The Lab
The Lab.
Working builds, public product dossiers, and private-lab field guides. Some systems are live demos. Others stay private while I share the data model, mechanics, screenshots, and failures worth learning from. Same hands that have been building AI systems commercially since 2016.
An A/B test can be broken before it spends a dollar.
Marketing Experiment Preflight. A private marketing and revenue lab that pressure-tests the audience, control, evidence, outcome, and stopping rule before traffic arrives. The application stays private. The field guide explains the data layer, synthetic rehearsal, connector roadmap, and the learning loop I am building for MarTech, RevOps, and sales.
Track model spend and review consequential output.
AI Operations Control Center. Inspect agent activity, spending, output checks and review gates in a scaled-back demonstration of the operating system. Figures synthetic, machine real.
A sports dashboard can have every number and still miss the one that changes the game.
Brokaw Bets. A public NFL matchup, DFS, and prop-research dashboard built from public data. The live product is useful because it keeps the game, player, market, source, and time context together. The build dossier shows what that teaches me about MarTech, RevOps, and AI decisions.
A working proof in the prospect's own brand.
The Outbound Killer. It researches a named company from public signals, builds them a personalized working proof of their own business in their own brand, and mails it with a code they can scan. When they engage, it reacts on its own. Synthetic prospects, real machine.
Personalize a first visit without on-site purchase history.
Cold-Start Personalization. A first visit has no on-site browsing or purchase history yet. This demonstration composes a page from the shopper signals available at arrival. Synthetic shoppers, live decisioning.