Most models of complex systems begin with a simplification: assume rational actors, assume stable conditions, assume the variables you can't measure don't matter. Mana begins somewhere different. Its simulation framework is designed to incorporate the full texture of human behaviour — the inconsistencies, the social dynamics, the feedback loops that emerge when millions of individual decisions interact at scale.
The systems we are most interested in — healthcare pathways, urban infrastructure, financial markets, public policy — share a common characteristic. They resist simple optimisation because the people within them adapt. Change one variable and the humans in the system respond, often in ways that partially or entirely offset the intended effect. Traditional simulation tools struggle to capture this. Mana is built specifically to model it.
At the core of the simulation framework is a behavioural layer — a representation of how individuals within a system are likely to act given their context, history, and the actions of those around them. This layer is informed by Hxly's Library: a growing collection of behavioural datasets and persona models that ground the simulation in empirically observed human behaviour rather than theoretical assumptions.
The question is not whether a system can be optimised in theory. It is whether it can be optimised in the presence of the people who inhabit it.
Simulation runs can be configured across a wide range of parameters — population size, time horizon, intervention type, environmental conditions — and are designed to be iterated rapidly. A policy designer might run hundreds of scenarios in the time it would previously have taken to model one. Each scenario surfaces not just aggregate outcomes but the distribution of effects across different groups within the population, making it possible to identify interventions that are both effective on average and equitable in their distribution.
The output of a Mana simulation is not a single predicted outcome but a probability landscape: a map of what is likely, what is possible, and what is sensitive to small changes in initial conditions. This approach acknowledges the fundamental uncertainty of complex systems while still providing decision-makers with actionable, structured insight.
Current simulation deployments span three domains: public health system modelling, urban mobility and infrastructure planning, and financial contagion analysis. In each case the goal is the same — to give decision-makers a more honest picture of how interventions will land in a world that pushes back.