Compliance monitoring · Production

Stripe uses ML to reduce SCA friction by 20% and fraud by 8%

The problem

SCA regulations in the EEA and UK mandate two-factor authentication for card transactions, creating customer friction and conversion loss; navigating the 20+ considerations for authentication and SCA exemptions is too complex to optimize without ML.

Workflow diagram · grounded in source
1
New card transaction arrives
trigger
“When presented with a new transaction, the model pattern-matches against this historical data”
2
ML model evaluates hundreds of variables
ai_action
“it takes into account hundreds of variables ranging from authentication-specific outcomes (such as challenge success rate and the probability of frictionless authorization being granted) to charge-level fraud risk”
3
Authentication decision requested
output
“requests the specific authentication decisions that optimize across conversion, fraud risk, and cost”
4
Periodic model retraining
feedback_loop
“We retrain and deploy a new version of the model every few weeks, adding new features, and ensuring that we're optimizing for the latest changes in network and issuer behavior”
Reported outcome

The ML authentication engine reduced two-factor challenges shown to customers by 20%, decreased fraud by 8% on average for eligible card transactions, and improved the average authorization rate by 61 basis points.

Reported metrics
Two-factor challenges shown to customers20%
Fraud on eligible card transactions8%
Average authorization rate61 basis points
Reported stack
authentication engine
Source
https://stripe.com/blog/using-ml-to-comply-with-sca-requirements
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The ML authentication engine reduced two-factor challenges shown to customers by 20%, decreased fraud by 8% on average for eligible card transactions, and improved the average authorization rate by 61 basis points.

What tools did this team use?

authentication engine.

What results were reported?

Two-factor challenges shown to customers: 20%; Fraud on eligible card transactions: 8%; Average authorization rate: 61 basis points (source-reported, not independently verified).

How is this compliance monitoring AI workflow structured?

New card transaction arrives → ML model evaluates hundreds of variables → Authentication decision requested → Periodic model retraining.

Related compliance monitoring cases