Decisions, rehearsed.

Flagship Demo

The Rumour Cascade

The scenario. A false rumour about a bank's solvency begins circulating among a subset of depositors. The Rumour Cascade simulates how the rumour propagates through the depositor population, which segments respond with withdrawal behaviour, and how the bank's communication responses either dampen or amplify the cascade.

What it demonstrates. Emergent bank-run dynamics that cannot be predicted from individual depositor behaviour. The global affective signalling layer produces collective anxiety responses that cross demographic and geographic boundaries. The Causation engine traces the causal chain from initial rumour to each withdrawal event.

Scale. 50,000+ synthetic depositors, realistic geographic and relationship topology, 6-month simulated window.

Applicable to. Retail banking, building societies, deposit-taking institutions, regulators modelling systemic stability.

Flagship Demo

Fraud Signatures

The scenario. A government tax authority needs to deploy new fraud detection rules against R&D tax credit claims without generating intolerable false-positive burden on legitimate claimants. Fraud Signatures simulates the deployment of candidate rule sets against a synthetic claimant population that includes both legitimate and fraudulent agents.

What it demonstrates. Adversarial adaptation. The synthetic fraudsters do not passively accept detection. They evolve their techniques in response to which rules trigger, producing realistic adaptation dynamics that static rule testing cannot capture. The Adaptation engine drives this evolution.

Scale. 10,000+ synthetic claimants, 500+ injected fraud signatures, 6-month simulated enforcement window.

Applicable to. HMRC, tax authorities internationally, financial crime detection, insurance fraud, compliance enforcement.

Flagship Demo

The Closure Study

The scenario. A FTSE-scale bank must close 40 of its 500 branches. The decision is not which 40 but in what sequence, over what timeline. The Closure Study simulates each candidate closure sequence against the bank's actual customer population topology, surfacing which sequences produce cascading deposit outflow and which achieve cost targets with manageable customer impact.

What it demonstrates. Nonlinear customer response to sequential events. A single closure may produce no effect; the same closure in a sequence of five produces a cascade. The global affective signalling layer captures the collective anxiety that builds when customers perceive a pattern of withdrawal from their region.

Scale. 50,000-500,000 synthetic customers, real geographic and demographic topology, 24-month simulated duration, 5-20 candidate sequences in parallel.

Applicable to. Retail banks, building societies, multi-site operators, any organisation sequencing location closures or consolidations.

Flagship Demo

The Trial Map

The scenario. A pharmaceutical company is designing a Phase III clinical trial. The Trial Map simulates the proposed protocol — inclusion criteria, endpoint definitions, site locations, recruitment strategy — against a synthetic patient population calibrated to real epidemiological distributions. Synthetic patients respond to recruitment, adhere or do not to protocol, and drop out for realistic reasons.

What it demonstrates. Behavioural dynamics in clinical trials that biostatistical modelling misses. Site-level variance, recruitment trajectory uncertainty, dropout patterns that correlate with protocol features rather than patient characteristics. The Causation engine explains why each outcome occurred.

Scale. 5,000-20,000 synthetic patients, realistic trial site topology, full trial duration simulated (18-36 months), multiple protocol variants in parallel.

Applicable to. Pharmaceutical companies, CROs, biotech, medical device companies, academic medical centres.

Coming soon

Policy Futures

The scenario. A government ministry is evaluating a policy change that will affect millions of citizens. Policy Futures simulates the proposed policy against synthetic citizen populations calibrated to real demographic distributions, projecting forward through months of simulated time to surface emergent behavioural responses, compliance dynamics, and political durability.

What it will demonstrate. Population-level behavioural response to policy that standard economic impact assessment cannot model. Trust shifts, compliance cascades, opposition dynamics, and the feedback loops between citizen behaviour and policy outcomes.

Status. In development. Available for early engagement with government and public sector organisations. Contact us to discuss scoping.