The $2B AI Startup Simulating Human Behavior with 85% Accuracy

Simile creates foundation models of human behavior by using generative AI to simulate diverse populations, enabling organizations to predict human reactions with over 85% accuracy. By training agents on representative data—including life stories and behavioral responses—the company provides a digital environment where businesses can test products, messaging, and policies instantly. This approach moves beyond traditional polling by modeling the complex optionality and environmental factors that drive real-world decision-making. Founder Joon Sung Park emphasizes that these simulations allow companies to identify counterintuitive insights and causal mechanisms, shifting the focus from simple prediction to actively shaping future outcomes. By replacing costly, slow research methods with scalable agentic workflows, Simile offers a new paradigm for understanding consumer behavior and societal dynamics across industries ranging from retail to finance.