Lefteris Ntouanoglou:
You’re trying to build the capabilities that move a specific business outcome. And that’s the whole idea.
Skills are not the destination. Skills are an intermediate step. And we spend as a market too many years in talking about skills taxonomies and building nice skills libraries and missing the mark. And what do we want to accomplish through those skills? What capabilities do we want to actually really improve here at the end of the day?
Steve Smith:
Hey everyone, welcome back to another episode of Work Tech Weekly. I’m Steve Smith, Managing Director of Growth at Rep Cap.
Skills have been the hottest idea in HR tech for years. HR teams have built taxonomies and libraries to track who has which skills, mostly for people who sit at a desk. On the front line, that kind of tracking misses what the business cares about. A worker can check every box in the skills library and still get the order wrong during the lunch rush. And every wrong order shows up in food waste and remake costs.
My guest is Lefteris Ntouanoglou, founder and CEO of Schoox. Schoox is a learning and development platform built for frontline workforces. Lefteris has a PhD in physics. He built Schoox from first principles, without benchmarking what the LMS market was already doing.
We talk about why skills are only one step on the way to a business result, and what AI can learn from the questions workers ask mid-shift. We also get into where training hits its limits on turnover, and what it means to run a company on doing the right thing when nobody’s checking.
If you’re building a skills strategy for the front line, start here.
Let’s get into it.
Welcome to the podcast. We’re really happy to have you here with us today.
Lefteris Ntouanoglou:
Thank you, Steve, for hosting me.
Steve Smith:
You live in one of my favorite cities in the world, Austin, Texas. And you’re not originally from there, are you?
Lefteris Ntouanoglou:
No, and I still actually go mostly back and forth. I still reside and I’m based in Thessaloniki, Greece, where I originated. We are headquartered in Austin. It’s my favorite city. I spend most of the time when I’m in the US in Austin, but I’m still based in Thessaloniki, Greece.
Steve Smith:
What’s the flight back and forth like?
Lefteris Ntouanoglou:
When I wanna fly to Austin and it’s not just one of my destinations once I’m in the US, then it’s a little bit challenging because I have to take three flights most of the time. So from Greece, somewhere in Germany, from Germany point of entry. Austin is not as much as a popular point of entry from international through international flights.
And then from the point of entry then over to Austin, so...
Steve Smith:
You built Schoox in Greece, but then, you know, you relocated the whole company to Austin. What’s sort of the short version of that story?
Lefteris Ntouanoglou:
Well, actually, the truth is that it was not really an expansion. That is a part that people sometimes get wrong. We were not a Greek company with customers and revenue moving to the US. We were building the business for the US market from day one while physically sitting in Greece.
Steve Smith:
Well, you know, this is a really interesting time to be in L&D right now. I think that obviously with a lot of kind of the shifts in the workplace tech market over the past few years, we’re at a point now where you have a lot more, I think, ability to dial in solutions for employees.
And I think that it’s fair to say that, you know, you look at the past 15, 20 years of the market, a lot of especially L&D technology was mainly designed for people sitting behind a desk, you know, really for knowledge workers. And a big focus of what you do at Schoox is really more on the frontline workforce.
How did you come to that focus? Was it just sort of organic and natural, or was there something else behind that?
Lefteris Ntouanoglou:
Well, I think it was a combination. I mean, my first job was as a high school student. I was working at a restaurant, so I got exposed to all the challenges that a frontline job has early on. Later on, when we built our first learning and development solution, we went after the restaurant industry, and this is where we gathered all of our expertise and learned and appreciated, you know, those industries.
And then we were just going after similar industries, always with frontline, because we really started understanding very well how to build a solution, how to expand the solution, how to serve those industries that have very unique pain points. I mean, very different industries and very different cases than white collar jobs, definitely.
Steve Smith:
You have a bit of an unusual background for even the work tech industry. You have a PhD in physics. What does that scientific background teach you about, you know, how people actually learn that maybe coming out of the, you know, HR and L&D space wouldn’t have?
Lefteris Ntouanoglou:
I think it taught me to think from first principles, Steve. In science, you learn to not accept assumptions too easily. You keep asking, what is really happening? What variables really matter, and what can we prove, and what are we missing? And so, and you get also comfortable with the idea that complex problems rarely have simple answers.
So I think that shaped me as a CEO more than as a founder, because if you do not understand the real source of a problem, you can move very fast and still solve the wrong one. So it is also why I did not wanna build a better version of what was already existed. I wanted to stand back and ask what learning and development should look like if you design it around how people actually grow and how businesses actually improve.
