This interview was originally published here.
Joey Gutierrez and Kyle Carriedo, Co-Founders and Managing Partners of Misfit Labs, are part of a new wave of operators redefining how modern technology companies are built and scaled through artificial intelligence. At a time when many startups and institutions are experimenting with AI but still operating within traditional development models, Gutierrez and Carriedo have focused on a more execution-driven approach, building companies designed around AI from day one. Through Misfit Labs, the two work directly alongside founders and organizations to design, build, and scale software products across healthcare, infrastructure, and enterprise technology. Drawing from their backgrounds across engineering, operations, product development, and venture building, they bring a systems-first perspective on what it takes to move from ideas to scalable execution in an increasingly AI-native world.
Based in Miami, they are focused on helping shape the next generation of company building while also working with institutions to launch internal venture studios and AI labs that drive innovation from within. In this conversation, they share insights on AI-native development, the evolution of venture studios, why execution remains the biggest challenge in startups, and how artificial intelligence is reshaping the future of technology companies.
Misfit Labs positions itself as an AI-native venture studio. What does “AI-native” actually mean in practice, and how does it change the way companies are built from day one?
Kyle: “AI-native” gets used pretty loosely right now, so we try to be specific about what we mean. For us, it’s not about using AI tools here and there, but rather that AI is part of how the product is designed and how the company operates from the beginning.
That changes things pretty quickly. You don’t build teams the same way, and you don’t plan timelines the same way. A lot of what used to require a full team can now be done by a much smaller group, which means the constraint shifts. It’s not about execution in the same way anymore; it’s about judgment.
You both come from different backgrounds across operations, engineering, and company building. How did those experiences shape the model behind Misfit Labs?
Joey and Kyle: The model is really just a reaction to seeing the same gaps over and over again. On one side, you have strong technical teams that can build, but don’t always translate that into a business. On the other, you have founders with a clear idea but no real execution behind it. We’ve both been around those situations from different angles. So instead of building something that sits on top of that, we built something that sits inside it. We work directly with founders as part of the team, helping make decisions and actually building alongside them. It’s much closer to a co-founder model than what most people think of as a venture studio.
Many startups struggle not because of vision, but because of execution. Where do you see founders breaking down most often when it comes to building and scaling products?
Joey: It’s usually not one big mistake, it’s a set of smaller ones that compound.
Teams try to do too much too early. They build things before they know if they matter. Early technical decisions get made quickly and then quietly limit what’s possible later. And ownership isn’t always clear, which slows everything down once there are more people involved.
AI doesn’t really fix those issues. If anything, it just speeds them up. You can move faster now, but if you’re pointed in the wrong direction, you feel it sooner.
Misfit Labs embeds directly into the companies it helps build. How does that level of involvement change outcomes compared to traditional venture or advisory models?
Joey: Most venture studios are still fairly capital-first. They fund, they advise, they take equity, but they’re not in the day-to-day of building.
We’re much closer to the work. We’re in engineering, product, and go-to-market with the team. That changes how decisions get made because you’re not speaking in hypotheticals. You’re seeing what actually works and what doesn’t. It also creates a tighter loop between strategy and execution. There’s less lag, less translation, and fewer things that sound good in theory but fall apart when you try to build them.
There is a growing conversation around smaller, more experienced teams building with AI. How do you see team structures evolving over the next few years?
Kyle: You’re already seeing the shift toward smaller teams, but it’s not just about size. It’s about the profile of the team. If AI is handling more of the execution, then what matters more is how people think, how they make decisions, and how they define problems. So you end up with fewer people, but a higher bar for each one.
People need to be able to move across product, engineering, and business, not just stay in one lane. And a lot of traditional management layers start to feel heavier than they need to be in that kind of environment.
You are working across sectors like healthcare, infrastructure, and enterprise software. What makes these areas particularly important, and what challenges do they present for builders?
Joey: These are areas where there’s real impact, but also real complexity. You’re dealing with regulation, legacy systems, and a lot of stakeholders. In many cases, trust and relationships matter more than speed. The challenge is that building the product is only part of the work. You also have to integrate into systems that don’t move quickly and don’t tolerate mistakes very well. That’s where a lot of teams get stuck.
Beyond venture building, Misfit Labs is partnering with institutions to develop internal venture studios and AI labs. What is driving that demand from larger organizations?
Kyle: A lot of large organizations know they need to evolve, but they’re not structured to build new things quickly. They have capital, data, and distribution, but their internal processes make it hard to move. So instead of trying to overhaul everything, they’re creating smaller environments where they can operate differently. That’s where internal studios and AI labs come in. It gives them a way to experiment, build, and launch without getting caught in their own systems. For us, it also creates a second lane of work alongside venture building, which is not something most studios are set up to do.
How do you think about balancing speed and quality when using AI in product development, especially in complex or regulated industries like healthcare?
Joey: You can’t treat speed as the only goal, especially in those environments.
AI makes it easier to move quickly, but it doesn’t remove the need for accountability. In areas like healthcare, mistakes have real consequences.
So the approach is more about being deliberate. Early on, you can iterate quickly and explore. But once you’re dealing with anything sensitive, you need more structure, more validation, and more oversight. It’s really about knowing where things can break and adjusting accordingly.
Looking at your early portfolio, what patterns are you seeing in the types of companies that are best positioned to succeed in an AI-driven environment?
Joey and Kyle: The ones that stand out are usually pretty focused: they’re solving a specific problem where AI actually changes the outcome, not just improves efficiency a bit. They’re built by teams that are comfortable moving quickly but still have discipline in how they build.
And they stay lean. They don’t assume they need to scale headcount in the same way companies used to, which gives them a lot more flexibility.
As you look ahead, what does the next generation of company building look like, and what will separate the teams that succeed from those that struggle?
Kyle: The biggest shift is that building itself is becoming less of the bottleneck. What matters more is deciding what to build and making good decisions consistently. The teams that do well will be the ones that can think clearly, move quickly, and adjust as they go. The ones that struggle will either overcomplicate things or treat AI like a layer on top instead of rethinking how they actually operate from the ground up.