Monitoring Layer
Alert-AI
Coming soonIntelligent alert triage and consolidation. Aggregates alerts from AML-AI, Fraud-AI and TxnMonitor-AI into a single prioritised queue with auto-deduplication, severity scoring, and MLRO auto-assignment. Reduces alert fatigue by suppressing low-confidence signals and grouping related alerts into cases.
This plugin is in development. The API schema below is planned and subject to change. You can design your integration now using this reference.
Endpoint
POST
/api/alerts/triageRequest parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
jurisdiction | string | Yes | One of: CA, AE, SA, GB. |
source_plugins | array | No | Plugins to pull alerts from: aml, fraud, txnmonitor (default: all). |
min_severity | string | No | Minimum severity to include: low, medium, high, critical (default: medium). |
auto_assign | boolean | No | Whether to auto-assign alerts to available MLRO (default: true). |
suppress_duplicates | boolean | No | Whether to deduplicate related alerts into groups (default: true). |
lookback_hours | number | No | Hours of alerts to triage (default: 24). |
Response fields
| Parameter | Type | Required | Description |
|---|---|---|---|
data.total_alerts | integer | Yes | Total alerts received before deduplication. |
data.triaged_alerts | array | Yes | Deduplicated, prioritised alert list with group_id, severity, source, and assigned_to. |
data.suppressed_count | integer | Yes | Number of alerts suppressed as duplicates or below threshold. |
data.critical_count | integer | Yes | Number of critical-severity alerts requiring immediate action. |
data.auto_assignments | array | Yes | List of alerts auto-assigned to MLRO staff with alert_id and assignee. |
data.ai_recommendation | string | Yes | Priority action summary for the compliance team. |
cURL
# Coming soon
curl -X POST \
https://shield.kodexa.systems/api/alerts/triage \
-H "Content-Type: application/json" \
-d '{...}'Code examples available when plugin launches