompt the AI so that it does it. The AI would refuse originally to do it just to not hurt the person, but it turns out that if you know prompting well, you can get around it. I think those are problems that are still the case in a lot of job tasks. People who know how to prompt and are crafty are going to get the job done better. I think a year from now, 2 years from now, the models are just going to be so good at understanding your intent that it's not going to make that big of a difference if you can be crafty or not, unless you're a builder of AI system where you need that level of craft. To your question around, are the skills of a manager of humans the same as the skill of a manager of agents? I think they're very similar, although human management has a much deeper emotional component where you need to care about how your message resonates with the person, not only in the content of the message but in the form. And in fact, we joke that you should not thank ChatGPT or AI because it uses unnecessary tokens and more energy and more compute. You should do it with a human though, because it is not useless. It tells them that you're grateful and that might have an impact on the way they complete the task or the way they interpret your feeling about their work.

We can all get better at maybe refining what we want from somebody. You know, here's the parameters, can you have it by this date? I want it to be like this and that. Would you recommend people practice with an AI to sort of practice at kind of refining what the thing is that they want?

Yeah, absolutely. Just practicing to make sure you're precise with your tasks and reuse those skills with humans as well. Just add a little bit of emotional connection to it. Also note that oftentimes, and it might be skills or generational, but some people treat AI as a search engine. They just search for information, they get a reply. Others use AI as a set of tasks that they want to complete. And finally, some use AI as a complete operating system, as one of my, my friend Omar Bawad says. And that's what we should all aspire to do. I try to practice that. And if you look at me within Workera, almost every information that gets to me somehow passes through AI. And any information that goes out from me also passes through AI, and that's just how it is. On top of that, AI is getting better at more difficult tasks. So what you might want to try is give AI a task that you think a human could do in 30 seconds, and then try to think of a task in your daily life that might take an hour and see if AI can do it. And then as you go through that, you're going to understand the reasoning capabilities and AI needing more time to complete certain tasks and today being more prone to errors, but tomorrow AI is going to get, uh, better at it. And what I advise people is to practice because it will give you a feel of what AI is capable of. And once you have that feeling, you'll know better when to use AI and when not to.

I wanna talk a little bit about mentorship. You guys have a special tool for that, which we will get into in a second, but I wanna talk just generally. You had a very special mentor at Coursera. Can you tell us about that? Why that was so shaping for you?

Yeah, I was lucky in my life to have a few great mentors that shaped how I think and the things I aim to achieve. One of them was Andrew Ng, who is my advisor at Stanford University. Professor Ng is the founder of Coursera as well as one of the leading founder of Google Brain. And there is a few things I learned from Andrew. Andrew has a really high bar when it comes to content development. With him, I had the chance through lecturing at Stanford or teaching online to teach millions of people around the world AI skills. And the reason it resonated with the community is because the bar was so high. In fact, we spent 80% of our time deciding what to teach and only 20% teaching it because we knew it was so important to not distract people with methods that they may not need, but just focus on the essentials. On top of that, I remember the first article or paper that I wrote and gave Andrew, he just looked at the first 30 characters out of 17 pages. He gave it back to me and he said, "You need to work again on it." And then every day I would come back and put that stack of paper in front of him and he might read a little more and, you know, mark what he liked and what he didn't like until we got calibrated and I felt I had a better sense of what his bar ease for content development. And over time it became also my bar to a certain extent where I was able to understand what a high-quality article or research paper looked like. And I think that served me all my life. On top of that, like starting an AI company requires specific knowledge or specific skills in terms of how to structure your team, what you know is going to work and not going to work. And I learned a lot of these skills with Andrew as well.

Is there a way that you structured Workera that you would never have done had you not had this experience with Andrew?

Yeah, for example, one of our values at Workera is learning, growth mindset, lifelong learning. And that comes from Andrew's lifelong learning mindset. And I'm really glad we kept that as one of our values and that every time we interview someone, we truly look for Do they have this growth mindset or this insatiable curiosity that you want to have when you're starting a startup in an uncertain environment and are ready to pivot before you can find what your true purpose is and scale it?

In your mind, what are maybe the characteristics that really effective mentors always share?

