LEX / Personal build / 2025 to present
A legal marketplace where merit beats media spend.
What I found
Legal directories sell visibility to whoever pays the most. A person in the middle of a legal crisis gets matched to a marketing budget, not to competence, and the attorney who wins the client is the one who won the ad auction. The incumbents have no incentive to fix this. Their revenue is the ad auction.
What I built
LEX: a marketplace where attorneys compete on verified performance instead of ad spend. AI-powered intake turns an unstructured crisis into a Case Passport, an organized case file that saves attorneys hours before the first call. Clients review and choose attorneys on transparent metrics. The business model is a flat subscription, designed from the start around attorney advertising and fee-splitting rules rather than retrofitted for them. I built it personally, end to end: client portal, attorney application, and provider application on a React front end with a Python FastAPI and Postgres back end.
What it made
A working platform, demo-complete across all three applications. This one isn't a revenue story yet, and I'm not going to pretend otherwise. It's proof of how I operate: see a market structurally rigged against its own customers, design the fix, and personally build it to working software instead of writing a deck about it.
The thesis in practice
The book argues that when platforms auction attention, the customer loses to the highest bidder. LEX is that argument turned into architecture: strip out the ad auction, and make performance the thing that compounds.