Why a $1M Startup Wasn’t Worth Building in the Age of AI

3 operating rules for building software in the age of AI: when profitable isn't enough, how to find hidden value nobody asked for, and why owning your data beats owning the model.

With Jean-Denis Greze

"We realized that the incumbents were going to be able to use the technology to automate tax prep for their existing customers faster than we could grab market share." — Jean-Denis Greze

Jean-Denis Greze, on the episode

Jean-Denis Greze was CTO of Plaid before he left to build an AI tax product for small businesses, a business that made nearly a million dollars in its first year and was tracking toward two and a half to three million. He shut it down anyway. The pop story here is the obvious one: founder pivots after a slow first product. The actual operating system underneath is sharper. Greze was running a model where the question was never just 'is this working,' it was 'does AI change who wins this market,' and the answer for tax was that incumbents with existing customers would always out-automate a startup trying to win share from scratch. That pivot became Town, an AI assistant that works out of email and calendar, and this protocol pulls the decision framework and product strategy Greze used to get there.