Effective mentors are able to do 3 things really well. The first one is assessment. If the mentor can't assess your skill, understand your skills and your psychological profile, they're not going to be able to help you very well or customize their guidance. The second thing is they understand what you're capable of, your potential, and they're able to help you dream bigger. You might be dreaming of something and you could be dreaming of that thing times 2, and a good mentor helps you identify that. And then once the mentor understands the starting point and the ambitious yet achievable goal, the mentor is able to close the gap. And good mentors have a lot of experience having mentored other people on how to connect the dots between those two points. Now, my point of view is that we're never gonna be able to scale human mentorship, and that's why I think AI helps a ton. AI configured correctly is a better assessor of skills than any human ever. I in fact think that in a few years from now, it will be unethical for a company to ask a human to judge another human's skill. Whether it's external for an interview or whether it's internal for a performance management process. AI is just going to get better. AI is being right now learning from many interviews, from many assessments internally at companies on when am I fair, when am I not fair. We have at Workera experts, PhDs in skills measurements that are teaching the AI how to measure and when they make a mistake, When the AI makes a mistake, correct them. And soon enough, it's just going to be very clear that the AI is less biased than humans. Right now, we have companies that have thousands of interviewers that are completely uncalibrated from each other. Even if the HR team tries very hard to calibrate them, you might see an interviewer that likes people who sound like them or looks like them or thinks like them. And all those things are things that AI is less prone of. I would say bias is a bug and humans have it, unfortunately.

Workera has an AI tool called an AI mentor called Sage. Can you tell us a little bit about Sage?

Sage is the AI mentor of Workera specialized in assessing skills. So you can think of Sage as an agent that can show up in any enterprise workflow, whether it is to assess someone for their career development, assess someone for their eligibility for a project, assess someone to simply give them feedback, Sagee just really good at assessing people. It's really fair and it's getting better by the day. And that's the purpose where Kera is really trying hard with our research to build universal agent that can measure any skill with the best practices.

And, and what problem does this help solve?

Immediately, it saves a lot of time and money for people in an organization. The cost of a bad hire is extremely high. Both on the individual that was hired and then fired, but also on the organization who could have hired a better fit. That measurement should have been right. A lot of untapped potential within organizations. We hear every day from employees who are saying, "I should have had that project. I just didn't know the VP personally, and this person got that project. I had no way to show my skills and to raise my hand and say, 'I want to take care of it and I'm going to do it well.'" That's significant untapped potential within organization. And then I talked already about the billions of hours of humans interviewing other humans or humans judging other human skills. Time wasted, unfair, biased, lot of problems to solve there. And finally, retention and attrition. Most people leave because of their manager, and oftentimes you realize the manager has never been trained at evaluating people's skills. Very few managers actually know how to evaluate people's skills well. And so without understanding the skills of their people, managers are assigning the wrong tasks, they're coaching them the wrong way, and then as a consequence, a portion of the population that you did not want to leave ends up leaving. One of the problems we solve on the longer term is a more meritocratic society. I truly believe in a skills-based society where work and rewards are based on who deserves it and who's capable of it, and we're able to support people to get access to those opportunities. That future of a more meritocratic society is only possible if we have a universal measurement engine that everybody trusts, that is scientific, research-backed, and we just know it's better than human judgment. Once we have that system, I expect a lot of things are gonna change in the way We assess people, we evaluate them, we promote them, we reward them, and it's going to be for the better overall because people are going to get placed in the right thing that they're good at or going to be coached to achieve the things that they're not able to do yet.

How would that coaching fit in? How will mentors need to work differently?

Human mentors will need to work on their emotional intelligence, and I've seen it in different ways. The first example is teachers. To date, we've trained so many teachers around the world. Even as we speak, there's teachers being trained to be teachers. The job of a teacher is extremely complicated. It requires to be facilitator and motivator and mentor. It also requires you to have a subject matter area in math and physics and chemistry and so on. It's a lot to juggle. Not only you have to teach, but you have to look at which student is disengaged or which student is ahead and adapt yourself to them. I think that that role will change. Subject matter expertise on a variety of topics can be delegated with AI. And the real role of a teacher or a mentor is to be the facilitator, to be the motivator, to be the person who's going to cheer you up when you need it and hold the bar really high when you need it. And that's a very human skill, and I think we'll need a lot more of that in society. Less so of the person who just knows the curriculum by heart and is teaching on a whiteboard the curriculum. And I think for a mentor in the enterprise, managers, you can imagine that they're going to be able to manage more people. Today, what's the ratio of a manager to employees? 1 to 5, maybe 1 to 7, and then it gets complicated. I think in the future you might have a manager that relies on AI agents that are specialized for measuring people's skills allocating the team to the right task in a good way, helping them with the variety of management tasks. And as a consequence of that, one manager might be able to manage 100 people.

I want to shift gears a little bit and talk a little bit about you. I read that you sometimes play computer war games. I was interested in that. Tell me a little bit about that and how do you think that has shaped how you look at strategy and long-term thinking?