So coming from an HR background would definitely have given me domain knowledge that I had to learn the hard way, but science gave me a different lens. Curiosity, skepticism, and patience for problems that take years to solve, I would say.
Steve Smith:
Well, and it seems like that... I can see where that’d be an advantage, because you also didn’t have a bunch of things you need to unlearn. Because it seems like, oftentimes, the obstacle for innovation is the inability to see a problem a different way. You know, you get locked into one way of seeing things, and you don’t necessarily ask questions about, “Well, what if... Why does it have to be this way? Why can’t it be some other way altogether?”
Lefteris Ntouanoglou:
Yes, that’s correct.
Steve Smith:
So, you know, was there a moment when you felt like, “Okay, at Schoox we’ve got... We, you know, we’re really onto something here.” This... Was there one of those kinda light bulb or aha moments where you kinda felt like, “Okay, we’re definitely on the right track”?
Lefteris Ntouanoglou:
Well, there was... You mean when we made a decision to build the product? No, there was no any dramatic moment, honestly. I should be honest about that. It was a frustration that built slowly while I was working around LMS products and HCM software. I kept feeling that category was solving the wrong problem.
It was very good at administering training but not really designed to help anyone get better at their job. So the thing that actually mattered was a decision, so not a realization. I made a deliberate choice to not study too closely what already was out there. Because if you benchmark against them, you build definitely something better most likely, but still fundamentally the same thing.
So we started from the question instead: if the goal is to help people grow and help the business improve, what should this actually look like? So that was actually what happened.
Steve Smith:
Well, you know, and I think that one of the things that you focus on that I’m in complete agreement on is just the idea of, like, starting with the strategy. What is the business objective that you’re trying to achieve? And not, you know, focus on the learning metric or the development metric in and of itself.
Why does it seem like the industry is always relearning that truism? You know, why is it that... Because it seems completely obvious that it’s just like you wanna start with what you want to achieve, but it seems like I’ve seen companies over and over again who never got the memo on that.
Lefteris Ntouanoglou:
I think this is true, and as much as it seems to be very obvious, it is still very difficult, and the market is not going... I think it is just much easier to measure and count things that you have access very easily. Measuring business impact or going that route is just a much more complex, I would say, challenge.
You know, training is an investment, and an investment needs a goal. So we reverse the usual order that everybody should do. But so we don’t start with what to teach. We, exactly as you said, we start with what are we trying to improve in the business? So you mentioned speed of service, guest satisfaction, order accuracy, safety, retention. What is it?
So let’s say it’s order accuracy. The instinctive answer is communication training. So that is where most programs stop, and this is exactly why most programs cannot prove anything afterwards. So you work backward instead. Which roles touch this number? What do those people actually have to be able to do?
In this case, would be listen properly under pressure, confirm the order back to the guest, know the menu and the modifier codes, use the POS correctly, and hold accuracy through a peak hour rush. So communication is in there, but on its own it does nothing. And order accuracy, by the way, is not an isolated number either. It moves food waste, remake costs, speed of service, guest satisfaction, and even frequently employee stress.
So only then do you design the training, because you’re not trying to improve skill in isolation, you’re trying to build the capabilities that move a specific business outcome. And that’s the whole idea.
Skills are not the destination. Skills are an intermediate step. We improve skills to build capabilities, and we build capabilities to move business outcomes. And that may sound kind of obvious, but the whole chain is a much more difficult kind of chain to build. And sometimes we have even situations where, you know, the final outcome is not that easily measurable.
And so I think this is actually the problem. And we spend as a market too many years in talking about skills taxonomies and building nice skills libraries and missing the mark. And what do we want to accomplish through those skills? What capabilities do we want to actually really improve here at the end of the day? So I think all of those just sum up to a problem that just keeps existing out there.
Steve Smith:
And I love everything you just said. I think that, you know, one thing I’m curious about is, you know, learning, you know, does need to be tied to business metrics, but what happens when some of those business metrics are harder to quantify? Like, you know, having a great culture or creating a safer workplace or improving, you know, situational judgment. You know, how do you approach that in your work to be able to kind of draw that direct line to those outcomes?
Lefteris Ntouanoglou:
Yeah, I think this is exactly how we built our solution, how we leverage AI, because from the very beginning, we wanted to connect learning with development and business outcome. Yes, some learning is more difficult to quantify, but hard to quantify is not the same with no impact. It usually just means the chain has not been mapped yet.
