The AI Co-Founder: When Agents Replace Headcount

Machine-generated transcript from the episode video. Speaker labels are not included and some names and product terms may be transcribed phonetically.

[0:00] I've learned that the best founders are always selling. Welcome to the GTM Vault podcast hosted by Rick Kleta. We uncover the strategies, the pivots, and the breakthroughs turning startups into giants. Let's crack open the vault and find out how. Today on GTM Vault, I'm joined by Ben Collins, co-founder and CEO of W, a YC company, an AI platform that enables anyone to build and scale a software business without writing a single line of code. We dive into how a team of AI agents can replace the need for a technical co-founder. What it takes to go from idea to market in days, not months, ascends from building AI first companies in the post chat GPT era. The future of entrepreneurship when anyone can build a company from scratch.

[0:54] Whether you're a founder, operator, or investor, this is a front row look at how AI is removing the barriers to building and scaling products. So, I want to get right into it. What was the aha moment that led to WAS, Ben? You know what I see a lot? Companies with solid strategy, but terrible execution. They've got dashboards, reports, tools everywhere, but no one knows what to actually do next to scale their business. That's why I've been paying attention to what Zoom Info is doing. They're not just the contact data company anymore. They built a full system of execution. They're calling it GTM Intelligence, and it actually works the list, writes the outreach, and triggers the play. No guesswork, no manual grind, just pipeline moving, predictable growth strategy that actually delivers. Check it out at zoominfo.com.

[1:46] Big question. Diving right in. Well, really excited to be here. Really excited to chat with you. The aha moment, man. I think there's like many, many aha moments along the way, right? My co-founder and I known each other for a long time. We were kind of ideulating in this space for for many months, explored many different ideas. And I think you're kind of zigging it zigging and zagging, you know, learning from every one of these ideas you explore and kind of like reccalibrating. And uh eventually we we kind of started with the idea that hey the way businesses are being built at least many types of businesses is changing with the use of kind of AI tooling that really accelerates and compresses different parts of the business building process. So what are the implications of that?

[2:29] implications are well we can experiment with many many businesses much faster. And so our first idea was how could we launch many many businesses at the same time in u kind of like a studio model. And then eventually we we started in kind of researching and archetyping what that tech stack might be to let us do that to launch and scale many many businesses at the same time. Um, over the course of a couple months, it just became apparent that the size of the opportunity in front of us was much larger than could be kind of captured in a portfolio of companies under one studio and that it made a lot more sense to kind of open up the platform and allow anyone to to build businesses on top of it. And so that was kind of the the evolution that that led us to to launching was last year.

[3:18] How did your time in VC and partnerships shape the way you approached to building this company? Yeah, that's a great question and I think it's going to be even more important as we move forward. I think a lot of lot of tools so far have seen massive massive adoption just because of the unique value they bring. There's a kind of a wow factor like an aha moment for a user, right? We've existed in a pre-AI world before, especially let's give the example of like a non-technical person writing code, right? That's always kind of been a barrier for you. Now all of a sudden you have the kind of cursor winds surflet lovable tools like this come in. Oh my god. All of a sudden I just built built a website. I just built a tool for my business. Like that was mind-blowing.

[4:02] And so there hasn't needed to be a lot of go to market or partnership thinking at those companies because they've just been like so propelled by the u awe and amazement of the users. So I think we're very quickly entering an age where go to market partnerships being really really creative in inventing new business models and new ways of scaling are are are just going to kind of kind of prove to be more and more valuable. And I'm definitely leaning on my kind of early career years working in venture capital, working in partnerships and growth to to figure out how we can we at WS can can leverage our strengths and our comparative advantages and tie those with partners who we can complement. So yeah, I think that's going to be a really really big part of the journey moving forward and something that will help differentiate us.

[4:50] WA's positioning is bold. Build and scale a software business without a technical co-founder. How do you deliver on that promise? Yeah. Uh well, fortunately, we have a team of really uh really experienced engineers who kind of uh in a way productize their expertise and now with AI, we can kind of kind of customize that expertise and allow anyone to to really tap into it. So, it's a lot. It's not that, you know, we have an incredible amount of technical expertise and experience. Our team has, you know, combined, I want to say, 10 decades of experience building and scaling products from zero to one, scaling them to, you know, tens of thousands of users. So, we've been there. We've looked around the corners and see the the pitfalls that exist and the traps you run into.

