Credit research is the core of risk control—but it’s also the most detail-choking, time-sucking step in lending. Take E.Sun Bank (玉山银行): with enterprise lending exceeding NT$1T, their corporate finance team must repeatedly re-check each counterparty’s credit and financial status. Yet data is scattered, gathering is manual, and writing a credit report often takes 10+ hours—new hires can take 20+ hours, meaning days to deliver a first draft. To make it worse, each relationship manager (RM) writes to different standards, so risk depth and judgment vary, driving endless revisions and back-and-forth approvals. Since 2023, E.Sun Group has gone deep into workflow integration, using generative AI as a real deployment scenario. They split credit work into 6 stages and built tools directly embedded into daily systems: - iCatcher: crawls public web info + connects via API to summarize in ~5 minutes - AI + OCR: structures financial statements faster than manual entry - ZhijieTong: drafts a full credit report in ~7 minutes using the “5P” framework - “Prompt 家” acts like a shadow reviewer to flag risks early Result: reduce report prep from 1–2 days to 2–3 hours, while turning expert experience into an iterative quality loop with LLM evaluation + human verification. #GenerativeAI #FinTech #CreditRisk #WorkflowAutomation #RiskManagement
Want to learn more? Visit Explore the world, stay updated on travel insights and international affairs, and discover authentic stories from real life
评论
发表评论