E375 | Jan Lynn-Matern, Emerge: Future of work & learning VC Emerge launches oversubscribed Fund II at $73M

32:35I can only imagine how big it is. But I was much more specialized two years ago in what I did on a day-to-day basis. Because now I can do what two people were doing for me before. I can have a computer do it. Yeah, exactly. Basically, you need to have probably at least two or three specialisms and then a very broad knowledge base, whereas previously you might have one specialism and a broad knowledge base to be a knowledge worker. Because, yeah, you need to be able to sort out the agent when it's bugging. So you need to have that certain baseline in multiple areas. And, yeah, it's super interesting.

33:19You have the capacity to have so many different things running at the same time that the puck ends with you. Where I think that there's so many people that I know that as long, you know, once there's more than five processes running that all answers to them and they need to kind of keep control of, it gets difficult. And that is where I'm seeing there's a big change in what we educate people to. And there's a big change in the skill set that each individual needs to have. Because a lot of this, what I call lateral thinking or execution before, I think there's a lot of people that's going to struggle there because they don't like the uncertainty.

34:00They don't like shifting from task to task so much. Yeah, I completely agree. Yeah, I like your term lateral thinking. I think another way of calling it is like interdisciplinary work. People need to not think in silos, but be able to look at multiple different fields, be able to quickly get to grips with a new field that they previously have now worked in and come up with solutions that combine mental models from different areas of the world. It's incredible. Maybe you could give us some examples of these non-traditional solutions that all of a sudden can kind of happen in the ecosystem. Yeah, so what we've done is we've actually come up with, together with our venture partners and through a lot of research and thinking and obviously just seeing the patterns in the companies that approach us, we've come up with 40 solutions.

35:02So these are our investment themes. And they broadly sit in these three categories, learn, navigate, work. And what we're doing is, as new companies get created, we're continuously mapping the entire EdTech, HR Tech, Future of Work ecosystem of tens of thousands of companies against these 40 ideas. And so we have a database. So actually, these 25 ,000 are not in thesis. And then these 5 ,000, they correspond to our 40 ideas. and we're eating our own dog food, by the way, because we have an AI agent do that for us. Of course. Of course. And as we briefly touched on, these 40 themes don't neatly fall into EdTech, HRTech or whatever.

35:52They're our own thesis. And what we're also seeing is if we compare them to traditional EdTech or HRTech, they're massively outperforming those categories. like in 2024, they raised at least 3x the funding of the average like ad tech company. And that's partly to do, we think, with the macro need that I've been talking about and partly to do with, you know, these themes are all very heavy on the AI application layer, which is obviously seeing a lot of interest right now. But yeah, so to kind of tell you more about it, yeah, I would love to go into each, like learn, navigate work, talk a little bit about the specific problems we see or the specific needs we see in each of these, and then give you examples of companies that we've backed or kind of the kind of ideas we'd like to back.

36:43Let's start with one where everything begins. Okay. Yeah, great. Yeah, so, I mean, at a high level, as we were discussing, to succeed in this new world, you need to be adaptable. You need to be able to learn things fast. And that requires two things. Firstly, you have to have a strong educational foundation. You have to be able to think. And then secondly, you need to have easy access to new learning opportunities or to learning resources when you want to expand. And neither of those two things are available to everyone right now. And I'd say kind of there are many reasons for that. But at the core, the main reason is that to deliver a high quality learning experience right now requires a lot of human interaction.

37:36And that's just super expensive. It's hard to scale. That's why a lot of people don't have access to that. And the big idea with AI is to start delivering developing systems that can take over some of those interactions, that can mimic the actions of a good tutor or teacher and that then can democratize access to learning opportunities for people at all ages, from kids to adult learners. And to be clear, ChatGPT is not there yet. ChatGPT is desperate to give you answers to questions, but a good teacher doesn't do that. A good teacher helps you figure out how to get to the answers yourself, But I think at the application layer, that's where the opportunities exist to build didactic models that end up doing these things and that grab this opportunity.