Culture, for example, influences retention. Safety influences incidents, compliance, and risk. Judgment, I think you brought up as well, influences decision quality, error rates, escalations. And curiosity could influence adaptability and internal mobility. So none of those are soft outcomes. They’re simply further from the number that we’re looking for here.
And take culture. Better culture improves retention, right? And better retention means more stable and more experienced teams. More experienced teams give customers a better experience, and better experience raises the likelihood for them to come back, and that raises customer lifetime value at the end of the day.
So those are five links, and most organizations cannot see past the first one. And so what we’re building is the ability to make those relationships explicit, connect skills with capability with a metric, and understand how one metric even moves another, and put, at the end, a value on the improvement. This is how it should be done.
Steve Smith:
So you’ve called the old HR tech learning stack a Frankenstein, which is, I think, sort of saying the quiet part out loud, and it might not be popular in some quarters, but I think it’s extremely accurate. You know, it’s, you have these solutions that are stitched together, legacy code bases, and, you know, rapidly getting into the unwieldy. How has Schoox avoided that fate and made sure that they didn’t fall into that trap as you grew and became a bigger solution?
Lefteris Ntouanoglou:
I think it all happened because the way we started building the solution, it made just total sense to bring in all of those activities that could support that kind of vision. And over time we didn’t even know exactly how to call our product, obviously, very often, because there was no clear product category where we were falling into.
And so we started as a learning solution. As I said at the beginning, I mean, I always deeply cared about learning and development. And very early on, I mean, in business this is how things are. I mean, we have to understand that when we invest in our people in any way, that we expect, of course, to get better business success and, you know, better business growth. This is what this is about.
And so learning is just another investment. And so when we train people, we train them to develop, you know, the capabilities. And when we develop the capabilities, we wanna move specific business outcomes. So if you think of this, the way we built the solution from day one was always putting pieces together in one unified solution.
So it was back in 2016 when we added skills in our learning solution. I do remember a question I got from a C-level executive from a big HCM firm who asked me, “Lefteris, why in the world did you add skills in an LMS?” And I thought this is a weird question, because I thought, why, how it... I mean, how in the world do you know what training your people need if you don’t know what skills are required?
And, you know, later on, of course, we understood very early on that skills alone is not enough. Those skills have to impact specific capabilities. And I think just one thing brought the other, and as we were exploring and tapping, we’re just going deeper into the field, we felt the necessity to add more stuff and more. But always within the context of what we wanted to accomplish. And naturally, some point is we ended up that way.
And what we had to compete very often against is we were one solution versus 10 applications out there. I mean, in what the market looks like today, you know, look at how companies supporting employee growth today. I mean, learning sits in one place, content in another, skills assessment somewhere else, performance, then goals, coaching, career development, engaging. You know, there’s like 10 systems in a software world.
And take the example of order accuracy rate. So a learning solution, a traditional learning solution, will show course completions. A skills system will show skills assessments. A performance management system would show manager feedback. And a business system would show the cost for remake.
Now, someone has to sit on top of this and try to figure out how to connect the dots to understand not only if there was an impact of whatever they did in order accuracy, but also what they should repeat and what they should stop doing. So because the way we always thought, this is how we... And yeah, we didn’t assemble everything with acquisitions. We just built it from day one as a native unified solution.
Steve Smith:
Well, you know, I think that part of the... It seems like the advantage is, as you’ve also pointed out, is that, you know, with a unified system, AI can be the thing that finally connects the skills part, the learning part, and the performance data. You know, what do you think that AI is able to do differently than traditional dashboards, and how is that going to ultimately be better for the end users?
Lefteris Ntouanoglou:
I think a dashboard shows you data inside a model that you have already built. AI is building the model. And so this is actually the difference. Let me give you a few examples.
In our solution, for example, it can construct the entire business impact profile. So it’s telling us which skills the role needs, which skills should be developed together and benefit from being learned together, what capabilities they add up, which business metrics those capabilities move, and how much improvement is realistically available. Then it can build the whole training, and afterwards explain what moved and what did not.
And so another way is when we talked about learning that is not that easily measurable, and how it can build a whole chain and help you understand, you know, the whole connection, which is very difficult to accomplish without AI. You may do it once for every job, but, you know, the business is changing, so it’s a very dynamic kind of role. So you would have to continuously rebuild this.
And also imagine, for example, systems in frontline organizations that are learning in the flow of work. So people get responses to questions they have and get operational assistance how to get the job done.
At the same time, though, our system, for example, is not only providing those responses to those people, it is also analyzing all of those questions at the back end and understands where are people getting confused, what gaps do they have, for example, what are they struggling with, what is the quality of your content where people get responses from and aren’t very happy about this, what capability gaps are there, or what knowledge are they missing?
