Ms. Lin Yuan

A Multi-Agent AI Framework for Intelligent Chargeback Recovery

Abstract:

Chargeback disputes are costly and time-consuming, often returning less than 15% of lost revenue after fees and reversals. This talk presents a multi-agent AI framework that automates chargeback resolution while maintaining transparency, fairness, and compliance. The system’s agents collaborate to classify cases, gather evidence, build arguments, and validate compliance in an auditable workflow. By separating reasoning from rule enforcement, it improves both trust and efficiency. The goal is to show how agentic AI can make financial risk management more sustainable and accountable.

Profile:

Lin Yuan has spent nearly two decades advancing the science of fraud prevention and risk analytics at leading fintech innovators such as PayPal and Block (Afterpay). Her work has shaped global risk frameworks, AI driven detection systems, and cross functional teams that safeguard billions in digital transactions each year. Now beginning a new chapter at OKX, she focuses on product and growth data science, applying her deep analytical expertise to drive user engagement, personalization, and sustainable platform expansion.