1. Podcasts
  2. The Business Leader Podcast with Sir Richard Harpin
  3. How Udemy and AI will revolutionise how we learn
The Business Leader Podcast with Sir Richard Harpin

1 Apr 2025 · 31m

Episode artwork

How Udemy and AI will revolutionise how we learn

In short

Host: Graham Ruddick

Guest

Hugo Sarrazin, CEO of Udemy

Description

Discussion on Udemy's role in reshaping learning through AI technology and the challenges faced as it scales as a tech giant.

  • Founded: 2010 in San Francisco, Udemy has grown rapidly in the online learning space.
  • Current Statistics:
  • Annual revenues around $1 billion.
  • 77 million users, including 17,000 businesses globally.

Written by AI. May contain mistakes.

A man walking through an airport, asking his phone a question

Ask about this episode

Get the gist, explore an idea or ask about something said.

Claude

On The Business Leader Podcast with Sir Richard Harpin, what are the key points of this episode?

Searching your listening

Transcript

Automatic transcript. May contain errors.

0:00This is Britain's growth moment. Do you want to be part of it? On 26th of March, 750 of the UK's most ambitious founders, CEOs and investors will gather in London to shape the future of mid-sized business. The company's driving our nation's success. The Business Leader Summit 2026 isn't just another conference. It's a day built around the three forces that define growth, ambition, policy and capital. And as one of our valued podcast listeners, you can be there in person. To secure a discounted ticket, visit the events section of our Business Leader website and use this code. BL Podcast, all one word, all uppercase.

0:41We look forward to seeing you there.

1:04That was Lando Norris of McLaren winning the first Formula One race of the 2025 season. I've come to McLaren's headquarters in Woking, Surrey, to meet one of its partners, Udemy. Udemy is revolutionising how we learn. It's based in San Francisco and is an online platform that offers courses for consumers and businesses. It has thousands of instructors offering courses ranging from how to use new AI tools like DeepSeek to one with McLaren about how to build high-performing teams. According to its new chief executive, it's only just getting started. We're on the cusp of a revolution in terms of how we learn.

1:52I'm Graeme Ruddick and this is the Business Leader Podcast. In this episode, we look at how Udemy is trying to revolutionize how we learn new skills, and how new AI technology is now finally ready to be used. To do so, I speak to Hugo Sarrazin, who joined as the chief executive of Udemy in March 2025. Udemy was founded in 2010, and Erin Barley, one of the co-founders, is the chief technology officer. The company already generates annual revenues of around$1 billion, And in 2024, it had a remarkable 77 million users, including 17 ,000 businesses around the world. But it looks like there's much more to come.

2:38What attracted you to this business and this role? Well, there's two reasons. The first one's a very personal one. My family, first time generation going to university. Both on my mom's side, on my dad's side, I got to see the impact of getting access to education, how it helps change the social context in which I was brought up. And then my dad was a university professor. My mom was a school principal. I ended up Canadian. I spent a lot of time reflecting learning. learning, but also I got the opportunity to help the prime minister way back in the early 2000s, think through education for Canada, passionate about it.

3:23When I went to the West Coast, Stanford had, we're trying to think about reverse classroom, got involved. So there's a theme here of how the impact learning has on society, how it can make a difference to help people realize their full potential. So I got very attracted by the mission. And that is always an important thing. And to me, there's clearly a mission that's really worth pursuing. The second, if I look at all the players in the space, I thought Udemy is uniquely positioned. It has many things that are very unique. The first one, it has a B2C business and a B2B. And the two go hand in hand.

4:08We get through a marketplace on the business to consumer side to serve millions and millions and millions. We're the largest one. And we attract a lot of content creators, the instructors. We're the largest one, too. So that little marketplace is like a fantastic platform that gives us access to great stuff that we can move on the business side. So again, now you move to the business, you can kind of curate the best content and do some amazing things. So that, to me, was very unique, very exciting. And I do think we're at a moment in the business world where we're trying to figure out what to do to help reskill the workforce.

4:52There's so much change going on. There's AI, obviously, but there's economic uncertainty, geopolitical uncertainty. You know, there's never been a moment in our history where there hasn't been as much change as now. So I thought, you know, being at Udemy and helping realize that would be like a really fantastic chapter. That consumer and business part of Udemy, it was initially a consumer business, really. And then this business side has grown rapidly. Was that a surprise? I mean, obviously you've just joined, but for the business, was that a surprise? And how has it changed the business's priorities?

