I didn’t start in AI leadership. I started in the trenches—co-founding two SaaS companies, watching one nearly die during COVID, building the other from a $10K month to $2M in annual recurring revenue, hiring over 20 people, and learning every painful lesson the hard way. Those experiences broke some of my worst instincts and built the ones that actually matter.
Today, when I sit across from a client who’s trying to figure out their AI strategy, I’m not drawing from a framework I learned at McKinsey. I’m drawing from scar tissue. And honestly? The scars are worth more than any framework.
Gleantap: From $10K to $2M (And Almost Zero in Between)
Gleantap was a customer engagement platform for fitness businesses—gyms, studios, personal training chains. We built AI-powered marketing automation that helped these businesses retain members and sell more services. The pitch was simple: stop losing customers you already have.
When we started, I was doing everything. Writing code, running sales calls, handling support tickets, building pitch decks at 2 AM. The first year was the classic startup grind—long hours, small wins, the constant question of whether we were building something anyone actually wanted.
We hit $10K in monthly recurring revenue and I thought we’d figured it out. We hadn’t. We’d just found a few early adopters who were willing to put up with a half-finished product because they liked us personally.
The difference between $10K MRR and $2M ARR isn’t more of the same. It’s fundamentally different. Getting to $10K requires hustle and a product that kind of works. Getting to $2M requires systems, repeatable sales processes, a product that works reliably without you babysitting it, and a team that can execute when you’re not in the room.
The COVID Pivot
Then March 2020 happened. Every gym in the country shut down overnight. Our entire customer base—hundreds of fitness businesses—went from paying us monthly to wondering if they’d survive the month. Churn spiked. Revenue dropped 40% in six weeks.
We had a choice: hunker down and wait, or pivot hard. We pivoted. Within three weeks, we rebuilt our platform to support virtual fitness—livestream classes, on-demand content, digital membership management. We shipped features in days that would normally take months. We slept in shifts.
The COVID pivot taught me something I now tell every client: your best product decisions will be made under constraints, not in brainstorming sessions. When you have unlimited time, you overthink. When you have three weeks to survive, you build what matters.
That pivot saved the company. But more than that, it reframed what we were. We stopped being a “gym marketing tool” and became a “member engagement platform.” That broader positioning opened doors we didn’t even know existed. Within a year, we’d not only recovered—we’d grown past our pre-COVID numbers.
Mastera: The Churn Problem That Nearly Ate Us
Mastera was different. It was a platform for content creators and educators to build and sell online learning experiences. The product was good. The market was real. But we had a churn problem that was slowly killing us.
10% monthly churn. For context, healthy SaaS churn is 3-5% monthly. At 10%, we were essentially replacing our entire customer base every 10 months. It didn’t matter how many new customers we acquired—the bucket had a hole in the bottom.
We tried the obvious things first. Better onboarding emails. Feature announcements. Discount offers for at-risk accounts. None of it moved the needle meaningfully. The churn rate budged from 10% to about 8.5%. Not enough.
The breakthrough came when we stopped looking at churn as a retention problem and started looking at it as a product-market fit problem. We weren’t losing customers because our customer success was bad. We were losing them because a segment of our customers were using the product for something it wasn’t built for.
We dug into the data. Customers who used our course-builder tools churned at 4%. Customers who signed up primarily for our community features churned at 18%. We had two completely different user bases with completely different needs, and we were trying to serve both with one product experience.
The lesson: Aggregate metrics lie. A 10% average churn rate hid the fact that half our product was working beautifully and the other half was failing. When a client shows me a dashboard with averages, my first question is always: “What does this look like when you segment it?”
We made the hard call. We doubled down on the course-builder audience, redesigned the onboarding to qualify users earlier, and gently sunset the community-only features that were attracting the wrong customers. Churn dropped from 10% to 2% in four months. Revenue per customer went up because we were serving a more focused audience better.
Hiring 20+ People (And What It Taught Me About Scaling)
Across both companies, I hired over 20 people. Engineers, designers, salespeople, customer success reps, a head of marketing, a VP of engineering. Some of those hires were brilliant. Some were expensive mistakes. Here’s what I learned:
Your first five hires define your culture permanently. I don’t mean in the mission-statement-on-the-wall sense. I mean in the “how do people actually make decisions when nobody’s watching” sense. At Gleantap, our first engineer was someone who shipped fast and iterated. That speed-first culture persisted through every subsequent hire. At Mastera, our first product hire was methodical and research-driven. That rigor shaped everything that followed.