38:31Could you tell me some of the subcategories within learning? Because, okay, so you just described it's the fact that AI can provide low cost personalized learning accessible for all ages. But what are the subcategories within this? Yeah. Yeah. So there's five subcategories. It's tech for young families. Then there is like consumer learning, consumer focused companies that provide learning support to kids in high school and university. It's co-pilots for educators who are just massively overworked and they're using AI a lot to help them do more with less. And then it's like self-improvement and language learning, which is more focused on adult learners.

39:14And if you look at those categories, today's spend is about 500 billion across those five. Why don't you have the old behemoth of schools, universities, like classroom teaching? Yeah. I don't think you have educator co-pilots. So it's within the world of formal education, it's the co-pilots that you focus on. It's not the whole disruption of the existing way of schooling. There's several reasons for that. I think high level, the formal education system that we have right now is becoming obsolete. We need to learn so much faster than a four-year university degree. And I think a lot of the things that you can learn at university, by the time they launch a course, it's taken them five years to approve.

40:18It's already out of date. It's probably out of date by the time the draft is written, but it's definitely out of date by the time they've approved it and actually launched the course. So that's one big reason why I think real change to how learning happens is not going to come through institutions. The other reason it's not part of our investment thesis is just because it's really hard to sell into educational institutions and get to any scale quickly. So it's not where we're focused. When it comes to learn, it's mainly consumer companies that we invest in. What is the expected date of death for the existing university?

40:57It's a really interesting question. Yeah. I mean, I think over the last decade, there has been a quiet death of a lot of lower tier and mid tier universities. It's not been a new story, but there have been several factors. You know, one is the real cost of going to university has been increasing at like a stupid rate compared to inflation. Universities are... In Europe as well? Because I know this for the US. Mainly in the US. Yeah. I mean, a lot of European university provision is obviously free. So it doesn't... Yeah, but it might be, you know, in the back end, the same problem. We just don't see it.

41:41That's true. That's true. Yeah. Yeah, I think, I mean, it is mainly in the US, even in the UK, for example, where more and more of the cost of higher education has been shifted to the consumer. The skepticism amongst consumers to basically invest in a university degree when it's unclear what the real impact on your career is going to be of a lot of these degrees has meant that some universities have really suffered. And the big brand names will always be there. But for the lower tier and some of the mid-tier universities, unless they really adapt, they're going to find it tough. Yeah. Teal fellowship all the way, man.

42:24Give us some examples then. What is it within these subcategories that you haven't mastered in or are looking at? One area that I find really exciting right now is learning support for kids at all, like whether they're in high school or university. because every kid faces exam pressure at some point in their career, but not every kid has the same level of support to basically succeed in these exams. And some kids just find it really tough to perform in an exam setting in the first place. And if you combine that with, you're maybe from a disadvantaged background where your parents can't afford to also pay for a tutor that prepares you for these situations, you're going to suffer.

43:06And then these things have long-term consequences because it determines what university you might get into and so on and so forth. So to level the playing field and give everyone equal access to good learning opportunities, democratizing access to learning support outside of school is actually really important. And tutoring is super expensive. But as we were just discussing, AI can start mimicking the actions of a good tutor. and we have a company called Solvely where we co-led the first institutional round for the company earlier this year. And they're basically an AI study buddy for all subjects used by both university students and by high school students.

43:47And contrary to ChatGPT, it doesn't just give you the answers. It helps you figure out how to get to the answer. The tech is essentially, they've figured out how to not hallucinate for really complex questions in maths and science, which is where ChatGPT falls short. That and the user interface is like their source of defensibility. And this company is like the fastest growing company I've ever seen. There is so much pent-up demand among students for these kind of tools. They were, in general, they were the first to adopt AI, you know, from basically January 2023, because contrary to popular belief, They're not just after the answers, they're actually after high-quality learning support.

44:34And TurnGPT isn't enough. They're looking for more. So Solvee is a really exciting company, a year old, and well, well, well into seven figures, and doing exceptionally well. and then like a second example in the same category as a uk business called my edspace which has a really exciting model so they're basically like you can imagine them as twitch for education so they're taking teachers that are that are streaming and they're turning them into celebrities they've built a machine that turns them into celebrities on social media and then they become cool and kids want to learn with them and you can subscribe basically to have a flat rate access to these streamers all year round.