5:29So, yeah, the consumer side was where the whole thing started. Our founder, who rejoined us and is the CTO, had this view of democratizing access to learning. And he had a view that you could also democratize access to the instructors. and instructors were not only at the best elite school, but everybody had a thing to say. So that was his vision back then. And it was a game changer. And that was the consumer play. And it worked really, really well. But then we start to see that businesses were taking advantage of that. So the blending of the consumerization of the business experience played big time.

6:13And we decided to, OK, wait a second. We may want to kind of organize this thing in a way that makes sense. So we created a separate group focus on it. We put together packages. We created the best classes. We created offers that enterprise could relate to. Then we realized that, well, beyond the content, what they want is things that integrate with their existing system. They want to link it to the business outcome. They want coaching on how to create programs to facilitate retraining certain parts of the workforce. So yeah, we ended up building this other thing and that took off because the existing providers were not meeting the needs and and we had a you know very very unique value proposition who were those existing providers what has where was it that you gained traction was it was it entirely sort of new offer or did you find that businesses moved away from so it's it's it's many thing uh the first one the most important one is the notion of most of the existing provider had catalogs and they were a publisher model.

7:18So you spend a lot of time determining what is the right content and you work on it, you do high production quality, and then you put in a catalog and then it stays. The marketplace, what's beautiful is we get 80 ,000 people who are fighting for eyeballs are keeping their content up to date. So we ended up with an offer that was way, way, way more powerful. It means it's fresh. And I'll give you an example. Earlier this year on AI, DeepSeek kind of like was a moment of reckoning. Well, go check on Udemy today. I think there's, I mean, we should double check. But, you know, last I checked, more than 200 classes on what is DeepSeek.

8:03It's in nobody else's catalog. I mean, that's the power of this marketplace as this network of folks. So that ended up being a value proposition that was incredibly, incredibly powerful. Hugo Sarrazin's task as the chief executive of Udemy is to help scale up the business around the world. Udemy has grown quickly, but like any business, it's had challenges scaling up. For example, last year, Udemy announced it would lay off 20 % of its global workforce. It's a business that is, you know, like many of these wonderful businesses based initially in Silicon Valley, tech disruption, founder, you know, insight on, and then took advantage of an incredible moment, you know, being there at the right time, right?

8:52Like many of these things. Then in the COVID period got an amazing acceleration. In some ways we pull forward many years of changing consumer in corporate behavior, you know, it was on a rocket ship, right? That's what this company had. It also made the transition to a global company. You know, 65 % of our revenues are from outside of the US. We have more languages than anybody else because, again, the instructors, there's somebody somewhere in some remote region of India who will offer a class in the native the thung of that region. And that's amazing. So a very global company already. And right now, part of the next chapter, if you will, it's about how do you continue to scale this business?

9:44Right. It's currently we're about$800 million of revenue. How do you think about the next chapter become a two, a three, a five, a$10 billion company and become this iconic adaptive learning platform. So that's where we're going. And that's what we need to build. What are the challenges of scaling for business like this? Well, there's always, you know, this too cliche for somebody who's representing, you know, a company in the learning space, but it's about people, right? It's about people. We really need to, you know, help attract, retrain and develop the best people to operate at the new scale.

10:24And, you know, what made us great in the past is not what's going to make us great in the future. And we need to constantly evolve. We had the pleasure today, we're at the McLaren Test Center, and we spent some time with Zach Brown on stage. I mean, he was describing the season. I mean, to me, that's one of the most incredible thing he said. the car in a season moves and 80 % of the parts or the car is replaced through a season. And I think that's something that we can learn a lot from. So as we think about scaling, we need to kind of constantly look at the processes we have in place that served us well.

11:05We need to honor them, celebrate them, but not be afraid of saying, that's not going to be good enough. you know if we I'm gonna you know how do we localize the content yes that's good we can do that with instructor we can do that by dubbing we can do that with translation but how do we localize the marketing how do we localize the method of payment how do we localize the support I mean you know something very simple as you know a global marketing campaign hitting I'm going to pick on India again, in the middle of a cricket game, it's not very, you know, geographically aware. So we need to build the muscle to understand, you know, act globally, but think locally.

11:48I think that's one of the things that we need to do. I think the other one here is given the amount of disruption AI is bringing on to the world, just in general, and we can talk a bit about, you know, what we're observing there, like it affects us too. So we need to be very thoughtful in saying, how do we use AI to transform our own business? How do we help our instructor build more adapted content? How do we expose our content more differently? I mean, we have a lot of amazing long form content, but now we have the ability to change the modality. Maybe it's going to be more short form. Maybe it's going to be more in the moment.

