Agent Fico Introduces Credit Scoring System to Prevent AI Transaction Fraud

A new tool called Agent Fico has been developed to address the high rate of false positive fraud detections for AI shopping agents. Currently, platforms like Stripe flag approximately 18% of first-time agent purchases as fraudulent because their buying behavior resembles bots. Agent Fico functions as a credit score for AI agents, analyzing 127 distinct transaction patterns including payment velocity, refund rates, merchant diversity, and session consistency. The score updates in real-time and allows merchants to underwrite transactions based on risk levels; scores above 680 are approved automatically, while those below 580 require human confirmation. Integrated with MnemoPay, the system uses a tamper-evident chain to create an audit trail, enabling agents to carry their reputation across different sessions and platforms. Early pilots with three ecommerce platforms indicate that agents with scores above 700 experience 12% fewer transaction declines. This innovation aims to streamline ecommerce operations for AI agents by providing provable history for disputes and eligibility for better payment terms.