Built by founders who spent their careers predicting what people perceive, prefer, and purchase.
Simulacra is led by Jason Cohen — previously the founder & CEO of Gastrograph AI / Analytical Flavor Systems (acquired by NielsenIQ in 2025), and co-founder Gerry Schweitzer, previously co-founder of Leblon Cachaça (acquired by Bacardi in 2015). After 15 years of building AI models for the world's largest CPGs, we built Simulacra because the same teams that ran our work asked, repeatedly, for a way to run scenarios on the surveys they'd already paid to field.
Four positions that shape our platform.
Causality matters.
Correlation isn't strategy. The questions CMI leaders care about are counterfactual: what changes if price moves, if the audience mix shifts, or if a claim changes? Simulacra is built to answer these questions using end-to-end causal AI, for real-time inference on the data you already have.
Response structure matters.
Predicting marginals is easy. The response structure — how every variable moves with every other — is hard. That's exactly what Simulacra learns, predicts, and validates.
Constraints matter.
All generations are valid by design. Simulacra refuses to invent questions outside your schema, and refuses to generate any scenario without prior predictive probability learned from your data. Efficacy is earned with accurate predictions; trust is earned with honest refusals.
Validation matters.
We validate claims against held-out data, not demos. Every benchmark includes the metrics, methods, and known gaps. Bring a study you already fielded, and we'll score Simulacra against the truth you kept back.
JC
Jason Cohen
Before Simulacra Data, I was the Founder and Executive Director of the Tea Institute at Penn State, an interdisciplinary research institute for the study and preservation of tea and tea culture; during that time, I did my research on human sensory perception and preference of flavor, aroma, and texture, leading to my first company Gastrograph AI (Analytical Flavor Systems). Gastrograph was founded in 2013 and acquired by NielsenIQ in 2025 — the technology behind all of NiQ Creative Product AI. During my 10 years as Founder & CEO of Gastrograph, I faced all of the industry's problems with consumer and market research. I founded Simulacra in 2023 to solve those problems: expensive fielding, small sample sizes, lack of coverage, and no foresight into the viability of the desired results.
Keynote speaker and published researcher in AI, sensory science, and consumer behavior. Selected interviews and profiles: The Atlantic, The Economist, The Guardian, The New Statesman, The Spoon, IFT Food Technology, Forward Fooding, the Center for Data Innovation, and the Data Gurus podcast.
In my free time, I write books about tea (teatechnique.org).
“We validate 100% of our claims while pushing the frontier of AI for consumer research. Let's validate our platform on your data.”
Gerard “Gerry” Schweitzer
Before Simulacra Data, I spent more than three decades in brand strategy. I started my career in strategic planning in Paris, later served as Strategic Planning Director at Ogilvy & Mather New York on American Express, and worked across spirits, CPG, financial services, and founder-led brands. Then I co-founded Leblon Cachaça, the category-defining U.S. cachaça brand acquired by Bacardi in 2015.
I first worked with Jason as a board advisor and consultant to Gastrograph AI, helping global food & beverage manufacturers use new technology to develop market-winning products. At Simulacra, I make causal AI usable for the teams that turn research into decisions: CMI, insights, marketing, brand, innovation, and commercial leadership. We help companies get more from the studies they already fielded — running the scenarios, cohort boosts, and target searches that fieldwork alone can't cover.
“The best research does not just describe consumers. It helps teams understand what will move them.”
GS
Bring a study you already fielded. The founders will walk you through the validation.
Three validations are public. Jason and Gerry will run the holdout protocol with you on a study you already paid to field.