45:25And again, the traction in that company is going bananas. There's a company in India called Physics Waller, which has established this model. They've just raised the$200 million Series B. And My AdSpace is basically bringing that same model to the West and now using this massive traction to also build AI learning into the product as well. So like between the streams, you'll have AI support to help you with homework. And then you go back to the big brand names to get your fix. So those are two examples from the learn category. To get your fix, yeah. To get your diploma, right? Because that's the value they provide.

46:06Yeah, yeah, exactly. Incredibly interesting. I remember something that you spoke about, which was the access to tutors or to, you know, knowing as a blue collar family member, you know, how to navigate the ecosystem of education and how to hack that system. I just, when you shared a bit about that, I thought back to my time at university, my first year. And I remember I went up to an exam. I knew that I'm fucking good at this. I went up and I was expecting to get an A and then I failed. I was like, what the fuck? What happened here? And then I took that and went to my teacher and said, what happened here?

46:51And then, you know, I said, well, blah, blah, blah. But then what I did get was I got three A-graded assignments, right? And then I said, okay, so what has happened here? And then I deduced what's the preparation for this type of exam to be able to then deliver this type of product, right? which then you know after i did that i went up and i got an a and i then did the same thing with the other courses uh and in that same semester and i got ace all around right and so i realized this is just about hacking the system it's a game yeah it's ridiculous so then you know i actually sped up my bachelor's so i did my bachelor's in halftime i was then working during my my master's So I think I showed up for the first and the last class in each or course there.

47:41But that was literally blew me away. The fact that as long as you know the methods of testing and you know how to hack that and you know how to learn, you've got what you need to be able to ace. At least in the in the social sciences, you've got what you need to be able to ace the exam. stories. It just absolutely blew me away. And the fact that, you know, you can now, I can only imagine like, it's very erratic to the university ecosystem to just be teaching methods to hack exams. But an AI tool that would just tell you, look, give me the three best assignments. I'll tell you exactly how to prepare for this type of thing.

48:25And then I'll run you through that. So simple. Yeah. Yeah. It's funny. I mean, yeah. Firstly, there's a ton of demand from students for that kind of support, but also contrast that with what you were talking about, about your son at the beginning of the conversation. Yeah, completely. Like that's not curiosity. And actually that's not what's going to help you build the multidisciplinary knowledge and lateral thinking that we were talking about. That's going to like help you succeed in this new world. So I think we need both. You need to exploit the system. You need to hack the system to get through the exams, but then you need to find the curiosity in yourself or else.

49:05I've never been more curious. Like chat GPT has allowed me to find my curiosity, like, or at least put it on steroids again, because I'm learning so much now. All right, let's get to navigate then. So this was learn. Let's get to navigate the second part of the learn, navigate work thesis of yours. Yeah, so this part is really about like how the labor market changes as a result of AI disruption. I think the fundamental trends are like more people will transition in and out of jobs than they are today. You know, when more jobs are being automated, people are going to transition out of them. When, you know, people will be looking for safer careers, people might be equally looking to shift more quickly or more frequently because as it becomes possible to learn new things more quickly, there's new routes that are going to open up for them to get better paying jobs.

50:00And McKinsey is predicting that we'll be at like 2x pre-pandemic levels in terms of job transitions. By what year? By 2030. By 2030. Yeah. And then at the same time, organizations will really feel the crunch for skills because they will want the kind of people that you were talking about earlier that can use AI to do more with less and they're rare still. And so you've had this world where more and more people will be switching jobs. Companies are desperate to find the right talent. But already today, the way we find jobs and the way that we hire people is stupidly inefficient. And if we're going to double the rate, we better do something about it.

50:53So the big idea here is that we can use AI to increase the efficiency of these processes and to have essentially like an impartial expert at every step of the way to help both people and companies make the right decisions about where and who to work with. So again, there's like a few subcategories here. So career navigation is one, helping people make these choices. Most people fall into their careers. They're not deliberate choices. So, and that's insane to me. And by the way, just while we're at it, I wanted to ask you how you ended up here. How did you make this choice? So that is, I've always just gotten involved in whatever I was close to.

