How it works
A large-scale computational system that models global processes as interacting systems rather than isolated variables.
Why systems
Most forecasting isolates a price and extrapolates it. That works until something moves two steps away. A gas disruption is a gas story for a week; then it becomes a fertiliser story, because nitrogen fertiliser is made from gas, and a few months later it is a grain story, because farmers planted less. Nothing in the fertiliser price series announced that. The cause lived in a different system.
The Universal Synaptic Engine does the opposite: it integrates markets, trade, policy, logistics, climate and behaviour into a single structure. When conditions change anywhere, the engine recalculates their impact across the whole — producing actionable signals, not isolated data points.
What the twin integrates
Each domain is modelled on its own terms, then wired to the others so that a change in one is felt where it lands, not where it started.
markets
trade
policy
logistics
climate
behaviour
We describe the engine by what it connects, not by how much it ingests. Coverage is published per product, where it can be checked.
How models compete
Many models run in parallel on the same question, continuously recalibrating against what actually happened.
The engine does not pick a winner and stop. It keeps the losers running, keeps scoring them against what actually happened, and promotes on evidence. Forecasts are never black boxes.
In Market Lens this runs as champion and challenger: for every commodity a champion publishes the forecast while challengers keep scoring against it, and a challenger that outperforms on observed results takes over. We name the mechanism, not the models — which model wins is a result, not a promise.
From engine to application
Market Lens is the engine's commodity layer and its first commercial application: 55+ commodities across 7 categories, forecast with confidence bands, scored in public.
The War Room is the same engine pointed at a shock. Stage a disruption and it propagates through the measured relationships between commodities — two hops out, so the second-order move is on the table before it is in the price.
Beyond commodities — logistics, infrastructure, policy, regulation — is Enterprise 360: the whole engine, pointed at your question.
We do not publish a single accuracy number here. We publish a per-commodity scorecard on rolling backtests and let you inspect it.
Research lineage
The Universal Synaptic Engine grew out of the Digital Twin programme of the Global Arena Research Institute, the Prague non-profit of which GARI Enterprise s.r.o. is the commercial spin-off. The Institute has conducted continuous R&D since 2017 with governments, public institutions and international partners; the engine is that work carried into a product.
Institute research behind the engine has been funded by TAČR (including the AID TWIGLOW project), the International Visegrad Fund, the Czech-Polish Forum and Konrad Adenauer Stiftung. The funders supported the Institute's research; they do not endorse any commercial product built on it.
The family
Commodity price forecasts, risk and scenario analysis for procurement and trading teams.
Demand-side intelligence — reading consumption, inventories and substitution before they move prices.
Early warning on supplier, shipping-lane and policy disruptions across your bill of materials.
Board-level scenario planning that turns the engine's signals into strategic options.
Turn global chaos into buying confidence.
Published pricing — no sales call to see it. Self-serve from €200/month; Procurement module and enterprise agreements on request.
Tell us the decision you are facing and the horizon you plan against. You will hear back from a person, not a pipeline.