12:30There's lots of opportunity and all of this on the table. It's like a beautiful, the next chapter is going to look different. We're going to build on the strength that made us who we are. But, you know, as you can imagine, you know, the future is wide open. What is your ambition for the business? We want to be the category. We want to be the one that everybody thinks when you think about, you know, reskilling and changing my career and being successful, both as an individual. Like I'm sitting and I say, I'm happy in the role I am, But what if I could do this that I say, well, Udemy is going to help me.

13:08Udemy is going to be there for me. It was there the last time I made a career change. I want that to be the case in large companies. I want them to be able to say not, you know, I have a catalog of wonderful stuff from Udemy. I want them to say, you know, the way Udemy supported my business, we got to reskill and do something. And it's linked to my business objective and I can demonstrate the value. And once we do that and we focus on our customer, you know, the answer is it's going to be a, you know, five, six, seven, ten billion dollar company. It's just going to be an outcome. It's just going to be a wonderful thing.

13:43But we need to be able to think that way. We need to think global. We need to, you know, it is already global. As I said, 65 percent of our revenue. And we need to continue to think globally. Now let's talk about AI, because this business is at the forefront of the technological change that is happening. The first thing about AI and Udemy to say is how it is changing what courses and skills people want. Udemy runs a load of practical courses on how to use tools like Microsoft Excel. But increasingly, people want to take courses on how to use new AI tools. It is one of our hottest categories right now.

14:21It's in the moment and we've got lots of instructors coming up with all sorts of classes. It means different things for different people, right? You have how to build LLM and how to do prompt engineering to, you know, 101. Like, hey, I'm not sure I'm a paralegal and I'm not exactly sure what it's going to do to my industry. So it is literally touching everything. And it has to be one of the main theme of reskilling the workforce. So we're obviously, you know, supporting our instructor as they do that. And then we're using it ourselves. So it is absolutely like right now, the core, one of the main area.

15:00Surprisingly, is the soft skills are going to become even more relevant. And that's another area where we've got a really nice set of offering, but we're going to see more interplay between the two. if you don't need to code in the same way in the future as you use in the past, because, you know, you can have some conversation with some wonderful tool and there are plenty of them out there and it gets you part of the way there, maybe not all the way. Or if you're going to end up doing some validation of the quality or you're going to get some integration, like there's going to be a whole lot more that you're going to be able to do with AI and therefore we need to kind of be at all these intersection point and be able to support our customers.

15:49How much more practical has the use of AI become over the last six months, I guess, really? Because we had the first sort of wave of LLMs, but it's interesting talking to businesses now as we do. It's now becoming clear that there's some really practical uses that businesses are starting to put. I think it's like any technology adoption curve. It goes through, first of all, this is not new. Let me just kind of just state, you know, it may have come to the forefront of everybody's collective, but, you know, the core paper by Google is, I think, 2016. You can go all the way back. I mean, I have a PhD in engineering in 1994, and, you know, we were using many of the same concept, not all of them.

16:33It sneaks up on you until at one point, poof, it becomes like a really big and important thing. And that's what we lived. And then everybody goes and says, oh, what does it mean? And the exploration is that. And I think OpenAI and then after that, Gemini and Anthropic and others did a good job kind of letting people play with the technology. So it's not surprising that the first wave of learning and training is, what is it? How do you experiment? How do you have fun? And then everybody was looking for the use case. Everybody was looking for the use case. I mean, I spoke to many, many CHROs here and then before and now they're business leaders and they were all kind of trying to play around and trying to figure out what does it mean for me?

17:19How do I think about this in customer support? Yes, I can have a bot, but you know, I think we're at a point now where, you know, some of the use case are more known. The battles, I don't want to say the battles of the models are done, but the, you know, the mega, mega, mega model, there's going to be continued improvement. But I think we're going to move from the training of large LLM to how do you create micro rags? I mean, I'm getting a bit too technical, but there's some specific ways to augment models for specific use case in specific vertical, a specific department with your own content.

17:56Let me kind of give you an example. you could train a model to understand your HR policy. So rather than to go and chat GPT and say, how can I work my retirement plan, and it gives you the answer from the internet, which is interesting. What if now you had the training on your own company? This is a use case that's kind of now super common. People know how to do that and it's rolling across the world. Now you can do similar things in customer support. You can do the same thing in sales. You can do the same things in contracting. And so the ability to take specific domain, embed that in the model, tweak the feature engineering function in a certain way to get you very context aware answers.

18:43I think we're there. There's lots of good stuff. And we're obviously moving a lot of the training to support that. But equally important, what's now interesting is like, how do I get the organization to move on this? You can have the best model, but if people are intimidated and they don't know how to use it, it's not helpful. So there's a lot of change management, same way that we did with digital transformation, the same way we did, you know, and every technology way brings the same level of learning. And, you know, you need to help people feel that they can let go of what made them successful today, embrace something new, and make a step forward, and then the second step and the third step.

