AI-assisted resilience analysis
A backup exists. Will the critical workload actually recover in time?
Apertus structures messy recovery evidence. Deterministic rules identify the first binding or unresolved condition. The Copilot then shows the next justified action—without turning missing evidence into confidence.
Apertus livePublic-safe synthetic demo
Main synthetic caseNorthstar Bank
Can the same critical workload recover inside its tolerance?
95 minObserved recovery
>
60 minRecovery target
First binding stateS5 RED · B2Repair execution and retest.
1 Scenario2 Apertus3 Rules4 Decision5 Why / Fix6 Systemic lens
Step 1 · Evidence intake
Describe the recovery scenario
Use normal language. Apertus extracts explicitly stated facts only; missing evidence remains UNKNOWN.
Public-safe demo only. The model cannot invent missing evidence or set the final decision.
Evidence disciplineWhen the evidence stops, the inference stops.
Synthetic public-safe prototype. AI structures and challenges; deterministic rules decide. UNKNOWN remains UNKNOWN.