Deployed at Tier 1 financial institutions since 2018.
Purpose-built for sanctions and financial crime compliance — not generic AI adapted to regulated environments.
Alert Interception
Payment alerts generated by the institution's screening engine are intercepted before reaching investigators, enabling earlier, automated assessment in real time.
Contextual Investigation
Each alert is analysed in full context, analysing unstructured payment message fields, resolving entities across all payment parties (originator, beneficiary, intermediaries, correspondents), and assessing jurisdictional exposure and counterparty relationships to determine whether genuine sanctions or restricted party risk exists.
Policy-Aligned Decision Output
Every alert results in a clear outcome — false positive clearance, escalation, or further investigation — supported by structured reasoning that demonstrates alignment with institutional screening policies and is fully auditable.
Up to
Alerts Resolved Without Manual Review
High volumes of payment screening alerts are investigated and resolved without manual intervention, removing the triage bottleneck and reducing reliance on investigator review.
Up to
Straight-Through Processing
Alerts are investigated and resolved by Iris 7, enabling payment decisions to be made without manual intervention and payments to move without unnecessary delay.
Automated Real-Time Decisioning
Payment screening alerts are investigated and resolved using structured, policy-aligned reasoning at transaction speed. Routine false positives are cleared automatically, while higher-risk matches are escalated for investigator review.
Payment Message Parsing
Unstructured payment message fields across all major formats — SWIFT MT and ISO 20022 (MX), ACH, SEPA, Fedwire — are parsed and relevant entities resolved automatically, including originators, beneficiaries, intermediaries, and correspondent banks.
Policy Editor and Simulator
Screening rules and decision logic can be configured to reflect institutional risk appetite. Policy changes are tested and validated in a controlled simulation environment before deployment.
Transparency and Explainability
Every decision includes structured reasoning, detailing the signals assessed and policy criteria applied — enabling full review and validation by investigators and regulators.
Continuous Quality Assurance
Selected alerts are sampled and reviewed by investigators to compare outcomes with agent decisions. Discrepancies are analysed to maintain alignment with institutional standards over time.
AI-Driven Continuous Improvement
Analyst feedback and decision outcomes are analysed to identify improvement opportunities, enabling ongoing optimisation of policies and investigative workflows.
Human-in-the-loop governance
Humans govern the policy framework that drives every decision. In production, human-over-the-loop oversight ensures the system operates within defined authority limits at all times.
