Mirror Particle launches foundation model to predict human behavior

Mirror Particle is building a foundation model to predict human behavior. The San Francisco startup is developing its world model from scratch to simulate how human behavior changes over time using longitudinal and revealed data.

Founded two years ago, Mirror Particle provides brands with an AI engine that predicts consumer behavior and the reasons behind it. Co-founder Thomson Yen and CEO Abhivyakti Ahuja lead the company. The model focuses on longitudinal data to track how people change, what triggers change, and to what degree. It emphasizes 'revealed behavior'-what people actually do rather than self-reported survey answers. The prediction engine provides the 'why' behind actions, including motivations, constraints, and context.

The company uses a proprietary combination of client customer data, current events, pop culture, and social media. It analyzes groups as systems where motivations and choices shift with experiences, not as fixed demographics. The long-term vision is to be a general layer for anticipating human behavior, moving from population-level to individual-level insights.

Contrasting with LLM-based approaches

Mirror Particle argues that fine-tuning large language models for demographic role-playing is insufficient for capturing real-world behavioral shifts. Abhivyakti Ahuja compared using LLMs for behavior prediction to "bringing a super soaker to Niagara Falls." She stated that LLMs are trained on hundreds of billions of data points, questioning how much behavior can be influenced by fine-tuning with small datasets.

Ahuja said LLMs model written language, but humans rely on visual perception, spatial reasoning, and social intelligence. Relying on LLMs gives insights based on what humans don't notice, which is irrelevant for predicting behavior. The company wants to capture the changing person, not the static person. If people aren’t changing, that is also a signal. Ahuja compared the model’s evolution to how a baby learns: from vision to language to body awareness to social intelligence.

Early validation and market positioning

In a pilot with a pet food brand, Mirror Particle found that brand perception-not packaging-drove sales. The issue was the brand being seen as mass market and cheap, not its packaging imagery. This validated the startup's focus on underlying motivations. Initial business areas are market research and brand and product strategy. The go-to-market strategy targets budgets in these sectors.

Funding trajectory and competitive landscape

Mirror Particle has raised an angel round and is close to closing its first venture round. The company is competing in Startup Battlefield 200 at TechCrunch Disrupt 2026. Ahuja studied at the University of Toronto and was inspired by AI pioneer Geoffrey Hinton’s work. She previously worked at Amazon Robotics. Co-founder Will Song has experience building sales personalization engines.

The competitive landscape includes well-funded rivals. Aaru raised $88 million at a $1 billion valuation over the past year. Humans& announced a $480 million seed round in January at a $4.48 billion valuation and launched Persimmon, a model to simulate human behavior.

Mirror Particle is set to compete in Startup Battlefield 200 at TechCrunch Disrupt 2026 in San Francisco on October 13-15, 2026.