Hire for the problem you have right now, not the one you think you’ll have in 18 months. I once hired a VP of Sales when we had 30 customers and needed 300. What I actually needed was a senior account executive who could close deals. The VP spent three months building sales processes for a team that didn’t exist yet, and by the time we needed those processes, the market had shifted and they were all wrong.
The people who get you from 0 to 1 are rarely the people who get you from 1 to 10. This is the hardest lesson in scaling, and it never stops being painful. Some of your most loyal early team members will hit their ceiling. Handling that with honesty and respect is one of the most important things a founder does.
Learning to Prioritize Ruthlessly
When you’re running a startup, everything is urgent. The customer who’s threatening to churn. The feature that a prospect says is a dealbreaker. The bug that’s affecting 5% of users. The partnership opportunity that needs a response by Friday. The team member who needs a career conversation.
Early on, I tried to do everything. I’d work 14-hour days and still feel behind. The turning point was when a mentor told me something that changed how I operate:
“If everything is important, nothing is. Pick the three things that will matter in six months. Do those. Let the rest be slightly broken.”
I started using a brutal prioritization framework. Every Monday, I’d ask: “If I could only ship three things this week, what would move the revenue needle most?” Everything else went into a backlog that I reviewed monthly. Some things in that backlog never got done. And you know what? It didn’t matter. The business grew anyway.
This framework is now central to how I run AI engagements. When a client has 15 AI initiatives they want to pursue, I make them rank-order them. Then we do the top three. Not because the others aren’t good ideas—but because doing three things well beats doing fifteen things poorly.
Why Founder Scars Make Better AI Leaders
Here’s what I believe, and I know it’s a strong claim: the best AI leaders are recovering founders. Not because founders are smarter, but because they’ve felt the consequences of bad decisions in their gut, not just on a spreadsheet.
When I tell a client that their AI roadmap needs to start with quick wins before tackling the transformational stuff, I’m not reciting a playbook. I’m remembering the time we tried to build a transformational feature at Gleantap before we had product-market fit, and it nearly sank us.
When I push back on a client who wants to build a custom ML model before they’ve validated the use case, I’m thinking about all the features I built at Mastera that nobody used because I was too in love with the technology to ask if anyone wanted it.
When I insist on user research before development, I’m remembering every dollar I wasted building things based on assumptions instead of evidence.
Founder empathy is an AI leadership superpower. When your client says “I know the data says X, but my gut says Y,” an outside advisor hears stubbornness. A former founder hears someone who has context that hasn’t been captured in the data yet. Both might be right. But only one of them will ask the follow-up questions that uncover the real answer.
What I’d Do Differently
If I could go back and give myself advice at the start of each company, here’s what I’d say:
Talk to 50 potential customers before writing a line of code. At Gleantap, we built for six months before doing serious customer discovery. At Mastera, we did three months. Both were too long. The insights that shaped our best features always came from conversations, not from internal brainstorming.
Hire a finance person earlier. I ran both companies without a dedicated finance function for way too long. By the time I understood our unit economics deeply, we’d already made pricing decisions that took months to unwind.
Say no to custom enterprise deals earlier. Both companies had moments where a big potential customer offered a large contract if we’d build custom features. We said yes more often than we should have. Those custom builds created technical debt that haunted us for quarters.
Invest in customer success from day one. At Mastera, we didn’t hire a dedicated customer success person until we had 200 customers. By then, the churn patterns were already baked in. If we’d started with proactive customer success at 50 customers, I believe we’d have caught the segment problem at least six months earlier.
I don’t regret a single one of those mistakes. Every bad hire, every wasted sprint, every feature nobody used, every sleepless night during the COVID pivot—all of it built the judgment I use every day with the companies I embed with.
When someone asks why I left the startup world, I tell them I didn’t leave it. I just found a way to apply what it taught me by embedding inside more companies, across more industries and more problems. The founder in me never turned off. I just pointed it in a different direction.
And honestly? Helping other companies avoid the mistakes I made—or at least make them faster and cheaper—is the most rewarding work I’ve ever done.