Combine supervised and unsupervised machine learning, billions of training transactions, and rules you control. Decline real fraud, approve real customers, automate the boring middle. Real-time scoring under 600ms, on every CX Pay charge. Backed by AML screening, KYB and KYC onboarding, and gateway-level velocity controls.
Fraud scoring backed by billions of training transactions, real-world fraud patterns, and machine learning models that refresh continuously, not quarterly.
Supervised models handle the patterns we already know. Unsupervised models surface the ones we do not. You set the policies. The engine does the rest.
Combines supervised and unsupervised models, trained continuously on regional fraud patterns specific to the Caribbean and Latin America.
Recognises identity-usage behavior across sessions and merchants. Increases accept rates on real customers, lowers review rates on borderline ones.
Cap the velocity. Block specific BINs. Require 3DS over certain amounts. Allow trusted recurring customers. Rules and ML run in parallel.
"What-if" replay against your historical transaction data. See exactly what a new rule would have done last quarter before you turn it on.
The engine continuously surfaces new patterns from your transactions and suggests rules. Review, accept, ignore, or modify.
Approve, send to manual review, or decline. Configurable thresholds, configurable handlers, dashboard queue for review.
Frictionless authentication on every transaction. SCA exemptions applied automatically where eligible. Higher approval rates, lower friction.
Filter by rule, by BIN, by region, by device. Review every flagged transaction. Approve, decline, or refund directly from the queue.
Enable on any CX Pay account with no engineering work. Or hit the API for full control, custom routes, and webhook delivery on every decision.
Pre-authorization runs your rules and ML models on every charge attempt before the request hits the card network. That means lower acquirer fees on the fraud you stop, fewer holds on real customers, and higher approval rates across the board.
Activated on day one, no per-feature charge. Configurable from the dashboard or the API.
Automatically assesses risk with advanced modeling tuned on regional patterns. No setup, immediate scoring on every transaction.
Flexible strategies based on your business needs. SMB defaults, custom rules for platforms, automated suggestions to refine.
Tools to quickly review suspicious transactions, approve overrides, see daily summaries, and export decisions.
For high-volume or complex businesses, our analysts step in. Design strategy, run reviews, handle peak load.
Complement your in-house team with our analysts. Help you design, implement, and maintain a fraud strategy customized for your business. Decades of regional experience.
Book a consult →Get expert help with acceptance rates, fulfillment, and fraud mitigation. Fully outsource your review process, or augment your team with overflow, after-hours, and peak-season support 24/7.
Get pricing →Fraud tools activate on day one. Talk to us about your current chargeback ratio, approval rates, and what good looks like for your business.