Building, Backing & Scaling in an AI-Native World with Kenneth Auchenberg, Partner @ AlleyCorp

0:00Kenneth Auchenberg:Hello and welcome to another episode of Riding Unicorns. Today we're joined by Kenneth who's a partner at Alicorp. Thank you so much for joining us. It's great to have you on. Kenneth, maybe we could start with a little introduction from you on yourself and also Alicorp and then I've got loads I want to ask you about so then we'll dive into some of the details of what you've been up to over the last few years in your career. Absolutely. Well, first of all, thanks for having me. I'm really excited to be here. I'm Kenneth. I'm a partner here at Alicorp, leading our investments in AI developer tools and infrastructure.

0:37And for those who don't know who Alicorp is, we're an early stage venture fund in New York. We're coming from a family office background, but we are a fully-fetched venture fund today. really focus on everything from starting a few companies every year with our incubations, but also doing Pre-Seed and Series 8. So we do everything early stage these days.

0:58Kenneth Auchenberg:Awesome. And you have an amazing background where you're a proper builder. You've really been involved in some incredible things from early days of video conferencing, code editing tools, payments, and so much more. Now you're more on the investment side, But what really made you want to be involved in the financing and the investing side of business, as well as, you know, having that core kind of builder attitude in you? You know, I started my career as a software engineer many, many years ago when I was 16. I also have a bit of a Danish accent. I'm originally born in Copenhagen, Denmark, where I really started building.

1:36And, you know, my career has really been focused on a decade of building as an engineer, then a decade of product management. And now here on the venture side. And in many ways, I think it's all about like building great companies and contributing to that in terms of like some people write code, some people write strategy. And now I get to work with a lot of early stage founders and helping them building the companies. It's really all about company building. and now I'm just on the other side of the table. But, you know, like somebody who started out writing code and was fortunate enough to be a part of a few great startups back then in Denlog, it's really just like a natural continuation of what I've been doing for the past, yeah, time flies, decades, right?

2:23Kenneth Auchenberg:Yeah, and I think as an industry, we always like to see more people who have actually been hands-on and have operational experience and engineering experience, et cetera, in the investment hot seat, what do you think it gives you as an investor from an individual perspective? What do you think that having all of that experience of actually getting into the code helps you with when you're making investment decisions? I'm obviously biased here, but I absolutely think having a board member on your team who has been in your shoes, who has been actually operating, building, hiring, who has been in the early days of sleeping in the office, of doing the all-nighters, and have really been in your shoes as a founder, just builds another level of empathy.

3:10But also, I think, with what we do at AlleyCorp, all of us are former operators, and we've been building companies. And I think you having that operational experience and credibility and the network that you can mobilize, I think is critical and gives us an edge into how we can help founders. but also I want to be the phone call. Like when the founder has a problem, I want to be the one they call because I can actually help beyond just writing them a check. And so far that has turned out to be the case with the founders I work with.

3:44Kenneth Auchenberg:That's great. That's awesome. And how do you make sure that you are that person to call? I've had other people come on the show and say that they make sure they engage within 24 hours of like wiring the money or something like that to make sure that they really open up that communication channel from day one. Is there anything you do to make sure that the founder really knows that they actually can go you with their biggest problem? You know, like all the founders I work with, we call each other, we text each other. The founder knows that I'm available. Just last night I was texting with one of the founders I work with.

4:17And I think it's really about establishing that relationship is that I'm part of the team. I'm here to help. Naturally, it is your company. you have to do your company, but I'm available when stuff needs to get done. And I think for some investors, they are like an email away. I want to be a phone call or text away and really build that level of communication. But I think it's also about what do you bring to the table? I think it's been particularly in the early stages of company building, you need to mobilize your extended network. And it's a part of that in my job is to mobilize It's our collective network and how we can help this particular founder, this founding team to get to the next level.

4:58Kenneth Auchenberg:Yeah, great. And Alicorp has a great model where you not only invest in companies, but you also incubate. Do you want to just explain a little bit how that works and what's the sort of decision making process around stuff that should be invested in or should be incubated? How does that all kind of play out in reality? I wish I could sit here and outline a very well-articulated framework, but I will spare you for all those details. Because really, the way we see this is that we get to work with a lot of great founders and we can back them with capital. But there's also opportunities that we observe as former operators and investors in the market where we've kind of been spotting opportunities and saying, you know, we have looked at 50 companies in this vertical, but we think the market opportunity is actually slightly different.

5:44And when we see those opportunities, we see tangible problems. We take that opportunity to put a team together and really incubate a company. And right now, I'm in the middle of incubating an AI agent company. I'm the interim CEO while I'm putting the team together, et cetera. But that's really us leaning into our operational expertise of company building. And that means right now, I'm busy investing, but I'm also vibe coding at night, building things. And I really think when we see those opportunities, we're putting money behind our conviction and starting companies in the areas. We incubate companies a few times a year, but what we mainly do is invest in great founders.

