Behavioural economics · Complex systems · Decision science
Rubicona is a research-led advisory in behavioural economics and decision science — for the high-stakes calls you make once, under deep uncertainty.
Founded by two senior researchers in behavioural economics and decision science.
Our thesis
Every consequential choice is an act of valuation under uncertainty.
Pricing a risk, allocating capital, setting policy, designing a product, deploying AI — each is a decision made with incomplete information and predictable human bias. The structure of a choice shapes its outcome as much as the facts do.
We bring together the behavioural science of how people judge value and the computational machinery to measure it and act on it — diagnosing where decisions go wrong and redesigning them so the better outcome becomes the natural one.
Behavioural testing of disclosures, journeys and products to evidence good consumer outcomes and sharpen risk decisions.
Explore our services →Behaviourally-informed policy and validated instruments that measure attitudes, trust and disorientation under stress.
Explore our services →Dynamic discounting, learning-over-optimisation and decision models for choices made with genuinely incomplete knowledge.
Explore our services →LLM evaluation, human–AI workflow design and stress-testing of where automated judgment helps or harms.
Explore our services →The hard part of a high-stakes decision isn't the analysis — it's knowing when you've learned enough to act.
We make that line explicit, weighing the value of more information against the cost of waiting, so the call to cross or wait is made on purpose. It is the difference between deciding well and deciding late.
Illustrative. Past a point, more information barely improves the decision — but always delays it.
Co-founder
Professor of behavioural economics; senior cognitive scientist (ENS / Institut Jean Nicod). Neuroeconomics, decision theory, the mind of money.
Co-founder
Computational decision scientist (PhD; ENS). Networks, machine learning, reinforcement learning, and validated measurement.
More information can make decisions worse, not better.
Read →Irreversible choices carry an option value most models ignore.
Read →Redesign the choice, not the chooser.
Read →