Each analysis passes a session through QUORUM's full nine-stage decision pipeline — the same logic, signal weights, and arbitration pathways used in the production system. No data is transmitted. All computation runs locally.
Select a Signal Test or Fraud Typology from the left panel. Each scenario represents a distinct session profile with pre-configured behavioral and network signals.
Press Analyze to run it through the pipeline. Results appear in real time as each stage completes — from signal intake through scoring, arbitration, and final verdict.
Campaigns run a sequence of related sessions to simulate multi-event fraud patterns — velocity escalation, synthetic identity rings, or governed rule rollback.
The narrative strip above the pipeline explains what is happening at each step and why. Campaigns pause between events so you can observe state changes.
Live Flow — real-time pipeline execution, tier evaluations, and verdict with risk score.
Forensic Replay — navigable step-through of the full causal decision chain after a run completes. Each stage shows cause, effect, and the internal state at that moment.
System Topology — static reference diagram of the three-tier AI architecture.
Nine sequential stages: Intake -> Normalize -> Engine -> Score -> Arbitrate -> Dispatch -> Evaluate -> Consensus -> Commit.
Each stage is causally dependent on the previous. A stage that completes in green passed. A stage that completes in red triggered a block condition. The gauge reflects cumulative risk across all three AI tiers.