6:30Kenneth Auchenberg:Awesome. And how does Vibe Coding change the way you think about incubating? Because it's really lowered the barrier to just prototyping something very quickly. So you mentioned, I think, you do a few incubations a year. Do you expect that number to go up because you can kind of take a few more shots on goal with wide coding? I think, first of all, it will be in one of the most exciting times when it comes to writing software. We are going through one of the few paradigm shifts of where we're actually redefining what it means to be building software right now. And as a former operator who has been spending decades building developer tools and code editors and those kind of things, We're finally seeing the capability unlock with AI and models that a lot of the things we were dreaming about in the 60s when it comes to computer science, we finally now have the capability to realize those.

7:17So it's an incredibly exciting time. But the second order effects of what it means when we suddenly have agents writing code for us is that I would say the early stage game has changed a lot. And we kind of call this AI-native founders, and we're seeing a generational divide between how AI-native founders operate and then the previous generation operate. And that has really changed our expectations to founders. And it means that if you are a founder right now racing for your pre-easeed seed round and you show up without a prototype, you're really sending a clear signal. If you spend more time perfecting your pitch deck than you by coding at night and not really building a prototype and getting that in front of your users, you're really demonstrating to the market that you're not AI native, that you have embraced these tools.

8:05So by coding and the ability to be able to build very quick iterations of your product is definitely changing the game. But also the big thing we're seeing with AI native founders is how capital efficient they are in the sense that we are basically seeing a reallocation of the early stage budget. In the old days, you know, I remember the days in Copenhagen when we were sitting in the basement building. I would need to go and hire a team of like a handful of engineers, maybe five engineers. I was in the basement crunch for like six, nine months and then finally have something that would work out.

8:40I could take to my customers. I can do that in weeks now. So it really means that what we're seeing with the AI native founders is that they might go out and raise a few million for the seed round, but they're not spending the money to go out and hire engineers. Rather, what they're doing is that they might hire one engineer, and then they spend the money on agents and tokens and basically spinning up like fleets of agents they're using to build. And then they're starting reallocating the money they have to actually investing in other areas like distribution, brand marketing, et cetera. So we are seeing a reallocation of how the money is spent because the cost of building has calmed down dramatically.

9:18Kenneth Auchenberg:Yes, so interesting. We're definitely seeing the same thing on our side. And it's not just the coding. We're also starting to see things like vibe marketing and things like that happen. So what do you think is the future of kind of software only businesses? How do you think about how do people build defensibility and moats and unfair advantages and all those sorts of things that we as VCs generally like when everything starts to become, you know, reduced down to unskilled people having superpowers and things that they previously couldn't do? We are definitely entering the era of mass-produced software.

9:59And generally, my thesis for the past many years has been that we are transitioning from a world where 80 % of all code was written by hand to 80 % of all code would be generated by a model. That's pretty clear that that thesis is kind of playing out, and that's kind of the world we're moving in. But I would say that this notion of us going from artisanal software to mass-produced software is kind of funny if you think about it. Like in the old days, you would have an artisanal engineer sitting and writing every single line of code. And today, the code code is going to do that for you. So the value of the code itself is changing a lot.

10:32And in many ways, we are back to a lot of the good old modes from like the 2010s about distribution, data, taste, brand, etc. But I think especially in this world where I can deploy an agent to look at my competitor's website and make a copy of the feature in an afternoon. I think a lot of the value going forward in particular will come from you being the subject matter expert of how you solve a particular problem. A company will buy your solution because you are the best way of solving this particular problem and you own the workflow. And by owning the workflow, that means you can provide the best possible way to how you do, let's say, issue tracking.

11:09The customer can also go and use like a carbon copy of your piece of software, But I go to linear because they're the best people in the world that really cares about how you do issue tracking, project management, et cetera. And that's kind of the edge. So in many ways, I think brand and trust and taste is going to be much more important going forward that you have to lean in because the act of building is not really what is going to keep your users around. But there's also elements of like, you know, you owning the workflow. So that's kind of been a philosophy of mine. It's like, how can you as a founder carve out a space where you can come in and land and expand and own the holistic workflow with the products you're offering?

11:48I actually think like workflows goes hand in hand with data. And that goes hand in hand with what we're seeing around RL and reinforcement learning is that if you're owning the workflow and you're actually facilitating the full journey of solving a particular problem, you are in a very unique position to collect all the data that is needed for you to train the AI systems and the agents actually do that job going forward. And by you owning the workflow, you collecting that data, that actually puts you in a very unique advantage and you build a mode around that data. Because like a generic AI company that does not know the specifics of how you do an accounting operation on how you, I don't know, review like inbound deals as an investor.

