China Artificial Intelligence China AI Application Tour Takeaways: Early Evidence of Workflow-Led Monetization

研报英文原文证据摘录

China Artificial Intelligence China AI Application Tour Takeaways: Early Evidence of Workflow-Led Monetization

Asia Pacific Equity Research

22 May 2026

China Artificial Intelligence

China AI Application Tour Takeaways: Early Evidence

of Workflow-Led Monetization

Eight meetings across autonomy, independent model development, enterprise Head of China Equity Research

workflow software and vertical AI applications point to early evidence of and Co-Head of Asia TMT

workflow-led monetization in selected use cases. The most recurring theme is Research

that value creation appears increasingly tied to proprietary data, workflow Alex Yao AC

context, customer integration and deployment know-how. Foundation models (86 21) 6106 6505

remain important, while many companies are using them as flexible inputs alex.yao@jpmorgan.com

within broader product and workflow stacks. A useful working framework is SAC Registration Number: S1730523020001

“own the context, rent the model,” although we would still treat this as a Olivia Xu

direction of travel that needs more public-market proof. (86-21) 6106 6138 olivia.w.xu@jpmorgan.com

This pattern appeared across enterprise data integration, insurance and SAC Registration Number: S1730525060001

Jiajie Shen, CFAfinancial-services risk, cross-border marketing, independent model

(86-21) 6106 6352

development and listed model-plus-infrastructure businesses. Several jiajie.shen@jpmchase.com

companies described routing tasks across different domestic and frontier SAC Registration Number: S1730520030006

models based on price, performance and use case. One listed model-plus- J.P. Morgan Securities (China) Company

infrastructure company also acknowledged that raw API switching costs can Limited