Yiren Digital Shifts to Shared AI Model for Multi-Industry Expansion

The company’s strategy centers on a common enterprise architecture that integrates proprietary large language models—specifically Zhiyu and Yizhi—with the MagiCube 2.0 multi-agent platform and the XuanJi execution engine. Rather than building bespoke AI for every project, Yiren Digital now treats components like fraud detection models and decision layers as reusable assets. This modularity allows the firm to adapt existing workflows to new business units without rebuilding core governance or infrastructure from scratch.

Chairman and CEO Ning Tang described the shift as a move toward an autonomous financial system. By validating these technologies within its existing credit and insurance portfolios first, the company intends to lower the cost of future innovation. The goal is to create a scalable foundation where standardized AI capabilities can be seamlessly injected into customer operations, asset recovery, and risk management, effectively turning isolated technical successes into a repeatable enterprise engine.