We had a handshake, a Notion doc full of honesty, and absolutely nothing to build.
Dan and I were finally co-founders. We had the alignment, the trust, the shared hunger. All that was missing was an idea. And for two people who’d spent years watching problems go unsolved, ideas were the easy part.
We had three.
The first was prompt injection detection. It was the early days of the AI hype, and at LayerX I’d seen firsthand how companies were scrambling to monitor what employees were doing with AI chatbots. It felt timely. But it didn’t feel like mine.
The second was browser-based process automation. Instead of building another workflow platform with a hundred integrations, we wanted to automate tasks directly on the user’s browser. Connect to enterprise systems like Okta for permissions, then handle the work right where people already were. No new tools, no migration. Just automation that lived inside the tab.
The third was process mining. At Tonkean, I saw the same thing over and over: customers had no idea what their actual problems were. They couldn’t tell you which processes were broken, which tasks were repeated, where the bottlenecks lived. But at LayerX, the data was already there. If you could surface it, teams could finally automate based on evidence instead of guessing.
The last two weren’t just ideas. They were problems I’d lived inside of.
We picked our target: operations teams. It felt obvious. My entire career had orbited that world. Tonkean was workflow orchestration for ops. The pain points I’d seen were their pain points. Of course they’d want this.
Obvious is the most dangerous word in a startup.
We started talking to people. Ops leads, process managers, heads of business operations. The conversations were polite. Sometimes even enthusiastic. But never urgent. Nobody leaned forward and said, “I need this. When can I start?” Nobody offered to be a design partner.
It didn’t help that we barely knew how to run those conversations. We were pitching, not listening. It took us a while to pick up The Mom Test, and even then, I adapted quickly but Dan struggled hard with it. Old habits. He wanted to explain the vision, not sit with the silence after an honest question.
We weren’t listening for what the market was telling us. We were listening for what we wanted to hear.
So we adjusted. Not the target, just the angle. We moved from process mining to automation to intelligence and back again. Same persona, different pitch. We renamed the startup. Butler AI became Jeeves AI became something else I can barely remember. Each new name felt like a fresh start. It wasn’t. It was a costume change on the same unsolved problem.
One hundred conversations. Two hundred. Five hundred. Every batch felt like “this time we’ll find the signal.” We never did.
These weren’t pivots. A pivot means you learned something and changed direction. This was pacing in a locked room. We kept changing the name like it was the name that was broken.
Nine months passed. Then twelve.
The excitement from those early days was gone. Dan was frustrated. I was frustrated. But neither of us said the word “stop.” I told myself we just needed the right angle, the right framing, one more batch of calls. Stubbornness dressed up as dedication.
I know now that I didn’t have one moment where I saw the truth. I just ran out of road. It would take another year and a half, and a completely different job, before I’d finally understand that operations teams were never our people. The ideas were real. The passion was real. But we’d been shouting into the wrong room.
Somewhere in the middle of all this, we convinced ourselves that what we were missing was a third person. Someone to own the business side. We thought we needed a woman to balance the dynamic. So we searched. For months. We met a few candidates along the way, people who could have been interesting. But Dan never felt it. None of them were right, he said. Eventually the search just faded out, one more thing that went nowhere.
By then, the ideas were stale. The passion that made us co-founders was now the thing weighing us down.
I could feel Dan drifting. Not in anything he said, just in the silences. The energy between us had changed. He was going through a rough patch with his girlfriend at the time, and I started to worry that everything was unraveling at once.
So I suggested we have dinner together. The four of us. My partner and I brought Indian food to his place. The evening was nice. But the tension between Dan and his girlfriend was thick. You could feel it in the pauses, in the way they didn’t quite look at each other.
The next day Dan was sick. He stayed sick for almost a week.
When he came back, he told me they’d broken up.
I asked him what he needed. More work, or less. He said he’d take a few extra days off. In a startup with two people, two weeks is a lifetime.
When he finally came back, he said he wanted to focus on his relationship. The relationship he no longer had. He was hopeful they’d get back together. They did, eventually. And broke up again. And got back together again. That cycle went on for about a year.
But I didn’t know any of that at the time. All I knew was that my co-founder was choosing something else.
He was ambiguous about it. Wouldn’t say the words. But one thing I’d learned from The Mom Test, and from Avi back at the campus, is that sometimes you need to be direct. You need to push for certainty, even when the answer might hurt.
So I pushed. And he confirmed.
He was out.
I don’t know exactly what I felt. Betrayal, maybe. Anger. Confusion. Disappointment in someone I’d believed in. All of it tangled together.
But one thing I know for sure: I was lost. And I was alone.