[5:37] We're able to build that into our platform and then by doing so allow others to uh to to hopefully avoid them. So, it's not that there expertise doesn't exist. It very very much does. We're just now able to kind of productize that and embed it into our platform. Walk us through how W's AI agents actually work in practice. Yeah, I think agents is kind of a word that gets tossed around a lot to mean a lot of different things, but like in its nature, it's actually kind of a a pretty simple thing uh in its truest form, right? You're connecting an LLM to a tool and allowing it to use a tool. So you can think of what the W platform is is, you know, it's kind of similar to how an engineering team and a product team work at any other company. There's people doing many many different tasks and then we basically combine that in a network. So there's a lot of agents that are doing design related tasks. There are agents that are doing coding related tasks. There are agents that are doing,

[6:39] you know, quality assurance related tasks. And so you kind of just add on many many layers of agents, many many layers of LM LLM using tools. You kind of chain those all together and an outcome is a a really, you know, you're you're able to produce really high quality things purely with with AI in a system that kind of looks quite similar to what a what a traditional company building a a software product looks like. How do you handle the handoff between automation and human decision making in the product? Yeah, that's a great question. And that's probably like the the core question um when it comes to the whole industry and the whole vertical of like AI generated code, right? We're still in the era where um there is a last mile that needs to be filled, right? And so most of the tools that have seen the most adoption so far, think about like a cursor or wind surf or they're they're intended for developers because developers can close that last mile, right? If the AI can

[7:43] generate a 80% version, well then the developer can come in and close that last 20%. Um, and so that's kind of the trick. That's the that's the core question that kind of faces itself to every every company building in an AI code related space is how do you close that last mile? So you either allow developers to close that last mile or you build sufficient engineering tooling um kind of around your product to either close as much of that last mile as you can and then like fingers crossed it's good enough or you plug in a human to do some level of quality assurance. We've seen some really interesting and unfortunate kind of situations occur on, you know, Twitter, LinkedIn where non-technical people use these tools that do the 80%. They try to do the last 20%. They're they're lacking kind of the the knowledge and expertise to do it and then uh their tool gets hacked, personal information is exposed. It's really not a good situation, right? So what we do at WAS is we always have a human layer

[8:48] of quality verification at the end before anything gets published. We just think that's the responsible thing to do and we can kind of sign our name next to the quality of products that are that are built on was because we require that level of human verification at the end. What's the hardest technical or product challenge you faced so far? Yeah, that's a great question. man there's many many difficult uh product and engineering challenges that have arisen and I think you know it's like why is that the case I think the answer is that this technology is just emerging and evolving very quickly so the ecosystem of kind of like infrastructure players and like tooling players that exist in more mature businesses like think if you're just like building something on the internet or you're building like some piece of SAS the all the tooling and the ecosystem around this already exists right so you can kind of tap into those existing providers and put something together in this era of like AI software building that ecosystem is is now just coming together, right? So the hardest part is is picking the right vendors to use for

[9:53] different parts of your stack and then the challenges occur when you are developing ahead of where all these other all the other vendors are, right? Your capabilities are far beyond what they're able to help you with. And so the challenges arise when you know you need to make a decision. Okay, do we kind of wait for this ecosystem to catch up to us? Do we invest in building more proprietary technology to kind of layer on top of the existing existing vendors and the existing ecosystem of of technology providers. Who's the ideal user today? Is it non-technical founders, small teams or even enterprise innovation teams? Yeah, I think in the future I hope everybody. Right now it's soloreneurs.

[10:36] So you can think health coaches, doctors, consultants, personal trainers. It it's largely people who are running services businesses and want to layer on an element of kind of AI and productization so that they can scale beyond kind of onetoone sessions with a client. Have you seen any surprising use cases emerge that you didn't expect? Tons, man. I'm I'm always so impressed by people's creativity and entrepreneurial entrepreneurial spirit. So, I think there's there's tons of examples where I've been just like blown away by human creativity. I think what's been most interesting to us is actually kind of the the inbound enterprise engagement or interest we get from, you know, big established companies who, you know, probably just like aren't aren't IT companies in general, right? they don't necessarily have the internal expertise or bandwidth to want to kind of build things themselves. So that that's been actually the most exciting and surprising has been kind of the inbound interest from the enterprise level.

