MoreThanTwins builds causal digital replicas of your business to simulate decisions, quantify uncertainty, and show you what happens — before it happens.
See how it worksWhen a decision maker faces a critical choice — pricing, market entry, hiring, investment — they use one of three tools. Each has a breaking point.
Works: Fast, familiar, zero cost
Breaks: Breaks when the problem has 10+ variables with feedback loops between them
Works: Deep analysis, senior expertise
Breaks: 50-200K per project. Weeks of analysis. A report that expires the day conditions change
Works: Accessible, fast, conversational
Breaks: No model of YOUR business. Hallucinates financial figures. Doesn't conserve quantities
No tool on the market simulates consequences with rigor, quantifies uncertainty, and learns from its results. MTT does.
Vending-Bench 2: managing a vending machine for 365 days, maximizing profit under real constraints. Twin vs human expert.
“The twin outperformed the human expert by 104%. That's what we know. Whether it scales to other domains — we're going to find out, not assume.”
Validated engine, solid narrative, zero euros of revenue. We show what works and what we don't know yet.
Each brain does what the others can't. Together, they deliver what no single approach can.
System dynamics + Monte Carlo simulation. "What happens if you raise prices 20%?" — with probability distributions, not a single number.
Evaluates whether it's worth analyzing further or if the answer is already clear. Identifies what information is missing and where to focus.
9 specialized virtual advisors — CFO, COO, CLO — review every recommendation with domain checklists. They don't calculate: they contribute judgment.
A simulator that only calculates is a sophisticated Excel.
An LLM that only reasons hallucinates numbers.
A committee that only opines is corporate theater.
MTT connects all three.
Research, validation, and lessons learned — building the operating system of the autonomous enterprise.
LLMs describe a hedge correctly but compute the parameters wrong — a 54-point gap between conceptual reasoning and numerical accuracy. This is why MTT separates calculating from reasoning.
The jump from $709 to $1,361 (+92%) came from correcting a single parameter: variety from 6 to 4 products. Configuration matters more than model sophistication.
Today we sell 'make mistakes for free.' Autonomy is earned, not sold. Each level validates the next — and we're honest about where we are.