Ordellum AnalyticsLaunching
Risk scoring
Every transaction and actor receives one score from 0 to 100, built from five dimensions and tied to a clear response.
- TransactionAmount, velocity, number of counterparties and gas anomalies
- 30%
- CounterpartySanctions, blacklists, mixer or privacy exposure, scams and fraud
- 25%
- BehaviouralStructuring, fast deposit-withdraw cycles, cross-chain hops and deviation from history
- 20%
- GeographicHigh-risk jurisdictions, FATF grey and black lists and chain jurisdiction
- 15%
- ProfileVerification level, account age, alert history and verified identity
- 10%
Move the slider to see the level, the automatic response and the human review window for any score.
51–70 · High
- Automatic response
- AI contextual review and a compliance alert
- Human review
- Within 48 hours
Analysis effort grows with risk. Scores below 51 without an explicit escalation signal are settled by the rule engine alone — no AI model is called, which keeps costs low and avoids unnecessary data exposure. Between 51 and 70, a contextual AI review is added. From 71, three independent models review the case and a suspicious-activity draft is prepared.
Some signals override the score: a sanctions or mixer match always receives the full review, regardless of the number.
Points added when funds move across chains
- Each bridge hop
- Added to the score: +10
- Moving to a privacy-focused chain
- Added to the score: +30
- Moving funds within 10 minutes of a bridge
- Added to the score: +15
- Touching five or more chains within 24 hours
- Added to the score: +25
- Using an unknown bridge
- Added to the score: +20
- The same amount appearing on a different chain
- Added to the score: +15
- Contribution
- How much each factor contributed to the final score, as a percentage.
- Decision chain
- The step-by-step reasoning that led from the evidence to the level.
- Counterfactual
- What would have changed the outcome — for example, an amount below the threshold would have meant medium risk.
- Plain-language summary
- A human-readable explanation of why a set of transactions is suspicious, suitable for audit.