Categorize Unknown Fraud Alerts
“For fraud alerts with missing fraud_type, what category do they most likely belong to?”
Book a demo and get a free trial of the full platform: research agent, fine-tune capabilities, and forward-deployed engineer support.
By submitting, you accept the Terms and Privacy Policy.
Loved by data scientists, ML engineers & CXOs at
A real-world example
For fraud alerts with missing fraud_type, what category do they most likely belong to?
50K alerts monthly, 15% missing fraud_type classification due to rule-engine gaps. Without proper categorization, alerts route to the wrong team, resolution takes 3x longer, and SAR narratives are incomplete. Proper categorization from transaction patterns saves 15,000 analyst hours/year.
How KumoRFM solves this
Graph-powered fraud intelligence
Kumo learns the relationship between transaction channel, amount patterns, account behavior, and fraud category. It sees that Alert FA02 (POS, $340, retail account) matches the pattern of card_not_present fraud, while FA03 (wire, $8,500, commercial account) matches first_party fraud.
From data to predictions
See the full pipeline in action
Connect your tables, write a PQL query, and get predictions with built-in explainability — all in minutes, not months.
Your data
The relational tables Kumo learns from
Fraud Alerts
| alert_id | account_id | fraud_type | amount | channel | timestamp |
|---|---|---|---|---|---|
| FA01 | A001 | ATO | 12,000 | online | 2025-01-05 |
| FA02 | A002 | ??? | 340 | POS | 2025-01-10 |
| FA03 | A003 | ??? | 8,500 | wire | 2025-01-12 |
Accounts
| account_id | account_type | risk_tier | open_date |
|---|---|---|---|
| A001 | Retail | high | 2022-03-15 |
| A002 | Retail | medium | 2023-06-01 |
| A003 | Commercial | high | 2021-11-08 |
Write your PQL query
Describe what to predict in 2-3 lines — Kumo handles the rest
PREDICT FRAUD_ALERTS.FRAUD_TYPE FOR EACH FRAUD_ALERTS.ALERT_ID
Prediction output
Every entity gets a score, updated continuously
| ALERT_ID | TARGET_PRED |
|---|---|
| FA02 | card_not_present |
| FA03 | first_party |
Understand why
Every prediction includes feature attributions — no black boxes
Alert FA02
Predicted: card_not_present
Top contributing features
Alert channel
POS
34% attribution
Alert amount
$340
26% attribution
Account type
Retail
19% attribution
Account risk tier
medium
13% attribution
Account open date recency
1.8 years
8% attribution
Feature attributions are computed automatically for every prediction. No separate tooling required. Learn more about Kumo explainability
PQL Documentation
Learn the Predictive Query Language — SQL-like syntax for defining any prediction task in 2-3 lines.
Python SDK
Integrate Kumo predictions into your pipelines. Train, evaluate, and deploy models programmatically.
Explainability Docs
Understand feature attributions, model evaluation metrics, and how to build trust with stakeholders.
Bottom line: 7,500 alerts auto-categorized per month. Correct routing saves 2 hours per alert = 15,000 analyst hours/year. Faster resolution, complete SAR narratives, better regulatory standing.
Related scenarios
Explore more fraud predictions
Topics covered