I love board games in general, and my favorite one is a variation of Risk. The board game where you have to conquer countries and strategize and make alliances. And I just have played that all my life with my friends, and we somehow got really good at it and hate each other for it a lot of times. So the game I like is called Warzone. People think of Warzone as Call of Duty: Warzone. It's not that. It's an indie developer who alone created a game called Warzone. It's an app you can download on your phone, and it's Risk, the board game, on steroids. It's a turn-by-turn game. It requires some thinking, and you have more time to think and establish your strategy. And on top of that, I oftentimes play with friends that are not where I am. I'm in San Francisco. My childhood friends are all in my childhood town in the suburbs of Paris. And it allows us to stay in touch and play together, even if we're like thousands of kilometers apart, because it's a turn-by-turn game. And they can wait on me tonight to play and they're in the morning and whenever they want, they play, and then we compare our results. What he taught me is probably thinking steps ahead, because to be good at a strategy game, you need to have a strategy. You need to know, what am I doing right now? What is my plan A? If that plan A doesn't work, what is my plan B? And what is it going to tell me about the third step I need to take? This adaptability and strategy, you can learn it through board games and social games, and that's part of why I enjoy it, other than the fact that it's just a good time away from work, if you will.

You founded Workera in 2019, and a lot has happened. Was there a particular turning point in the last 6 years that was really, really shaping for you? Maybe even something early on where you're like, oh, this is what we're doing now, you know, like where you just everything locked in.

Yeah, I was always passionate about human capital as a whole. So when I graduated from Stanford, I wanted to work in education, human capital, talent. But the problem within human human capital is so vast. And at some point, I remember I really felt strongly about the problem of measurements. I said, if there's one problem I want to solve within that whole human capital ecosystem, it is build the best measurement tool that you can. And it turns out that it will contribute to the ecosystem vastly. It will help what used to be our competitors, companies that are education companies. It will help large companies developing talent. It will help certification boards that are trying to make their certification better. It will— it's just measuring skills really well in a flexible manner as a universal capability helps the ecosystem a ton. And I remember having that realization and deciding to focus on that problem and not get distracted by all the other problems that others might be solving. I also think that in the early days of the company, because of how competitive I was, I was not prone to sharing a ton, or I was maybe cagey of like partnering with people that are in the same space. But over time I realized that, hey, we're all working on something that is meaningful for the world, and our space happens to be full of mission-driven people. And now I'm very comfortable sharing, and in fact I've seen that as an opportunity to partner closely with many impactful players in our market.

Is there one problem that you think leaders should be focused on for the next 5 years, and how should they be thinking about that?

Their people. That's the number one. People is the biggest technology problem of the next decade. Leaders should check in on their people very frequently. Leaders should inspire their people who are going through so much uncertainty. If you think your people are doing well, they're probably not. Even if they're doing well, it's better that you as a leader think they're not and check on them to confirm that they're doing well. Because we're asking workers to do more with less. They're worried that AI is going to take their job. There's a lot of uncertainty ahead that then flows back to their families and their living situations. And so as a leader, the best thing you can do is to check on them, bring clarity, and be transparent about the business outlook ahead of you so that they can prepare for it.

Is there a piece of advice that you've always been grateful for?

Yeah, I really like the one-way door, two-way door decision framework that I believe comes from Jeff Bezos at Amazon, and Andrew Ng taught me that framework. Decision-making is complex and a lot of people stress about it, including myself. But decisions can be categorized into one-way door or two-way door decisions. Some decisions are going to be one-way door, meaning once you make the decisions, there's no going back or going back is really, really hard. On the other hand, there's two-way door decisions where it's actually quite simple to go back and forth and change your decision later. If it's a two-way decision, trust your gut, go for it, it's okay. If it's a one-way decision, think more deeply, make the decision, and move forward. And I think that framework has helped me a lot at work when I'm stressing about a decision. Is it a one-way door? Is it a two-way door? It gives me a framework. It allows me to calm down and make a more sound decision at a reasonable speed.

That was Kian Katanforoosh. Thanks so much to him, and thanks so much to you for listening. For more podcasts, including my colleagues' program Radio Davos, go to wef.ch/podcasts. This episode of Meet the Leader was produced and presented by me, Gerry O'Hansen, and Taz Kelleher as editor, Edward Bailey as studio engineer in New York, and Gareth Nolan driving studio production. That's it for now. I'm Linda Lucina from the World Economic Forum. Welcome to 2026.