Compliance monitoring · Production
Stripe Radar uses machine learning to reduce fraud-driven chargebacks by 11%
The problem
Online fraud causes unnecessary losses for most online businesses—threatening viability for some—and as the online economy grows, fraud is becoming more sophisticated.
Workflow diagram · grounded in source
1
Real-time signal collection
integration
“fraud signals detected by Stripe.js and our mobile SDKs. These signals are most commonly used to identify "bots"—scripted web browsers that make fraudulent transactions. Signals we pay attention to include things like screen resolution a…”
2
ML fraud analysis
ai_action
“Radar uses advanced machine learning (optionally augmented by human-supplied rules) to shield Stripe users from hundreds of millions of dollars of fraudulent transactions every month”
3
Chargeback reduction output
output
“they reduce the expected number of fraud-driven chargebacks a business will receive by 11%”
Reported outcome
Radar's real-time fraud signals reduce the expected number of fraud-driven chargebacks a business will receive by 11%, shielding Stripe users from hundreds of millions of dollars of fraudulent transactions every month.
Reported metrics
Fraud-driven chargebacks reduction11%
Fraudulent transactions blocked monthlyhundreds of millions of dollars
Reported stack
RadarStripe.jsmobile SDKs
Frequently asked questions
What did this team achieve with this AI workflow?
Radar's real-time fraud signals reduce the expected number of fraud-driven chargebacks a business will receive by 11%, shielding Stripe users from hundreds of millions of dollars of fraudulent transactions every month.
What tools did this team use?
Radar, Stripe.js, mobile SDKs.
What results were reported?
Fraud-driven chargebacks reduction: 11%; Fraudulent transactions blocked monthly: hundreds of millions of dollars (source-reported, not independently verified).
How is this compliance monitoring AI workflow structured?
Real-time signal collection → ML fraud analysis → Chargeback reduction output.