Startups promising to predict human behavior are attracting substantial venture funding. Over the past year, Simile raised $200 million at a $2 billion valuation, Aaru secured $88 million at a $1 billion valuation, and Humans& launched Persimmon with a $480 million seed round in January at a $4.48 billion valuation. Two-year-old Mirror Particle, based in San Francisco, operates in this space with a different philosophy.
Mirror Particle believes the standard approach using large language models fundamentally misrepresents how humans make decisions. Co-founder and CEO Abhivyakti Ahuja compares it to "bringing a super soaker to Niagara Falls." LLM behavior is shaped by hundreds of billions of training parameters, making fine-tuning with limited data largely ineffective. The company argues these models capture written language, not human experience.
Ahuja states, "LLMs are modeling written language, but humans are made of visual perception, spatial reasoning, social intelligence." Relying on them produces insights based on what humans don't notice, which misses the mark when predicting actual behavior.
A foundation model from scratch
Mirror Particle is building what Ahuja calls a "world model" developed from the ground up. This model simulates why humans act as they do and how behavior shifts across time. The goal isn't to capture a static person but the changing person — tracking what triggers transformation and to what degree. Even stasis registers as a signal.
The system combines proprietary data: clients' customer information, current events, pop culture, social media, and more. Rather than relying on self-reported surveys, the model focuses on "revealed behavior" — what people actually do. This data streams form a demographic segment treated as a system that evolves through experience.
Practical applications
Initial market focus sits where budgets already exist for behavioral insights: market research, brand and product strategy. Mirror might help a beauty brand not only craft better ad copy for makeup appealing to Gen Z but also determine whether that demographic desires the product at all.
Ahuja asks, "What if [the target demographic] doesn't want eyeshadow palettes? Maybe blush is a better option to go for if you want to sell a product to this market."
The prediction engine also supplies the "why" behind behavior recommendations — motivations, constraints, and context that justify suggestions. This helps brands make smarter decisions. In one early pilot with a pet food brand, the technology identified the real issue: packaging imagery of chicken, beef, or vegetables was beside the point. The brand's recognizability as mass market and cheap was blocking sales until perception changed.
Development origins
Ahuja's interest in modeling the human brain stems from neuroscience and computer science study at the University of Toronto, where Geoffrey Hinton's neural network work inspired her. After school, she joined Amazon Robotics, building robots that build other robots. There she met co-founders Will Song, who spent years building sales personalization engines, and Thomson Yen, who used deep learning to study how AI agents understand human behavior.
The long-term vision positions Mirror Particle as the "general layer for anticipating human behavior," moving from population-level analyses to individual-level insights. Ahuja concludes, "We just need a better model of humans if we're going to work alongside AI and with each other."
The company has raised an angel round and is close to closing its first venture round. Mirror Particle also competes the week of October 13-15 in Startup Battlefield 200 at TechCrunch Disrupt 2026 in San Francisco, where a winner will be decided by VC judges on Thursday, October 15.