A trader that can't explain itself
Rick Cockerham has been building A.N.N.I.E since 2013: a neural network and a genetic algorithm, written from scratch in C, that reads the market each day and ranks more than 1,200 stocks for the week or two ahead. It trades with real money and has supported his family. It's also a black box. Even Rick can't say exactly how it reaches a call. He had a draft outline, shaped with his own GPT, and people ready to open doors with investors. What it needed was a story.
Proof before promises
His outline was already most of a spine. We filled it out from his notes, reordered it around proof, then designed the deck on top of it. Instead of trying to explain the model, the deck leads with what it has done: real-money results, not forecasts, under one line from the title to the close, "Performance, not prediction." The founder story, a comparison where ANNIE checks every box, and a moat built on time follow from there. (Figures on the slides here are placeholders.)
Her, not it
Everyone calls ANNIE "her", and an illustrated robot head, drawn from the logo, was already part of the brand. We kept Rick's wordmark and palette, enhanced and refined the robot head, and made her the face of the deck. Rick liked our version enough to ask for it for his own home page.
Built for the conversations that followed
The engagement ran four weeks, with the first full deck in review nine days after kickoff, about a week ahead of our own estimate. The last weeks were for Rick's first investor conversation: when its feedback came back, an updated deck followed within days.