We failed. Twice. Before we built something people actually wanted.
I’m going to tell you the raw, unfiltered story of our three pivots in 12 months. You’ll see the fights, the fears, and the framework that finally put us on the path to product-market fit. This isn't a sanitized success story. It's a look at the messy reality of building an AI company when it feels like the ground is constantly shifting beneath your feet.
Everyone’s talking about moats in AI. About how to build a defensible business when the next foundational model could wipe you out overnight. I’ve lived that fear. I’ve also come out the other side. And I can tell you that the traditional moats are not what they used to be.
The Siren Song of the Shiny New Toy
Our first idea was born out of pure excitement. This was back when the first whispers of large language models were starting to get loud. We thought, “Let’s build a better tool for movie marketers.” My first company, MovieLaLa, had been acquired by Gfycat, so I knew the space. We had the connections. We had the domain expertise. We saw a problem: marketers were spending millions on campaigns with little to no data on what was actually resonating with audiences. We figured we could use AI to analyze trailers, posters, and social media chatter to predict box office success. A crystal ball for Hollywood.
It was a beautiful idea. We built a slick dashboard. We had charts. We had sentiment analysis. We pitched it to our network. They loved it. They all said, “This is amazing! So innovative!”
And then they didn’t buy it.
Why? Because it didn’t actually solve their core problem. Their problem wasn’t a lack of data. It was a lack of time and the inability to act on the data they already had. Our fancy AI dashboard was just another thing to look at. It was a vitamin, not a painkiller. We had fallen in love with the technology, not the problem.
Lesson #1: Your personal experience is a starting point, not a destination. Don't assume you know the customer's problem better than they do.
The “We’ll Just Add AI” Fallacy
So, we pivoted. We were bruised, but not broken. We looked at what we had built and thought, “Okay, the movie industry is too slow. Let’s go after a real business need.” My next company, RemoteTeam, had been acquired by Gusto, so I had a front-row seat to the explosion of remote work. The problem was obvious: managing a global workforce is a logistical nightmare. Payroll, compliance, onboarding… it’s a mess.
We thought, “Let’s build an AI-powered HR platform.” We’d automate everything. We’d use AI to answer employee questions, to generate contracts, to manage payroll across a dozen different countries. We were going to be the “Workday for the remote world.”
This time, we got closer. We had customers. We had revenue. But we were constantly fighting fires. The AI was good, but it wasn't perfect. And when you're dealing with someone's paycheck,
'almost perfect' is a recipe for disaster. We were spending all our time manually fixing the AI’s mistakes. We were a services company masquerading as a software company.
Lesson #2: AI is a tool, not a product. Don't just slap an AI label on an existing problem. Find a problem that only AI can solve.
The Pivot That Saved Us
I was ready to give up. Two failed pivots. A burned-out team. I felt like a fraud. I'd had two successful exits, but this AI thing was a different beast.
We had one last shot. We sat down as a team and we didn't talk about AI. We didn't talk about technology. We talked about pain. Real, visceral pain. What was the most painful, expensive, and recurring problem our customers had?
It wasn't payroll. It wasn't compliance. It was hiring. Specifically, hiring and retaining top talent in a global market. Companies were spending a fortune on recruiters, on job boards, on endless interview cycles. And they were still getting it wrong.
That's when it clicked. We could use AI to build a new kind of hiring platform. Not just a database of resumes, but a system that could actively identify, vet, and match candidates to roles. We could use AI to predict culture fit, to assess skills, to automate the entire top-of-funnel process.
This time, we didn't build a dashboard. We built a workflow. We became an extension of our customers' hiring teams. We didn't just provide data; we provided results. We found them the right people, faster and cheaper than anyone else.
And it worked. We found product-market fit. We have a real business, with real customers, and a real moat.
The New Moats in the Age of AI
So what did we learn? What are the new moats in the age of foundational models? It’s not about having the best algorithm. It’s not about having the most data. It’s about these three things:
Process Power: This is our biggest moat. We are so deeply embedded in our customers' hiring process that it would be incredibly painful for them to switch. We're not just a tool they use; we're a part of their team. We've automated a critical business process, and that's incredibly sticky.
Cornered Resource: In our case, our cornered resource is our proprietary data on what makes a successful hire. We have data on thousands of hires, across hundreds of companies. We know what skills, what experience, what personality traits lead to success in different roles. This data is our unique asset, and it gets better with every new customer.
High Switching Costs: This is a result of the first two. Because we are so embedded in their process and because we have so much data on their hiring, it would be a massive undertaking for a customer to leave us. They would have to start from scratch. They would lose all the institutional knowledge we've built up.
Stop Chasing Rainbows
Everyone is chasing the next big thing in AI. The next foundational model. The next breakthrough algorithm. But the real opportunity isn't in building the technology; it's in applying it to solve real-world problems.
Stop chasing rainbows. Stop falling in love with the technology. Fall in love with the problem. Find a painful, expensive, recurring problem and then, and only then, figure out how AI can solve it. That's how you build a real business. That's how you build a moat in the age of AI.
Frequently Asked Questions
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.