51:40Playing computers a bit haphazardly when I was 13, kind of, well, I really enjoyed this. Then I did World of Warcraft. Like, you know, so I was one of the best on the server, which was quite a lot of grinding back in the days. And what does being the best mean? is it like that means that means you know killing people so we literally had in there was a this was pvp so player versus player and we literally had a ranking system in world of warcraft at that time where you would you would have grant marshall which there was only one that was made grant marshall every week uh on the server on each side of the uh alliance or the horde and then you'd have two that were would be made field marshall and then you'd have i think it was four or five that would be uh the the tier below that i was literally playing so much that and then i i woke up it was Wednesday that it was updated the rankings and then i was i woke up i was getting in i was hoping i was planning to make field marshal so the you know in the top threes at least uh on the server i had been playing eight to twelve hours straight you had you were competing against other people doing bots uh like that was grinding all the time for them people sitting in china playing for them and stuff like that.

52:58I got up. I was expecting to make Field Marshal. I had not. I was like, what the fuck? And then I went out and found the two people that made Field Marshal. The one that made up before me was like three kills more than me. And we had like 250 ,000 kills in the last seven days. I was like, what the fuck? I stopped playing that day. I stopped playing. But then back to back to back to how I fell into this. So at university, I became student president, was close to the vice chancellor. He asked if I wanted to to help build the university incubator. I did. And then, you know, someone came from a pre-seeded fund and asked if I wanted to join there.

53:41And then, you know, I was on this route. Actually, I was about to decline that offer to to help build because I was super excited about university education at that time. So I was about to decline the offer of going the more innovation route. But then I realized, wait a second, before I went the education route, I had actually very much steered towards the entrepreneurial route. So I realized I'd only gotten myopic on education because it was what I was in. And I had this personality of doubling down on one thing, whereas my own personal inclination was actually probably entrepreneurship. So that's why I ended up going that route after all.

54:23I was just about to fall prey to falling into something instead of doing what I probably should do. And yeah, that's fascinating. And also, I think what you've described just proves the point that whatever we do as our first job is the highest signal predictor for what we end up doing in our careers. And it has a huge impact on earnings potential as well. and like whatever we first do is often the least deliberate choice yeah it's so crazy that's a great point i'll say that to any young person i know uh yeah like it has a huge predictor of your final earnings like ah that can make anyone think about what they do as their first yeah yeah for sure yeah so like you know one of the ideas under navigate is to basically build AI tools that understand you, understand your goals, but also have a really good understanding of how you can fit into the labor market and fit into career paths that have a potential to be fulfilling for you, whatever that means.

55:34Earnings potential obviously being an element of that and who can guide you in the right decision, but maybe not even stop there. like we could be eventually in a world where everyone has an agent that like goes around and applies to jobs for them and constantly figures out like what's the you know next best move for you and obviously that kind of career support exists in the world they're really expensive and they only work with the highest earning people but uh if an ai can do that at near zero marginal costs, why wouldn't that be available to a far greater proportion of the workforce? There's one idea.

56:16Some examples, some concrete examples of the most interesting things you've seen built in this space. Yeah. Yeah. I mean, so I guess an example is an investment in a company we made recently called Pop. They are more on the B2B side, but they essentially automate a lot of the recruiting process. Part of the problem in recruiting is it's really expensive to hire the wrong person because firstly, you invest time in the process. And then if you fire them, you have to do it again. Secondly, there's the opportunity cost of what the value you could have created had the right person been in the role from the start.

56:54And then, of course, you have redundancy pay, etc. Companies waste billions on bad hires. But for some reason, the way we acquire talent is still in the dark ages. It's like people pinging hundreds of emails to each other. Inevitably, the majority of the candidates that apply to a role don't hear back. The TAs, the talent acquisition people, the hiring managers, bring their own bias into the decision-making process. there are really rigorous science-backed methods for making better decisions, but they're so expensive, you would only ever use them for like really senior roles. It's just, it is ripe for disruption.

