Simile
Stanford spinout building human-behavior simulation for enterprises. Trains a foundation model on real interviews, transaction history, and behavioral-science literature, then populates multi-agent simulations with consent-based "digital twins" to forecast how populations respond to decisions before they go live.
