AI merchandising copilot powered by causal customer simulation
Elevate your merchandising by leveraging a causal customer simulator trained on your live transaction data that accurately predicts lift, cannibalization, and margin impact in minutes.
Increase gross margins by 10-15% via prices, promotions, and assortments that capture live intent of customers
Mimetra represents a paradigm shift in how a retailer uses data to power merchandising. Rather than relying on stale, historical data, Mimetra uses real-time transaction data and market signals to simulate live customer intent and behavior.
The result is a living simulation of your customer base that estimates complex substitution, cannibalization, and churn patterns to power your merchandising decisions.
Explore the Mimetra Simference™ platform
Simulates: Price elasticity of incremental demand and revenue; incremental revenue gradient; cross-elasticity of cannibalized demand
Use cases: Rank products to prioritize price changes; verify directional lift in incremental revenue; estimate revenue loss from adjacent products
Simulates: Promotional lift in incremental demand and revenue at configurable discount levels for any promotional bundle; incremental promotional elasticity and cross elasticity
Use cases: Rank promotion bundles by uncannibalized revenue lift; make reprice vs. promote decisions
Simulates: Substitution and churn behavior upon adding or removing items from an assortment
Use cases: Rank items by incremental impact on revenue; determine the revenue optimal assortment under shelf constraints; identify redundant products; make SKU add/drop decisions
Mimetra addresses a critical gap in existing solutions: the lack of a causal model of customer response rooted in real data
ML-native merchandising solutions rely on historical data that is
Stale — Doesn’t reflect the needs and wants of the customer right now
Confounded — Reflects the correlation between what the firms chose to do and the demand, not causation
Limited in scope — Reflects only a tiny fraction of policies that firms have actually tried
and hence, is structurally incapable of accurately evaluating what would happen under new pricing, promo, and assortment strategies that have never been tried.
Modern AI-native copilots lean heavily on AI foundation models without rigorous and scientific grounding in real data, and hence, are unreliable for mission-critical decisions.
Extracting causality from data is one of the hardest problems in ML. Using AI to solve this problem is frontier science.
That is why we are the right people for the job. Founded by career academics with over three decades of research and industry experience in AI and causal inference, Mimetra is defining the frontier of AI-driven causal simulation science.
We don’t need big data
Works for Retailers of Any Size
Retail datasets vary dramatically. A global retailer may have billions of transactions, while smaller retailers may operate with fewer customers, SKUs, and purchase histories.
Mimetra’s customer simulation engine is designed to adapt across this spectrum. By combining foundation model adaptation with decades of consumer behavior science, the system learns meaningful demand patterns even when transaction volumes are modest.
Designed for Practical Deployment
Privacy-First, Cost-Controlled Simulation
Mimetra builds customer simulations directly from transaction behavior, without relying on sensitive demographic data. By learning from what customers actually buy, the platform preserves privacy while capturing the signals that matter most for retail decisions.
The platform also lets you control the cost of decision optimization. Choose from multiple simulation fidelities—from fast, low-cost simulations for quick exploration to high-precision simulations for high-stakes decisions.
Privacy by design. Simulation on your terms.
About us
Vijay Kamble — CEO
Vijay Kamble is an Associate Professor of Business Analytics at the University of Illinois Chicago whose research focuses on applying reinforcement learning and causal inference methods to optimize retail and e-commerce systems. He previously spent a year as a visiting scholar in Amazon’s Pricing and Promotions organization, where he redesigned promotion sourcing algorithms for large-scale sales events such as Prime Day. At Amazon, he saw firsthand how counterfactual evaluation based on observational data is a central challenge in pricing analytics. He co-founded Mimetra with Varun when their research showed that AI synthetic data can effectively address this challenge.
Education: PhD in Computer Science, University of California, Berkeley; Postdoc at Stanford.
Formerly at Amazon, Technicolor, and Livsyt.
Varun Gupta — CTO
Varun Gupta is a Professor of Operations and Information Systems at the University of Utah, and an expert in optimization and machine learning for large-scale e-commerce platforms and digital marketplaces. His research focuses on developing algorithms for complex decision systems operating at scale. At Mimetra, Varun leads the company’s scientific and technological efforts, advancing the frontier of causal customer simulation from transaction data.
Education: PhD in Computer Science, Carnegie Mellon University.
Formerly at Microsoft, Google, Bell Labs, University of Chicago Booth School of Business.