Simulacra is a generative causal AI platform that fits a causal model to existing consumer‑research data, allowing users to intervene on any variable and instantly generate synthetic respondents that reflect how all downstream metrics will re‑equilibrate.
Funding
Funding not disclosed
Founders
Product
Problem
Consumer‑research teams must repeatedly field costly surveys to test how changes in price, audience mix, or product attributes will affect market outcomes, yet traditional methods cannot predict downstream effects without new fieldwork.
Solution
Simulacra fits a generative causal model to an existing consumer‑research dataset, learning the full response structure of the surveyed population. Users can then intervene on any variable (e.g., price) and instantly generate synthetic respondents that reflect how every dependent metric—volume, share, sentiment, segment composition—would re‑equilibrate. The platform also supports conditional generation to boost low‑incidence cohorts and scenario modeling that returns the most likely upstream conditions for a desired target outcome. All synthetic data preserve the original survey schema, respect learned constraints, and are validated against held‑out data to ensure fidelity. The engine is delivered via an interactive Synthetic Data Studio or a headless API, providing real‑time analytics, automated visualizations, and exportable results for insights teams and agency partners.
Target Audience
Primary customers are consumer‑insights teams, marketing agencies, and data‑science partners that need to extract actionable scenarios from surveys they have already fielded.
Features
- Generative causal AI that learns and reproduces the full joint distribution of survey variables, enabling interventional “do(X)” queries
- Conditional generation to create statistically indistinguishable synthetic rows for thin or hard‑to‑reach segments
- Scenario modeling that identifies the most probable combination of upstream factors to achieve a target outcome
- Real‑time cohort boosting, causal intervention, and scenario diagnostics within the Synthetic Data Studio UI
- Headless API for programmatic integration into agency workflows or data‑science pipelines
- Single‑tenant, in‑memory architecture with SOC 2 Type II and ISO 27001 compliance; no cross‑tenant training or data sharing
- Built‑in validation suite reporting marginal and multivariate MAE, variance preservation, and infeasibility diagnostics