Finance ops · Production

Stripe uses AI and issuer collaboration to create adaptive Radar fraud rules that increase payment success rates

The problem

Radar's default rules blocked transactions whenever CVC or postal code verification failed, causing legitimate low-risk transactions to be blocked alongside fraudulent ones and leaving revenue on the table.

Workflow diagram · grounded in source
1
Transaction fails CVC or postal code check
trigger
“Historically, Radar's default rules have allowed you to automatically block transactions whenever a card's CVC or postal code verification fails”
2
AI risk score generation
ai_action
“our AI models use data across our network of millions of businesses and tens of billions of transactions to develop a risk score that predicts whether a payment is likely to be fraudulent”
3
Route by initial risk score
routing
“Based on this score, we block the transaction, trigger a manual review, or send it to the issuer”
4
Issuer authorization with additional data
integration
“The issuer then either authorizes or declines the transaction, and sends us new, additional information with their decision—such as whether the CVC or postal code was incorrect”
5
Radar combines issuer signal with risk score
ai_action
“Radar combines the issuer's response and our original risk score. For example, depending on your other Radar rules, we would block higher risk transactions with an incorrect CVC, while allowing lower risk transactions with an incorrect CVC”
Reported outcome

Businesses migrating to the adaptive rules see a 1.3 percentage point increase in payment success rates with minimal changes to fraud rates, representing the potential for billions of dollars in additional collective revenue each year.

Reported metrics
Payment success rate increase1.3 percentage point increase
potential additional revenue across Stripe networkbillions of dollars in additional revenue each year
Reported stack
RadarRadar for Fraud TeamsEnhanced Issuer Network
Source
https://stripe.com/blog/using-ai-dynamic-radar-rules
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Businesses migrating to the adaptive rules see a 1.3 percentage point increase in payment success rates with minimal changes to fraud rates, representing the potential for billions of dollars in additional collective…

What tools did this team use?

Radar, Radar for Fraud Teams, Enhanced Issuer Network.

What results were reported?

Payment success rate increase: 1.3 percentage point increase; potential additional revenue across Stripe network: billions of dollars in additional revenue each year (source-reported, not independently verified).

How is this finance ops AI workflow structured?

Transaction fails CVC or postal code check → AI risk score generation → Route by initial risk score → Issuer authorization with additional data → Radar combines issuer signal with risk score.

Related finance ops cases