19:31Just looking at other skills on the recent reports that you've done, what other areas are people looking to upskill in at the moment? I mean, as I said earlier, there's a lot on soft skills. It is just a crying need. And even the funny part, I mean, we serve, we're very, we're the most pervasive in the tech industry and the professional service industry. We basically have the most professional service firm because they know, because they have this workforce that constantly being deployed against different customer needs. So they need to reskill them all the time. But yes, they do care about technology.

20:09But what they want is they want their managers to be good managers for the frontline. They want to have people with good, you know, critical thinking. They want to have people who can negotiate. And so those are the kind of topics that pop up all the time when you look at the courses, both on the consumer side and on the business side that are bestsellers, if I can use that language, top of the list. Even in areas where you think, oh, they want to hear about, you know, AWS certification or something. They do. And they do take many certification, but they just want good managers to be there to to motivate them, inspire them, create the right work culture, make an engaging culture, and there's a lot in there.

20:49Do you think that's going to be a sort of change across Udemy as we move through the months and years? Because at the moment, as you say, a lot of your courses are practical. I think the most popular still are practical. So is that something you expect to change? I think we're going to follow our customers is the simple answer. You know, we'll see where. But if I put my hat, I suspect our customers are going to be moving us in that direction because there's a big, big, big need. We're going to also do a lot in the, we have these academies where we're kind of helping people with very specific purpose.

21:23We, you know, change jobs, right? If you want to become a data scientist in finance, there'll be an academy and there's like going to be an overall program. So we're going to, you know, help shape that ourself. And rather than purely like go pick and choose the classes, we're going to kind of create the bundle to help you with a career change. So there's going to be a bit of that. And in each one of them, there's going to be soft skills and there's going to be more domain specific or technical skills. So we're going to also try to, you know, shape the curve, if you will, knowing what we know the market needs.

21:56I mean, some of these massive changes are predictable. I mean, we're in the UK right now, right? So there's a huge digital transformation that is on the agenda for the public sector. So it would behoove us not to kind of put a curriculum to say, hey, we got an answer to help you. Or you want to reignite growth. What do you need to do? Well, there's a whole lot of AI enablement. So we're going to package that to kind of meet the needs. And we're lucky enough in the UK, we had 2.8 million learners on the Udemy platform over time, which is pretty awesome. If you take a look at the UK population or active population, it's a reasonable number.

22:40Now our goal is to make sure that that's not just a learner at one point in time, but it's like an ongoing dialogue and ongoing engagement. And people feel, you know, as passionate, maybe. I mean, maybe I'm too I'm dreaming too much about, you know, the relationship they have with Udemy as they do with their Spotify or their Netflix relationship. This is a tool that my family uses to be successful. This is the tool I need to be successful at work, or this is the tool, you know, the UK, as you think about growth and where you want to position the country to the future, it's part of the answer. It's the way to get there.

23:18That's a summary of the skills that people want to learn via Udemy. But how they learn them is going to change, because this is where AI can make a huge difference. I'm going to anchor it first on personalization. And the personalization is your ability to ask question of the learner of what they know, what they don't know, and then tailor the content to specifically their needs rather than have this linear content that starts at the beginning and then just kind of runs you through what is supposed to be the best way to learn, that has limitation in many ways. First of all, it's not very efficient.

24:01You kind of go over things. Two, we all learn differently. And this is optimized for the average learner. So that cannot be the best way to have the best outcome. So having the ability to personalize based on an understanding of an assessment, an understanding of how you're doing in the moment, because you can now do more, you know, specialized learning tidbits, and then you have a feedback loop. So that to me is a real big game changer. The second is you can, instead of having one modality, which is long form, you know, here's a three hour course that we've now nicely broken down into X number of modules.

24:42We can make it so that in the moment when you need it, we make something available to you to consume. and it will be very easy. And this is actually doable. It just needs to get done. For example, once we find out that you're a manager and you've got book on your calendar, your performance review, being able to pull up your company specific training around that. And it's not like a compliance training. It is a, here's the three tips. Let me give it to you as a friendly reminder in a way that is easy to consume. in the moment. That's a pretty interesting things in terms of like delivering a better engaging employee experience and better outcome.

