Modeling Human Actions
Startups that promise to predict how humans will behave are having a moment. Over the past year, Simile raised $200 million at a $2 billion valuation; Aaru raised $88 million at a $1 billion valuation; and Humans&, an AI startup that announced a massive $480 million seed round in January at a $4.48 billion valuation, launched Persimmon to model human behavior.
While the current standard for predicting human behavior relies heavily on large language models prompted or fine-tuned to role-play as specific demographics, San Francisco-based Mirror Particle argues that this approach is fundamentally limited.
“It’s like bringing a super soaker to Niagara Falls,” says Abhivyakti Ahuja, co-founder and CEO of Mirror Particle, which provides brands with an AI engine designed to predict consumer behavior and underlying motivations. LLMs trained on massive text corpora cannot be sufficiently altered through small-scale fine-tuning to reflect dynamic human choices.
Ahuja notes that LLMs model written language rather than the multi-dimensional nature of human experience, which encompasses visual perception, spatial reasoning, and social intelligence.
A Longitudinal Approach
Mirror Particle is building a foundation model from scratch to simulate why humans act the way they do and how their behavior evolves 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.”
Mirror Particle has raised an angel round and is preparing to close its first venture round.
Data Sources and Market Strategy
The startup utilizes a proprietary combination of client customer data, current events, pop culture, and social media to model demographic segments as evolving systems. The methodology emphasizes revealed behavior—what people actually do rather than self-reported survey data.
Mirror Particle’s initial commercial strategy targets market research and brand strategy budgets. The system can assist companies in determining whether a target demographic actually demands a specific product, rather than merely optimizing marketing copy.
“What if the target demographic doesn’t want eyeshadow palettes?” Ahuja said. “Maybe blush is a better option to go for if you want to sell a product to this market.”
The prediction engine aims to supply the contextual motivations and constraints behind consumer actions, informing brand decisions.
In an early pilot involving a pet food brand, Mirror’s technology indicated that packaging imagery was secondary to consumer perception issues regarding brand ubiquity and pricing tier.
Founders and Long-Term Vision
“The way we see our model evolving is like how a baby learns about the world,” Ahuja said, noting a progression from vision to language, body awareness, and social intelligence.
Ahuja studied neuroscience and computer science at the University of Toronto, influenced by neural network pioneer Geoffrey Hinton.
She subsequently worked at Amazon Robotics, where she met co-founders Will Song, who specialized in sales personalization engines, and Thomson Yen, who focused on deep learning applications for AI agents.
The startup’s long-term objective is to serve as a general infrastructure layer for anticipating human behavior, scaling from broad population analysis down to individual insights.
“We just need a better model of humans if we’re going to work alongside AI and with each other,” Ahuja said.
Check out Mirror Particle and many other innovative startups that have been vetted by TechCrunch’s editorial team next week at Disrupt in downtown San Francisco. The winner of this year’s Startup Battlefield will be decided by our slate of VC judges on the afternoon of Thursday, October 15.