I mean, this is a live picture of your operational reality. There is no dashboard that can show this, because a dashboard would have to deterministically be defined from the very beginning what we wanna measure. And so this is a completely different story here, I would say. It’s not just a reporting layer, it’s an intelligence layer that sits on top of everything.
Steve Smith:
That’s huge. I think another thing I’m wondering is, you know, when you think about, you know, how AI should work in a solution like yours, is it something that needs to be transparent to the ultimate end users, the frontline workers? Or is it something that they shouldn’t even know exists, it just works? What’s sort of your take on that?
Lefteris Ntouanoglou:
Look, I think transparency and letting people know whenever AI is producing something that they consume is definitely part of the policy that we should adopt. But at the end of the day, I mean, we have to build our solutions using AI as we did in the past as well. We used machine learning, we used semantics. It doesn’t matter what we use. I think we shouldn’t really emphasize so much on what it is.
I mean, we have to build solutions for their reality, not for ours. So these people are serving customers, they’re running equipment, solving problems on the floor. The goal is not to get them use AI. The goal is to use AI to... so that their work is easier and help arrives faster.
Practically, what does it mean? It means in industries like frontline, it means, for example, mobile, role-specific, contextual short answers in the flow of work rather than a course. So someone should not have to really go through and read a 30-page PDF manual to find an answer, or be just recommended, suggested a structured course that they would have to go and understand how they should, you know, act or deal with a certain situation.
I mean, this is kind of how AI should come into the work. The technology has to adapt to the worker, not the other way around, I would say.
Steve Smith:
There’s been a lot of conversation, I think, in, you know, our industry circles about employers cutting back on investment in people because they want to invest in bringing more AI into the workplace or more automation into the workplace.
But I mean, a story from the front line that was just in The Wall Street Journal last week about how fast food restaurants are starting to roll back a lot of the AI ordering and automation, because what they’re finding is it’s taking the humanity out of the workplace, and it’s making it not as a customer-friendly environment.
Do you think that maybe we’re going to be heading toward a sort of a reset on, okay, maybe we’ve gone way too far in favor of investing in AI and we need to get back to investing in people? Or are you seeing something different? Are you seeing that hasn’t changed at all and your customers are totally on board with keeping up that people investment?
Lefteris Ntouanoglou:
I think that what’s going to happen is that at certain periods of time we will just pause and reconsider and correct things, adjust things, and then just keep going. Technology is evolving. It will never stop, and AI is just... It’s a genie that came out of the bottle. I mean, it is what it is.
I mean, that will keep evolving, and we can slow down the pace. Decisions were made, suggestions were made, but sooner or later, whether we are going to get where we would be in a year or we’ll get there in 10 years, at the end of the day, we will get there.
So I think it is always important to understand how AI can be used in a wise way. And yes, the truth is that wherever we remove the human aspect and, for example, damage things that we said are much more difficult to measure the impact, like culture and other things, that will definitely have another impact, so we will always have to balance everywhere. I think this is something we will learn out of that evolution.
Steve Smith:
What I think is interesting about, you know, the, you know, frontline workplace situations is these are such incredibly difficult employment environments just from a training and an investment perspective.
When you think about, you know, in hospitality or in, you know, any sort of service environment, you might have turnover rates that are 50, 75, sometimes 125, 150% annually, where it’s just like you’re constantly having to bring in people every 60 to 90 days because of the turnover, and that’s just, that’s brutal from a training perspective.
How have you kind of seen the ability to have a solution like Schoox that brings in that, you know, in-the-moment contextual training that, you know, can be mobile-based? How have you seen training and development have a meaningful impact on turning those numbers around and making those work for employers?
Lefteris Ntouanoglou:
I think learning alone cannot solve the problem of the high turnover in those industries. I wouldn’t claim otherwise. People leave for pay, scheduling, their managers, workloads, things happening in their personal life that have nothing to do with us. But it can definitely move part of that number, and the part it moves is early.
So I think most frontline turnover happens in the first few weeks, and this is usually not a pay decision. That is someone who felt lost, unprepared, and not set up to succeed. So clear onboarding at the beginning, this is highly important, and a manager who actually knows how to coach can change that specifically.
And yes, and because there’s a limit on how much time we can spend on training those people, those are high-paced environments, and people don’t have time to sit and go through long structured training. And of course, so many things change within that organization, so they would have to go again and again. And no matter how much time you spend on training those people, at the end, they will still have questions.
