Mirror Particle Is Building a ‘World Model’ of Human Behavior

Mirror Particle is developing an AI model designed to predict consumer behavior and explain the reasons behind it. The two-year-old San Francisco startup says it is building a “world model” from scratch, based on how people perceive and interact with the world.

Co-founder and CEO Abhivyakti Ahuja argues that systems based on large language models (LLMs) fall short because they model written language rather than human visual perception, spatial reasoning and social intelligence. Mirror Particle instead aims to model how people change over time, using longitudinal data to track shifts in motivations and the experiences that may prompt them.

The company says its system combines client customer data with information such as current events, pop culture and social media. It focuses on “revealed behavior”—what people do rather than what they report in surveys—and targets market research, brand strategy and product strategy.

In an early pilot, a pet food brand asked what imagery to put on its packaging to boost sales. Mirror Particle’s analysis indicated that the imagery was not the issue: the brand’s highly recognizable image was perceived as mass market and cheap, and sales would plateau until that perception changed.

Mirror Particle has raised an angel round and says it is close to closing its first venture round. It is also competing in TechCrunch’s Startup Battlefield. Ahuja’s longer-term vision is to move from population-level analysis toward individual insights.