[11:34] How do you ensure users not only build something but also find a market and scale it? Yeah, look, that's an area where where we intend to to help out as much as we can. We can help we can help people reach their customers by leveraging our brand and leveraging like the playbooks that we've built. But I think that's kind of our whole vision for for what the future of AI enabled entrepreneurship looks like is that that ability to to find your customers to to build a brand that's really what's differentiating and I think that comes down that still comes that's still the responsibility of the entrepreneur. So that that's uh that's how entrepreneurs can can differentiate in this new era for sure.

[12:14] WA is itself a GTM playbook in a box. How do you approach your own GTM? Yeah, that's a that's a great question and I think, you know, I probably won't share all of our our secrets and plans here, but you I think we can lean on we can lean on some examples from from past companies that have built kind of like platforms with verticalized solutions on top of the platform. We can think of something like I mean I think Shopify is kind of the the the best example and what they built for e-commerce in the age of the internet. But figuring out what those verticals are, I think that's probably the the biggest the biggest key decision in in the go to market. What are the right verticals to build on top of our massively enabling kind of horizontal platform? And then how do we incentivize an ecosystem to to help us build those verticals?

[13:04] What lessons have you learned from launching and scaling an AI product in a crowded space? Yeah, you have to move incredibly quickly and you always have to be reassessing your your hypothesis, right? I've never been a part of a technology landscape that's moving so quickly. I I don't think maybe there's never been one that's existed. And you always have to be reassessing collecting new data and and adjusting course. I think that's probably the the fundamental thing that is different about building this company than than building any of the previous companies I've been involved with. How do you think AI native companies will defer in their GTM compared to traditional? Yeah, look AI I mean it's going to be it's it's it's massively deflationary across the board. So you do need to redefine what your go to market is, right? The economics I think of, you know, going to market with a big sales team won't necessarily work out for many AI first companies. So yeah, there there needs to be a total re redefinition of of what that go to market playbook looks like. I think we're seeing a lot of

[14:06] companies already start to try to redefine that in terms of building a brand first and then figuring out what products to build. That's one angle that has been largely explored. But even in the last 6 months, we've seen kind of an evolution of like what works and what doesn't work. And when one thing works, it's almost like there's a whole herd of fast followers who replicate it. And then and then it's no longer interesting and it no longer works. So then you got to come up with your next kind of got to come up with your next uh go to market hack or trick. Yeah, incredibly competitive and people are incredibly creative when it comes to putting together these go to market hacks. How close are we to fully autonomous autonomous company creation?

[14:48] I mean we're you could build something very simple that's fully autonomous today. Will it be will it be highly successful and will it be useful to many people? I mean probably not, right? I think if you ask an AI to to spit out some version of a product or company today, uh it will basically take the average of what it's learned across all the data it's trained on. So essentially the internet, right? And it'll spit out like the average of that. Now, what really makes these businesses unique is what what kind of unique insights uh you add as a human. And so for an AI to do that, it kind of needs access to data that isn't necessarily on the internet.

[15:32] So I I think look, will there be certain use cases where you where there's some corpus of data that isn't on the internet and you use an AI on that data to then prompt AI to create products and software around that like unique ideas maybe. But I think for anyone to build a kind of unique differentiated business, uh there's there's going to be human involvement for for a long time. And what do you think this means for VCs, accelerators, and startup ecosystems? We're seeing kind of a a new dynamic come out um in the venture landscape for sure, right? funds are are dumping huge amounts of money into these kind of like uh you know we'll call them like early winners in the AI space and like we're seeing companies raise uh you know three or four rounds in in a span of six six to nine months which is kind of unheard of before. So yeah, we're seeing a concentration of capital to um the

[16:35] largest kind of or concentration of deals going to like these large multi-stage funds. At the other the other end of the spectrum, you have kind of YC dumping, you know, investment into a couple hundred startups a year. And the nice part of that is that using AI tooling, we're in a period where you can kind of get to profitability and growth without needing that much capital. So I think there's there's probably going to be like a you know a a barbell effect where you have a lot of companies that raise very little money and grow and become you know profitable decentsized companies. And then on the other end of the spectrum you have these kind of just massive massive we're calling like anointed winners who are raising hundreds of millions billions of dollars from these you know large multi-stage funds. And I think we're just going to continue to see that dynamic play out even more. We're talking in 2030. What does W look like?