57:40And so like Pop takes a lot of that process and automates it, like from the moment someone applies to a role all the way to where they get like that final interview. It's all done by AI agents that Pop has created who will go through the materials and figure out whether there's a fit. These agents will have a human-like conversation with the candidate on whatever channel the candidate wants to ask them questions and match them against the required skills. They'll collect documents and verify them, and they'll even do the scheduling for you. Have you seen any, what you might call, career co-pilots.

58:19I've met one founder, I think he was probably in Berlin. That's at least where I met him. Doing what I would probably call a career pilot, meaning that it would sit on your computer, see everything you do, give you feedback both in terms of, ah, it looks like you're less engaged. I can see from your email that your reply times have increased by X. And I can see in your calendar that so and so and i can see the tone is going this way maybe you should think about all that all that kind of self-development stuff as an individual in your professional life um and then for them the business the monetization route was actually to the company because the company would then get high level reporting on worker engagement and that type of thing i i love that i'd love to speak to the company that you're thinking about but yeah i've seen like 10 or 15 career co-pilots in the last six months and it's an area that we're really excited about we haven't pulled the trigger on anything but it's really exciting what's what's happening there these things are you know helping you figure out where you should focus they're helping you build the best resume for that job, you'll put your resume and the projects you've done into the engine and then it figures out how to position it to the job based on the requirements in the job.

59:46And the craziest thing I've seen, which I think goes to a moral gray zone where we probably wouldn't invest, there's even companies that are like a live teleprompter during a job interview. So imagine... I guess, yeah, I like that. Like they listen to the interviewer and like you just get like this is what you should say and can i get that for my podcast it would like i would do three podcasts every day i would just read aloud from the teleprompter dude like that if i'd be surprised if that doesn't exist like riverside should definitely be doing this but like i i'm like 95 sure that the founder that was pitching me that idea was literally using his own product to just say whatever the product said, you know, is recommended he say in response to my DD questions.

1:00:36And it would actually be the best demonstration as well, right? I couldn't make up my mind. What did you do is he should train his nine-year-old or whatever daughter to go on his calls and blow everyone away that, how can she, like, and then ah, okay, it's that powerful. I think, I think or just have Joe Biden do it. Yeah. You know, if he'd done it in a transparent way, I would have, it would have been like, respect your products. Right. But he was trying to get away with it. And yeah. Okay. So that's navigate. Now let's get to get, let's get to work. That's the final one of the, the learn navigate work framework.

1:01:17So going back to what we were saying earlier, to be successful in a world of AI disruption, organizations and businesses need to be constantly adapting as well. Every month, every week, the technology shifts and businesses need to respond to that. But in order to be able to do that, you need to have upskilling as a core competence in your organization. And a core competency also needs to be tooling your people up with the latest AI tech to actually use it in your business. And where we are today is so far away from being able to say that the majority of enterprises have upskilling as a core competency.

1:02:01Like high quality corporate kind of learning opportunities like coaching or really high quality training is only available to like the top 1%. And everyone else is getting access to LinkedIn learning, some static content library that they never use. And good luck. And I think this is genuinely a massive, is going to be a massive problem. because when these new AI tools come in, it's not going to be obvious how to use them. Like famously, like how old is Excel now? Like 40 years? I don't know. But so it's surely we've figured out how to use Excel by now, right? But the majority of people don't know how to do a VLOOKUP function, like let alone write a macro.

1:02:55And that's probably... I stopped at VLOOKUP. Yeah, you should definitely get into index match. It's way better.

1:03:07Anyway, so if you have an all-powerful AI agent that comes in that can do way more than Excel, it can actually do the data analysis for you, but you have no way of actually learning how to do it. How are we going to enable people to take advantage of these things? So that's the big idea. And I think it requires probably these tools themselves to also be providing the learning. Like the learning has to happen in the moment of work rather than in a hotel room, you know, in like a conference room in a hotel where you send your people away for a two-day workshop where they forget everything and then they come back to work and be like, okay, how do I do this?

