I've been catfishing my users for 6 months. Full disclosure: the hot wellness oppa thirst traps? All me.
The AI handles health advice. But every image, storyline, and narrative that makes Maibel feel like a companion who gets you? I've been manually creating all of it.
Shipped 3 versions in 6 months. Still no moat, but I'm finally learning what not to automate.
Broke my 3-week hermitage at Future Foundry last week with a panel on AI-driven product building. Rapid-fire moderation by David Isaac Mathews, Jay Teo on moats, Ayan Sircar on problem clarity, Idris Murat Atakli, PhD asking:
"Can we rehumanize this whole process?"
Here's what I've been doing wrong: trying to automate everything at once.
My original vision for Maibel: self-generating narratives, automated images, dynamic storylines (think: spontaneous thirst traps by your wellness oppa) PLUS clinical-grade wellness advice for women.
Version 1: Failed.
Version 2: Failed better.
My problem wasn't execution... it was believing I could hit every milestone simultaneously.
Jay shared how proprietary data is one component of a strong moat.
But I was so busy engineering the perfect moat on all fronts, I forgot to solve ONE problem really well first.
Ayan crystallized it: being clear about the problem to be solved.
I wasn't clear. I was ambitious. (Turns out delulu is NOT the solulu.)
So I disappeared. Three weeks of productive isolation.
Stripping it all back taught me: I can't automate emotional resonance yet.
The storylines. The images. The feeling of having a companion who gets you? I had to painstakingly design it all. Manually.
Loris Lunardi talked about clear acceptance criteria. Mine clicked: emotional resonance FIRST.
Make wellness feel fun, not another metric to optimize. Progress and accountability without tracking anxiety.
Version 3 focuses on one thing: making women feel seen. Consistently.
50+ women in closed beta. 14-day retention holding.
Not because the AI got smarter. Because I stopped trying to solve everything and focused on this: how do I make women actually want to come back?
Idris asked about rehumanizing AI.
My answer: know what NOT to automate.
While everyone's racing to AI-ify everything, I'm learning which parts need human intention over algorithmic efficiency.
The self-generating oppa? Version 4 territory.
Building in isolation has its season. So does returning to community.
This is my season of failing in public, disappearing to fix it, resurfacing when there's something real.
Version 3 isn't final. But it's the first where I know what success means: women coming back because they WANT to, not because they should.
The moat isn't technology. It's understanding which problem deserves solving first.
To founders in the messy middle: what are you building all at once that's better built in stages?
And if you're building in AI x women's wellness, let's talk.
Hi, I’m Mabel.
I’m the founder of Maibel. I write about emotional AI, women’s wellness and learning to build the things I wish existed.
More about me
