Rabbit Hop

Rabbit Hop sends AI shoppers with researched personas to shop live beauty stores and measures how faithfully each one behaves like the person it represents. I'm Akshaya, an Executive MBA student who loves skincare and makeup. Across customer discovery work in different domains, I have always wondered whether agentic infrastructure could automate part of it. The problem: customer discovery is slow and manual, and surveys capture what people say they do, not what they do on a real store. Generic AI shopping agents optimize for the best product, not for behaving like a specific shopper, and vary run to run. How it works: five researched archetypes are versioned and locked. Each runs on Muse Spark with computer use, plans as its persona (where to start, how to search, what evidence it needs), optionally draws on a 100-purchase shopping history, and shops Sephora or Olive Young US in a real Chrome window, recording every decision with a reason. A blind judge and repeated runs under identical conditions score persona fidelity and consistency.