12:31If you don't know how those workflows are done, you cannot collect the data and you kind of build the genetic systems for it and so in many ways we're really down to a lot of the basics of how we thought about modes a decade or 15 years ago yeah and i would love to hear a little

12:48Kenneth Auchenberg:bit about your experience of building agents because i think maybe a lot of the audience have played with lovable or cursor or repli and vibe coded a website or an app or something quite front-end heavy. But how are you going about building agents? Is there a platform that you like to use? What sort of tools work really well for when you're actually building agentic software? I think the running joke right now is that if you're building agents, what you're really building is workflows. You're looking at a set of actions and you're changing them together. And then the most advanced agents, they have a bunch of calls to the lens for your decision-making tool use, et cetera.

13:28Right now, some of the things we are building is simply just changed operations and workflows that is calling out to LLMs when you need it. We use frameworks like Mastra, that is a JavaScript framework on how you build agents in JavaScript. But most of the pieces are really not, we don't use any particular framework, et cetera. We are just building these workflows in-house. And that can be a workflow on, let's say some of the internal tools that we have and how we analyze inbound deals, how we score founders, those kind of things. But when it comes to like bytecoding tools like Lovable, Bold, Replit, I think those are not really like about like building agents, but like using agents to build software.

14:10And, you know, every software engineer these days are probably sitting in a tool like Cursor, GitHub, Copilot, et cetera, and using the agent inside the code that is to build software. And I think CodeCode is really taking off and we're seeing everyone kind of orchestrating many, many cloud codes at the same time, building software for them. And it's pretty fascinating to kind of see how some founders are really just replacing the early stage engineering team and using agents instead. One of my favorite pitches from a few months ago was a set of founders that introduced their founding team. And it was five Devons from Cognition and MultiDevon, the engineering manager.

14:48And that was the founding team. It cost them$2 ,500 a month instead of going out and hire five people. They were just using Devins to get the work done, right?

14:58Kenneth Auchenberg:Very cool. Very cool. Yeah, it's all changing in front of us. Have you been pitched by an avatar yet? I haven't, but has that happened yet? I have not. And I'm also like, it's a bit daunting, isn't it? Like you're suddenly having an avatar due to doing a pitch. But I think it is rather ironic that everyone are using some sort of AI in either end. I think that the running joke with interviews is that you have somebody being prepped by their AI agent and what to say. And then the other person interviewing, they have a note taker. And then it's all get scored by AI in the end. So why are we even talking to each other, right?

15:34I don't think we are moving into a very interesting kind of world where we're all going to be AI augmented and assisted. But will I truly be just talking to your future AI agent that is going to interview me here for the podcast? I'm not sure that that kind of feels right.

15:49Kenneth Auchenberg:I'm not sure we'd have the same zing. let's hope not anyway um well that's interesting and you've mentioned one of the companies i think you're an investor in linear and i think you're also an investor in resend and um resend is i think the preferred email solution for lovable so it if you ask lovable to build an email flow for you it will tell you to create a resend api key and put that in that's amazing for distribution for a business like Resend. So what is the future of kind of infrastructure distribution when connecting up to Vibe coding tools like Lovable, Cursor, Bolt, et cetera? How does that become a really incredible distribution mode if you're the kind of default preferred partner for a particular task used by Vibe coders?

16:43Yeah, you know, like I think there's a lot of talk about like what is CEO in this day and age in the world of AI, but also like taking it down to the more fundamentals. We've been talking a lot in this year about like agent experience. And in my prior life, when I was at Stripe doing a lot of the work for our developer platform, we thought a lot about developer experience. And really, we think about the journey we've been through of like, in the old days, it's all about your customer experience. Then we were focusing on the developer experience. And these days, it's really about the agent experience, because the reality is that our teams are changing.

17:20We're going from a world where it was mostly humans, and now we are in this hybrid reality that you're going to have like five devons on your engineering team. You're going to have some cloud coach. You're going to have a marketing agent doing a bit of the work. So I think the target of who you're serving is also changing, right? In the sense that if you are Resent building email infrastructure, the reality is that a big part of your users that used to be human developers are now going to be Lovable's agent. And you need to think about what are some of the tools and abstractions that you're providing that is agent first and how can you optimize that experience?

