Trade Finance and
Documentary Credit Notes
Articles from our team on documentary credit processes, ICC rules and RegTech.
The Model Comes to the Data: Federated Learning for Private AI
Federated learning lets organisations improve a shared AI model without moving sensitive data outside their own infrastructure. By combining local LoRA fine-tuning, federated averaging, secure aggregation and differential privacy, institutions can learn from each other's data while keeping documents, corrections and records private. For our document-processing pipeline, the strongest starting point is federated fine-tuning of the vision extractor, where client-specific layouts, stamps and scan conventions create the most value. Yapıştırılan metin
Read Article →
A New Era in Letter of Credit Document Examination: Scaling Human Expertise with Artificial Intelligence
The examination of letter of credit documents—one of the most critical processes in foreign trade—is still carried out manually in most parts of the world today.
Read Article →