The best products I've ever built started with one specific person and one specific problem. This one started with my girlfriend, and a question she had been carrying her whole life with no one to answer it.
A Problem That Existed Before She Did
My girlfriend is bi-racial — her father is African American, her mother is from Hiroshima, Japan. She is, by every measure, a beautiful person. And for twenty-one years, she didn't know what to do with her hair.
That's not a small thing. Hair, especially for Black women, is identity. It's heritage. It's the accumulated knowledge passed from grandmother to mother to daughter — which products, which techniques, which rituals protect the texture and celebrate it. It's something that travels through community and family, through aunties who sit you down and show you, through cousins and salons and shared language that builds up over time into something that works.
Her mother, raised in Japan, didn't have that knowledge. Not because of any failure on her part — because it simply wasn't hers to have. Japanese hair and African American hair are not the same, and the techniques and products that work for one do not carry over to the other. Her mother loved her deeply and had no idea how to care for her daughter's textured hair. That gap wasn't anyone's fault. But it was real, and it had consequences.
For twenty-one years, she navigated it mostly alone. No auntie. No community with that specific knowledge. No family tradition to draw on. She experimented, got it wrong, got it wrong again. Straightened it into submission, avoided styles she might have loved, made peace with a kind of helplessness around something deeply personal. The information existed somewhere in the world — curly hair communities, specialist stylists, decades of hard-won knowledge — but it wasn't organized in a way she could access it, and it wasn't tailored to her specific texture, her specific situation.
What Inspiration Actually Looks Like
I'm not going to pretend I had a clean, cinematic moment where the idea arrived fully formed. What I had was a long, gradual recognition that someone I cared about had a real problem, that the problem was shared by a lot of people I'd never meet, and that the tools to address it — the AI capabilities, the available knowledge bases, the ability to build something quickly and put it in front of people — were sitting right there.
The thesis was simple: what if the auntie, the community elder, the specialist with decades of experience — what if all of that accumulated knowledge could be made available to anyone, tailored to their specific hair type, accessible from their phone with no appointment needed?
The problem with existing resources wasn't that they were wrong. There is excellent curly hair content online. The problem was discoverability and specificity. A YouTube video for Type 4C hair doesn't solve the same problem as one for Type 3B. A product recommendation for someone in a humid climate doesn't work the same in a dry one. Generic advice applied to specific texture does damage. The knowledge needed to be personalized — not just aggregated.
Building the Thing
I built curly-hair-ai.com as an AI-powered curly hair analyzer: a tool that takes what someone shares about their hair, understands the texture type and specific challenges, and returns personalized guidance — care routines, product recommendations, styling approaches — along with links to the right communities, tutorials, and specialists for their specific situation.
The goal was never to replace the auntie. Community knowledge, passed person to person, has a warmth and a specificity that no model captures. The goal was to be the thing that gives someone a starting point when that community isn't available to them — when you're bi-racial and your mother didn't know, when you're far from your family, when you're at the beginning of understanding hair you've been struggling with for years.
What I cared most about in building it was the quality of the output, not the sophistication of the system. A technically impressive AI that gives bad hair advice is worse than useless — it sends someone down the wrong path with confidence. So I spent the majority of the time on what the system knew, how it asked questions, and how it connected people to real, vetted resources. The technology serves the guidance. Not the other way around.
Who It's Actually For
My girlfriend was user #1. She was the first person to use it and the first person whose feedback shaped what it became. Watching her interact with it — seeing the moment when a recommendation actually matched what she needed, when something that had been confusing for years suddenly made sense — is the closest I've come to the pure version of why I build things.
But she's not the only one. There are women who grew up in countries where their texture was in the minority and the products were never stocked on shelves. There are adoptees whose parents love them and have no frame of reference for their hair. There are women who cut it all off because not knowing what to do with it felt worse than not having it. The specificity of the problem — bi-racial, no community, no auntie, no one to teach you — is not rare. It's just quiet.
The best products solve a problem that exists before you build anything. The problem here existed long before curly-hair-ai.com. I just got close enough to one person's experience of it to understand what was actually missing — and lucky enough to be in a position to do something about it.