17:58So Sinnoh from Resend is actually leading the pack here. They're doing a lot of work on how can I optimize and make sure that Resend is easily consumable by agents so they can be the favorite choice, but also the easiest kind of piece of infrastructure that can be integrated. So they've been doing a lot of work on simple things. How do I get an API key for an agent that actually needs to use my service? And even my former team at Stripe has also been pushing the edges there of having a Stripe MCP server that Claude or Cursor can talk to to understand what's going on. But it's really true that I think for any kind of service provider that in the old days just had an API now needs to evolve to a reality of like you need to treat agents as a new kind of client type.

18:47Just similar in the old days, we were optimizing our experiences for the desktop experience. And then we all were trying to retrofit the desktop website to our mobile phone. And we all know that didn't work out as well, right? And then we started treating the mobile app as a dedicated client. The next leap we're seeing is people that treat the agent as a first-class citizen and really optimize the experience around that. So recently, it's one of my favorite examples of them now having a massive kind of inbound from Lovable, Bold, all the other ones. And I think in the database space, I think we saw Neon that got acquired.

19:21but they had the vast majority of their new signups and databases provisioned. They were coming from agents.

19:27Kenneth Auchenberg:Yeah, I'm sure, you know, Lovable's growth must be driving amazing growth for Resend, Superbase, Stripe, etc. Because they're all so integrated into Lovable, which is the one I use. So that's why I know about it. Actually, the Stripe experience is great. And I guess it also links to, I think, Stripe's sort of pitch initially was, will grow as the GDP of the internet grows. And I guess Resend must be going, will grow as long as the GDP of Vibe Coding grows sort of thing. It's like being that closely associated with such a sort of clear economical growth point is amazing to be so correlated to that.

20:11Our strategy at Stripe in the early days was always about how can we capture 100 % of the YC startups? How can we enable them with payments and just take that off their place so the founder could go back and just get busy building their company? And how could we enable us to grow with them? How could we retain them and enable them to grow? So in the early days, for example, Stripe Connect, that is the marketplace product on how you move money from marketplaces. We originally built that for Lyft because Lyft came to us and was like, hey, we're building a marketplace for writers, but we don't have the infrastructure to do it.

20:44Can you help? So we built that. And now Connect is one of the biggest businesses for Stripe because we enabled all the marketplaces. But more importantly, we retained the early customers and enabled them to grow, right? I think in many ways for the new kind of infrastructure providers that are emerging in this agent first world, Argentic world, I think if you can hook in and enable the agents, there's going to be many more agents and humans doing our work, right? And we think that there's an opportunity for you to kind of ride that capability on lock. But at the same time, we're also seeing companies like building payments for agents because I want my personal agent to go shop for me and order like all the things.

21:23So I think we're going to see agents playing a lot of different interesting roles, but the people that can enable agents from an underlying infrastructure perspective, with everything that they're going to have a very interesting market ahead of them.

21:35Kenneth Auchenberg:Yeah, it's really, really interesting. And definitely lots to think about that for founders who are kind of building new solutions. How do you become top of mind for the agentic ecosystem? And Kenneth, you've obviously spent a lot of time in Copenhagen, but you're now kind of New York fund and you've obviously had experience across both the US and Europe. What do you think is the kind of difference between the two ecosystems in the tech world at the moment? And, you know, there's a lot of doom and gloom about Europe at times, and some of it's completely blown out of proportion. But I think we are always a little bit behind on the US on certainly their ambition and their sort of positivity and all those sorts of things.

22:22Kenneth Auchenberg:What is missing? What's misunderstood? Give us your perspective from having great context of both environments. Yeah, you know, like as a European who now spent the past decade in the US, I definitely have dipped my toes into both cultures. You know, I think fundamentally, like the difference between between America and Europe, and I'm from Denmark, and we kind of have the tall poppy kind of syndrome is that you should not dutch out. And Denmark's success is to be a middle manager at Nova Nordisk, and then you kind of made it, right? Which is great, don't get me wrong. But if you want to start a company, you're definitely the odd one out.

23:00That has changed in the past 15 years, and we're actually celebrating entrepreneurship much more in Europe, which I think is very healthy. But the difference in North America is that people here just take inherently more risk, and then people celebrate risk. and then everyone you talk to want to create something, etc., which is to celebrate culturally. I think when it comes to European tech, and especially what I'm focused on when it comes to infrastructure and AI, if I were to generalize in very rough kind of ways, I think Europe traditionally from a tech perspective has been using primitives and platforms built in North America to verticalize SaaS.

23:38And an example would be that in Europe, you can go to any kind of country and there's like the big accounting player running on American cloud infrastructure, etc. I think the opportunity for Europe ahead and also what we're seeing now with the macro environment is really Europe building its own setup infrastructure, building like the right primitives to how to build software, etc. And I think the whole European cloud infrastructure decoupling and the sovereign cloud in this new geopolitical reality, I think is definitely going to create a lot of opportunity for Europe to kind of build its own infrastructure and own primitives.