Stripe Radar Expansion Tackles Multi-Account Abuse and AI Fraud Patterns.

Stripe announced a major expansion to Radar, its fraud prevention platform, introducing new tools to detect multi-account abuse, pay-as-you-go fraud, and bot-driven payments. The update, unveiled at Stripe's Sessions conference last month, represents the company's largest Radar expansion to date.

The numbers underscore the urgency. Across Stripe's network, over one in six sign-ups at AI companies are now linked to multi-account abuse, patterns that have historically been nearly impossible to detect automatically. The new Radar capabilities aim to address this and other emerging fraud vectors that have proliferated as AI-powered services scale.

The expanded Radar now offers global payment coverage with new multiprocessor signals, allowing businesses to apply fraud detection across payment providers beyond Stripe. Companies can also build custom fraud models tailored to their specific risk profiles, rather than relying solely on Stripe's default ruleset.

For platform businesses, Stripe added account fraud protection to help identify suspicious users before they transact. This includes detecting coordinated account creation, fake identities, and other abuse patterns that target platform sign-up flows. The company claims these features can catch fraudulent accounts earlier in the lifecycle, reducing downstream payment fraud.

The update also introduces defenses against usage-based billing abuse, where attackers exploit pay-as-you-go pricing models to rack up charges with stolen payment methods or synthetic identities. This has become a particular pain point for AI companies offering token-based APIs, including platforms like ElevenLabs, which recently expanded its billing capabilities on Stripe.

Stripe also revamped its dispute management tools with automated evidence libraries. When a chargeback occurs, Radar now pulls together relevant transaction data, customer communication logs, and delivery confirmations into a pre-formatted evidence package. The system uses machine learning to prioritize the most relevant evidence for each dispute type, aiming to increase win rates without manual intervention.

The update follows Stripe's recent focus on AI-native features, including a one-prompt payment setup for AI agents and integration with Amazon Bedrock AgentCore. As AI companies scale rapidly and attract more sophisticated fraud, Stripe appears to be positioning Radar as the default fraud infrastructure for this next wave of internet businesses.

The new Radar features are rolling out now to Stripe users. Pricing details were not disclosed in the announcement, though Stripe typically charges per transaction screened or as part of its broader platform fees.