So operational assistance, you know, learning in the flow of work, you know, bite-sized training, short answers, this is highly important, because onboarding is definitely important. But then to really help them be successful and do the job right, it is very important. And if this cannot easily be accomplished through structured training, then this kind of training, bite-sized and in the flow of work, is highly important.
I believe people stay longer when they feel capable and can see a path. Okay? So if learning is a compliance checklist, as it used to be all of those years, it cannot move that retention at all. If it makes people good at their job much faster and can keep assisting them in doing the job, you know, the right way and be successful, then it becomes part of the retention strategy.
Steve Smith:
And I guess another thing I’ve been wondering is, when you think about, you know, the frontline workforce as a monolithic term, it’s pretty all-encompassing, because it’s just like, we think about manufacturing versus retail versus hospitality, these are very different, you know, workplace environments. Is it possible to build a solution that can serve all of those audiences, or is the reality that they’re really more the same than they are different?
Lefteris Ntouanoglou:
Genuinely different. And of course there is also some similarities. A machine operator, a hotel front desk agent, a restaurant shift leader, a retail associate, they do not have the same day, definitely. They have different risks. They have different pressures. They have different compliance and different metrics.
But they do share, you know, something that’s more structural than it looks. They are desk-less. They work under time pressure. They cannot sit through structured training, as we talked about. They need the answer now, in the flow of work, and not in an hour, not even in 10 minutes. They depend heavily on their manager and on the person who sits kind of, standing next to them.
And as we said, what they do lands directly on the business. So safety and quality in manufacturing, the guest experience in hospitality, and sales and service in retail. So in our solution, we do not treat them necessarily, all the frontline, as one audience, but we treat it as a set of roles where individual capability shows up immediately in operational performance.
The need is the same: workforce readiness. The context changes completely by industry, by brand, sometimes by location. But our platform was built for that difference, and it works, you know, along, you know, with... despite the fact that it has those differences.
Steve Smith:
Let’s talk a little bit about the founder part of your work. You’ve, you know, been doing this for a while. You’ve got some, I think, you know, apart from, you know, the business success, but just the... When you think about, like, the culture success you’ve had, you’ve got a pretty high rating just in terms of, like, your CEO approval score when you look at, like, employee rating websites. When you think about, you’ve been a finalist for Entrepreneur of the Year awards.
It’s just like, this is a... You know, being a founder is, it’s a hard thing to do, and you’ve got some things that you can definitely point to and show as legitimate, you know, major accomplishments. What do you think that, when you think about what your team says about you, you know, is there anything you push back on, or do you think, “Nope, they pretty much got it all right. I am doing a great job”?
Lefteris Ntouanoglou:
Look, I mean, I think they would say I’m demanding and that I go very deep into the details, which is true. I mean, I’m an engineer and I’m a product CEO, and I want to know, you know, how exactly something works and whether it genuinely can solve the problem. I think they would also say that I care too intensively about the product, sometimes being neurotic, and I think this is true as well.
But I hope that what sits underneath is visible, too. I mean, our culture rests on two Greek words: philotimo, which means doing the right thing and being responsible for it, even if nobody’s checking, and meraki, which is meraki, putting your soul, your heart, and your pride into your work. And so those are not slogans from me.
They are the reason I’m demanding, and I think this is also... even though very, very different cultures between Greece and the US, definitely something, you know, I had to learn, yeah, you know, early on. But yeah, I think this is what they would say, and this is maybe why we’re successful as much as we are.
Steve Smith:
Well, this has been a great conversation. Thanks so much for taking some time to talk to me, and congratulations on all of your success at this point, and I wish you nothing but the best going forward.
Lefteris Ntouanoglou:
Thank you so much, Steve. I highly appreciate it, and thank you for your time as well and for hosting me.
Steve Smith:
The skills movement was right that skills matter. It stopped too early.
On the front line, a skill only counts when it moves a number someone is already tracking, like order accuracy or remake costs. That’s a tougher standard than a skills profile. You have to know which roles touch the number and what those people actually do under pressure. Then you build the training.
Hard-to-measure outcomes like culture still count under that standard. They sit a few more steps away from the number.
The front line makes all of this urgent. When turnover in some of these jobs runs past 100 percent a year, nobody has time to build a perfect taxonomy. You need people ready to do the job in their first few weeks.
So before you map one more skill, figure out what it’s supposed to move. If nobody can answer that, it’s just another line in the library.
If you enjoyed this episode, make sure to subscribe to Work Tech Weekly on Apple Podcasts, Spotify, or YouTube. And I’ll see you next time.