1:03:55so it's like the thesis is that working and learning will fuse and it will happen inside the tools that we use every day and an example here is a company we just backed called Mendo just announced it last week they're a French business that you can imagine them as a co-pilot for your co-pilot so they sit inside chat GPT for enterprise or inside Microsoft co-pilot or like Office 360 that's now also AI enabled. And it observes the way that you work with these tools. It knows your goals and knows your tasks. It knows your role. And then based on those two things, the observation and its knowledge about you can suggest how you can get more out of the tools.

1:04:43And it's basically purchased by big companies like PwC, EY, BDO as an add-on to the co-pilot they buy because they know that unless they have a tool like this in the process, they're never going to get the productivity gains from these co-pilot tools that they're buying. That's super interesting. I literally was typing a question here at the same time because I thought I wanted to ask you if you think that or to what level the workforce is ready to use AI. And I say that because even working with my own team but talking to other people as well that are non-VCs, it's incredible how low the adoption is.

1:05:27Even when people have tried ChatGPT, realized they can learn anything, realized that it can write anything, and then you realize they're not using it. It's like, what? Yeah, no, absolutely. That is the, you know, when I was talking earlier about like, there is potential for huge inequality. This is what I'm talking about. It's like, there are people like you who are fundamentally changing how they run their business, because they're using AI to the max. And then there's the rest who are not there yet. And we need to really rapidly invest in upskilling folks to take advantage of these tools, or there's going to be problems down the road.

1:06:10Do you have, from doing the work on Mendo here, do you have some stats top of mind on adoption across companies that do these large projects with AI? Yeah, there's a few interesting stats that came out of the diligence process. One is that only 8 % of users of LLMs come from large enterprises. you would think that it's way more because that's where the real gains can be made these large enterprises are massively behind the adoption curve and they're all freaking out I mean it depends on who you talk to but they all know they need to do something but people are worried about risk and security and if they are adopting it they don't know how to measure whether it's actually increasing productivity and so So the other interesting stuff that came out of it is as a result of this confusion, they're all hiring McKinsey and BCG to do projects for them.

1:07:13And McKinsey is doing like 30 to 40 percent of its annual turnover from AI projects right now. It's just mental. So there's, I think, a huge opportunity in driving bigger adoption in the enterprise of these kind of tools. But yeah, it requires an investment in upskilling. and that's hopefully what some of the companies that we invest in in the work segment can do. Amazing, Jan. This was so much fun. I'm really looking forward to doing it. I can just announce it now at the end of the episode. We're going to do a couple of episodes more together because I'm such a big fan of this space so it's going to be incredible.

1:07:50I just cannot wait to... And I think there's one deep dive we need to do at some point and that is the obituary of universities or formal schooling. I think we need to do a deep dive and really map out what are the problems, where do we see them? Do you know A16C? They've done it. They did a three-hour show. Ben Horowitz and Mark Andreessen did it. I think it was two or three episodes of two to three hours each where they dissected the higher ed crisis in the US. I love to see that for Europe because it's much more fragmented, very different dynamics. but I imagine that there's something happening similarly but just behind the scenes.

1:08:35We don't see it because we don't have the same exposure from having the fact that there's people paying. One of our LPs is the CEO of Europe's largest university which no one has ever heard of. They serve 8 % of the German university higher education market which is massive. like hundreds of thousands of students, billions in revenue. And it's 100 % online. They take three weeks to launch a new course where other universities take three years. They have a bunch of degrees that they've called the Copilot School, which they have built together with Microsoft, where they give you a degree in how to use Microsoft Copilot.

1:09:21It's the most innovative university I've ever heard of. and they are disrupting the space. So yeah, I should put you in touch with him and we can do a - We need to bring him on. Yeah, let's do it. Thank you. Everyone, on the note of me realizing that I was being, what is it? I have a meeting AI co-pilot that tells me that I'm not being inclusive because I say fucking shit too much. So here I am being mindful about myself, lack of inclusionary terminology. Jan, thank you so much for joining this podcast. It was incredibly cool to dive deep with you on this. Amazing, Andreas. Thank you so much. And yeah, it was a lot of fun.

1:10:03This would have been a final. Tear down this wall. It's more than just an ally. This is a union of values. Let's start acting.