25:26I can have another one if you assemble a team for a project and then you're the person sending out information to start the project. You could have a little coach on the side say, hey, by the way, do you want a reminder on the best practice to kind of have the kickoff meetings in the moment, right? You've got this mentor, you've got this coach supporting you who's context aware. That's pretty cool. The third one, and then I'll stop, is it helps with the pacing. Right now we have this very specific way to think about pacing, which is, again, optimized for the average learner. In the future, you can tweak the pace, and you can kind of also tweak the reinforcement learning that's embedded in the learning.

26:16So we're building all that in the platform. So our platform, you know, again, we're very lucky because it's a platform versus a publisher model. We're creating a lot of these capabilities in the tool and we're making it available to consumer, which is a good base of training. Not everybody has the ability to train on almost 80 million people. We use that training to kind of optimize the experience we deliver on the business side. And then we offer these tools to our instructor to optimize the way they build the content. And then the platform on the enterprise side is also capable of supporting the customization large enterprise want to do or medium enterprise.

27:00The reality is the way I want to onboard at the end of the day is specific to my company needs. And if I have now tools to facilitate my company onboarding so that my frontline associate can get productive more quickly, I mean, that's worth a lot of money. I mean, everybody is hiring and training and there's like, especially in an industry that are seasonals or industry where there's a lot of churn and attrition, facilitating the onboarding with company specific training. It's pretty awesome. So AI is going to make it much easier to build and deliver that in a way that we know is effective. How quickly are you going to be able to roll out these new products?

27:44I know you touched on this morning some of the new things that are coming. You just talk a little bit about your product road. Yeah, I mean, listen, we have an incredibly ambitious product roadmap, very ambitious across all the dimension that we touch and more. And some of it is in beta mode already. So we're pretty excited by that. And many of the things I'm discussing is going to be available this year. It's literally, and that's, again, back to your original question, what got me excited? I got a sneak peek at the product roadmap and I said, holy crap, I want to be part of this adventure. It's going to be a game changer.

28:22You've spoken in the past and written in the past actually about the importance of trust with AI. How do you ensure that here and when you're putting everything together that you maintain that trust and what you're offering? Yeah, I think it's so important. So important. And I think we talk a lot about trust. I think we need to continue to talk a lot about trust. it's a version of the same for every technology, but this one has an added degree of nuance. And to me, I think it is important for the users to understand what is doable and not doable. So having the information, for example, if you have specific answers to questions, being able to know where the answer comes from, being able to influence and provide feedback on that, you know, that builds trust.

29:16You kind of like, I'm not fully in control, but I've got at least some transparency. I've got some understanding. So that's one thing. The second, you know, the data to be clear with the user, where is this data going? And, you know, I don't think we always do. I say the role we like the tech industry do a great job explaining that. And I think you need to be forthcoming with that and make it very clear so the user can understand what is being done. And at the end of the day, listen, you build trust when you have empathy with your customers and demonstrate humility. This technology can do amazing things and will have shortfalls and you need to be willing to accept it.

30:05I mean, the biggest one that everybody knows is the hallucination that these large language model can deliver because, again, I'm going to get a bit too technical, but it's, you know, these language model are nothing more than statistical models that predict the next word that comes after a certain word. Now, they happen to do it on, you know, billions of data points, and therefore they get pretty good, but it's still a statistical model. So it is prone to error and you need to kind of build the right feedback loop, put the human in the metal to improve and correct, to eliminate them. I think when we get into these RAG models, which I talked earlier, that are vertically specific or domain specific, you have the ability to constrain the answer set.

30:49That's where basically you're just controlling what goes into it. Yes. And that makes it better. But we need, I still think we, and here I'm using the Udemy we not the tech we need to kind of be very clear with with our users what this is what is being done how it's helping and then be open to you know help our our user community our instructor community our partners to make it better you've been listening to the business leader podcast with me graham rudik our producer is anushka tate for more business case studies and masterclasses, please check out businessleader.co.uk.

31:40The Business Leader Summit 2026 is set to be bigger, bolder and more ambitious than ever. Don't just watch Britain's growth story unfold, be part of it. Secure your ticket today by visiting the events section of the Business Leader website and using the code BLPODCAST, all uppercase. It's your chance to join the leaders who are shaping the future of business.

From the publisher
Udemy is the San Francisco-based company changing how we learn new skills. Udemy offers online courses ranging from how to use Microsoft Excel to how to build high-performing teams. Now it plans to use new AI technology to provide personalised coaching and training. In the latest episode of the Business Leader Podcast, Graham Ruddick meets Udemy chief executive Hugo Sarrazin to talk about its plans and the challenges of scaling-up a new tech giant

Hosted on Acast. See acast.com/privacy for more information.

More from The Business Leader Podcast with Sir Richard Harpin

All episodes