Map Fraud Exposure Profiles
“For each account, which fraud typologies will they be exposed to next month?”
Book a demo and get a free trial of the full platform: data science 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 each account, which fraud typologies will they be exposed to next month?
Fraud rules treat each fraud type independently — separate models for ATO, card fraud, check fraud, wire fraud. But accounts are often exposed to multiple fraud types simultaneously. A multi-dimensional fraud profile per account lets you deploy layered defenses and prioritize the accounts facing compound threats.
How KumoRFM solves this
Graph-powered fraud intelligence
Kumo’s multi-label prediction returns a different-sized set of fraud typologies per account. It discovers that A002 is exposed to phishing AND card skimming AND check fraud — three countermeasures needed simultaneously. Rules treating each separately would miss the compound risk.
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
Accounts
| account_id | account_type | channel_mix | risk_tier |
|---|---|---|---|
| A001 | Retail | online+branch | medium |
| A002 | Commercial | online+wire | high |
| A003 | Retail | mobile | low |
Fraud Events
| event_id | account_id | fraud_type | amount | timestamp |
|---|---|---|---|---|
| FE01 | A001 | phishing | 3,200 | 2025-01-10 |
| FE02 | A002 | card_skimming | 1,800 | 2025-01-12 |
| FE03 | A002 | check_fraud | 8,500 | 2025-01-14 |
Write your PQL query
Describe what to predict in 2-3 lines — Kumo handles the rest
PREDICT LIST_DISTINCT(FRAUD_EVENTS.FRAUD_TYPE, 0, 30, days) FOR EACH ACCOUNTS.ACCOUNT_ID
Prediction output
Every entity gets a score, updated continuously
| ACCOUNT_ID | CLASS | SCORE | TIMESTAMP |
|---|---|---|---|
| A002 | phishing | 0.85 | 2025-02-01 |
| A002 | card_skimming | 0.72 | 2025-02-01 |
| A002 | check_fraud | 0.61 | 2025-02-01 |
| A001 | phishing | 0.78 | 2025-02-01 |
Understand why
Every prediction includes feature attributions — no black boxes
Account A002 (Offshore Ltd)
Predicted: 3 fraud typologies: phishing, card_skimming, check_fraud
Top contributing features
Fraud events (90d count)
5 events
36% attribution
Channel mix
online+wire
24% attribution
Account risk tier
high
20% attribution
Distinct fraud types (historical)
2 types
13% attribution
Account type
Commercial
7% 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: Multi-fraud profiles reveal compound risk. A002 faces phishing AND card skimming AND check fraud — deploy all three countermeasures. 30–50% better fraud prevention through layered defense.
Related scenarios
Explore more fraud predictions
Topics covered
One Platform. One Model. Predict Instantly.
KumoRFM
Relational Foundation Model
Turn structured relational data into predictions in seconds. KumoRFM delivers zero-shot predictions that rival months of traditional data science. No training, feature engineering, or infrastructure required. Just connect your data and start predicting.
For critical use cases, fine-tune KumoRFM on your data using the Kumo platform and Data Science Agent for 30%+ higher accuracy than traditional models.
Book a demo and get a free trial of the full platform: data science agent, fine-tune capabilities, and forward-deployed engineer support.
