Mastercard spent decades training its fraud system to see bots as thieves. Now bots are the ones doing the buying.

Mastercard says its fraud models, built over decades to block bot-driven transactions, must adapt as AI agents begin making purchases on consumers’ behalf. The network scores 175 billion transactions a year in under 100 milliseconds, and says generative AI is helping detect far more high-risk fraud without adding consumer friction.

The rise of agentic commerce will force payment networks to distinguish malicious automation from authorized AI buyers, a much harder trust problem than traditional fraud scoring. Mastercard’s shift suggests risk frameworks across payments, identity, and merchant systems will need to evolve quickly as bots become normal participants in checkout flows.