Stripe Radar Expands Global Coverage, Adds AI‑Powered Fraud Models and Smart Dispute Tools
Enable Stripe Radar custom fraud models and multiprocessor signals to block high‑risk transactions across all payment methods and reduce fraud by up to 71%.
Enable Stripe Radar custom fraud models and multiprocessor signals for all payment methods, and integrate Smart Disputes AI recommendations to improve dispute win rates.
Summary
Stripe announced a major expansion of Radar at Stripe Sessions, extending protection to all supported payment methods—including bank debits, BNPL, crypto, digital wallets, real‑time payments, and cash vouchers—and adding new fraud types such as multi‑account abuse, pay‑as‑you‑go abuse, and bot‑driven payments.
Radar now blocks high‑risk transactions across the entire payment stack, and its new multiprocessor signals provide early fraud warnings and dispute predictions for off‑Stripe transactions. Custom fraud models allow businesses to feed structured metadata—product catalog, loyalty status, behavioral metrics—to Stripe’s global network, detecting at least 15% more fraud without increasing false positives.
In a five‑month period, Radar reduced suspected fraud by 71% for merchants using third‑party processors like Affim, Cash App, Klarna, and PayPal. The platform also introduced a bot score for Stripe Checkout, enabling anti‑bot policies, and new signals for fraudulent websites, merchants, and delinquency risk.
Stripe’s Smart Disputes AI now recommends evidence fields such as tracking numbers or usage logs, and an evidence library auto‑selects documents based on dispute reason codes, boosting dispute win rates by 3×.
The company released a public roadmap with hundreds of features through Q1 2027 and highlighted plans to add signals across the payments stack.
These enhancements give merchants tighter control over fraud and disputes, leveraging AI to reduce losses and manual effort.
Key changes
- Global coverage now includes bank debits, BNPL, crypto, digital wallets, real‑time payments, and cash vouchers
- 71% reduction in suspected fraud over five months for merchants using third‑party processors
- Custom fraud models detect 15% more fraud with no increase in false positives
- New multiprocessor signals provide early fraud warnings and dispute predictions
- Multi‑account abuse detection uses device fingerprints, IPs, and email domains
- Pay‑as‑you‑go abuse prediction allows intervention before billing
- Bot score for Stripe Checkout enables anti‑bot policies
- Smart Disputes AI recommendations increase dispute win rates 3×