CLPS Achieves 90% Accuracy in AI-Powered AML Review for Chinese Bank…

Event summary

CLPS completed an AI-assisted AML review project for a major Chinese bank, achieving over 90% accuracy with a fine-tuned small LLM.
The project was conducted within a Hong Kong regulatory department's Gen A.I. Sandbox.
CLPS outperformed larger models with 400% more parameters, which achieved just over 30% accuracy.
The solution required only limited GPU resources, lowering deployment barriers for financial institutions.
CLPS developed a standardized financial AI model training framework for horizontal scaling across banking scenarios.

The big picture

CLPS's success validates the commercial viability of compact, fine-tuned models for vertical use cases in financial services. The project demonstrates that specialized AI capabilities can solve specific business pain points while keeping computing costs manageable, a trend likely to gain traction as financial institutions navigate cross-border compliance complexities. The standardized framework developed by CLPS could set a new benchmark for RegTech development in the banking industry.

What we're watching

Model Scalability: Whether CLPS can successfully apply its standardized framework to other banking scenarios like loan approval and credit monitoring.
Data Dependency: How the continuous ingestion of high-quality real-world business data will impact future model performance improvements.
Competitive Positioning: The pace at which CLPS can transition from project-based delivery to product-driven enablement in the financial AI sector.