The status quo for predicting human behavior relies on prompting or fine-tuning large language models (LLMs) to role-play as target demographics. Mirror Particle CEO Abhivyakti Ahuja argues this approach is fundamentally broken. 'It's like bringing a super soaker to Niagara Falls,' she told TechCrunch. 'LLMs have been trained on hundreds of billions of data points. How much can you influence its behavior by fine-tuning with such a small amount of data? It's still stuck in the past.'
Ahuja emphasizes that LLMs model written language, but humans are shaped by visual perception, spatial reasoning, and social intelligence. Mirror Particle instead builds what she calls a world model that captures the evolving person, tracking what triggers change and how motivations shift over time. 'We don't want to capture the static person,' Ahuja said. 'We want to capture the changing person. That means capturing the longitudinal data on how people are changing, what triggers are changing them and to what degree.' If they aren't changing, she added, 'that's also a signal.'
The model incorporates clients' customer data alongside current events, pop culture, and social media to simulate a demographic segment as a dynamic system. Much of the focus is on 'revealed behavior' what people actually do, rather than self-reported survey answers.
Mirror Particle has already raised an angel round and says it is close to closing its first venture round. The company is also competing next week in Startup Battlefield 200, a startup competition at TechCrunch Disrupt 2026 in San Francisco from October 13 to 15.
Like its rivals, including Simile, Aaru, and Humans&, which have collectively raised hundreds of millions of dollars in the past year, Mirror Particle's initial go-to-market strategy targets existing budgets for market research, brand strategy, and product development.
A broader shift is underway in the AI behavior prediction space, as startups move beyond static role-playing models toward systems designed to capture the messy, evolving nature of real human decision-making. Mirror Particle's approach, if successful, could offer brands a more dynamic tool for understanding consumer behavior, though it also raises questions about privacy and the use of personal data in such simulations.