A series · 6 parts
Building LEANSpark
Building an AI co-founder in public, from the idea spark to a shipped product, starting November 2025.
- Part 1 Why I Waited Three Years to Build an AI Co-Founder ChatGPT launched in November 2022 and I spent three years saying no. What changed my mind was not a better oracle. It was a loop.
- Part 2 Why I Stress Test Business Models Before Talking to Customers An advisor once killed my business model in five minutes, nine months too late. Now I run that conversation first, across seven dimensions, before any customer hears the idea.
- Part 3 Why I Plan in 90-Day Cycles Your 3-year goal is too far away to drive daily decisions, and your task list does not add up to a strategy. The missing middle between them is a 90-day cycle.
- Part 4 How I Signed 209 Customers Before Building the Product The Demo-Sell-Build campaign from the inside: a concierge MVP, a demo of software that did not exist yet, and 40 days from zero to 209 customers.
- Part 5 Why I Stopped Building an AI Coaching Tool Two months of problem discovery interviews revealed that founders were not starving for frameworks. They were drowning in tools with amnesia.
- Part 6 Why I Build for Wants, Not Needs Features are needs and outcomes are wants, and customers care far more about their wants. How that one distinction shaped an entire MVP.