[17:28] I think W is going to be a of its hands in a lot of different products touching a lot of different companies. So I think this model is going to be really partnership heavy and uh I hope that we can help amplify many many established players in growing their business and also help bring into the entrepreneurial world many people who who've been left out in the past. Yeah, our ambition our ambition is big and uh there's a lot of work that needs to be done, but uh there's never been a better time to uh to do it. Segue over to the rapid fire section of our podcast. First business you ever started? First business I ever started. Yeah, this is awesome. I come from I I have a lot of cousins and we grew up together on the east coast and we would spend the summers together and so I'm the oldest of all my cousins and so what I would do is on um recycling and trash day I would borrow my dad's pickup truck and bring put like three or four of my cousins in the back back of the pickup truck and then they would go collect bottles and

[18:31] cans throw them in the back of the truck and then I would go return them um and then pay all the pay give my younger your cousins a couple bucks. But yeah, that was a that was a pretty great business when you've got cheap labor working for you. One GTM channel you double down on right now if you only had 10K. I think YouTube right now is um is is is a really strong channel for for multiple reasons. If if you kind of have a knack for for creating content, I'm seeing I'm seeing really really strong results from YouTube. Biggest AI tool you personally can't live without other than WAS. I mean I I think probably for for most consumers it's chat GBT and for me I'm on it you know probably multiple hours a day using it as a a thought partner andor kind of research agent almost all the time but yeah one overhyped AI trend founders should ignore there is so much AI generated content I think founders should avoid avoid really leaning into the AI generated content

[19:35] for kind LinkedIn, Twitter, etc. It's at a point where it's kind of very obvious what's AI generated and what's not. Or that that may or may not be true, but much content I much much of the content I see is obviously AI generated. Like maybe there's some really good stuff that I'm not picking up on. And for me, I just I just feel like that detracts from from the brand. Um, and so I try to lean away from that. a founder you think is building something underrated in Oh, I mean there there are so many. I mean I I was kind of amazed by the quality of our peers in our in our YC batch last winter. Um the thing is it just takes time for for them to emerge, right? Uh it takes time for all the the big winners to emerge. I would say that like founders who are going after kind of these old sleepy industries and just building AI first companies for for a sleepy industry whether it's in real estate or insurance or back office medical etc. I feel like th those are

[20:38] kind of going to be going to be home runs. It was just the the right play at the right time and I think if they can execute those will those will be great businesses. One book, podcast or resource that's influenced your GTM thinking? Man, I'm a huge consumer of of podcasts. Um, so I don't know if there's one in particular, but I I love kind of stealing lessons and snippets from as many as possible. Learning about I love the Acquired podcast and learning about kind of the sometimes non-intuitive way these iconic companies have been built. And I think oftent times those lessons like really push me to think more creatively about about go to market. I also love the the the Bill Gurley um and Brad Gersonner podcast. It's about less about go to market, more about thinking about the kind of market holistically um and what that means for our business at was and and kind of what lessons we can glean from the AI players that have gotten out fast um and have been kind of

[21:40] anointed winners. So yeah, I think you you take a little bit I don't know what's your what's your favorite book or or go to market podcast? I recently started reading you all Noah Harrari's new book Nexus. Okay. About like the history of communication and how storytelling evolved from clay tablets to a religion. one of the earliest vehicles that was used to kind of get people aligned and and have a common story that they can share. Helping me better understand the power of storytelling by getting a better understanding of the the history of how it emerged. Maybe I'll have to look Maybe I'll have to look at that one. Awesome. Thanks for the recommendation.

[22:25] Why is W's mascot a wizard? Awesome. I'm glad you asked that. Um yeah, it there's there's two parts to it. one is that we think the experience of using W should be a bit magical and so we wanted to convey that and at the same time I think a lot of what we're doing is imparting kind of decades of experience and expertise and wisdom into our system and so we want our users to be able to have a magical experience but at the same time feel like they can trust in the guidance that they are are receiving and that it'll be guiding their business toward success In one sentence, the future of building companies is never been brighter. Man, there's so much opportunity and I think there's never been a better time to build a business.

[23:09] Ben, this has been a master class on building businesses with AI. You can learn more about Ben's work and try W for yourself at withw.com. If you enjoyed this conversation, subscribe to GTMvault on Spotify, Apple Podcasts or YouTube for more deep dives with the builders shaping the future of go to market. Thanks Rick. Tech founders and VCs careers lessons GTM