SWIFT & Azure confidential computing: Innovating financial transaction security

Who: SWIFT & Microsoft Azure

Amidst the growing concerns over fraudulent cross-border transactions, SWIFT, the leading infrastructure provider for financial messaging services, has been innovating to secure its community of over 11,500 institutions. In partnership with Microsoft, SWIFT leverages federated learning techniques, Azure Machine Learning, and Azure confidential computing to develop a state-of-the-art anomaly detection model—without transferring data from secured locations. The shared vision is to set a new standard in reducing financial crime while optimizing security, privacy, and cost efficiency.

What: Anomaly detection model for financial transactions

The project aims to establish a model that identifies anomalies in transactional data using Azure Machine Learning. Rather than centralizing data, the model for training is dispatched to participants’ local compute and datasets at the edge, converging the results in a foundation model. This unique approach fosters robust insights and fosters enhanced anomaly detection, leading to heightened financial transaction security.

Before: The challenges in the financial industry

Cross-border transactions and instant payment networks are growing rapidly, increasing both the risk and cost of fraudulent activities. As the industry standard for financial communication, SWIFT’s global financial messaging network handles more than nine billion financial messages annually, making security paramount. The heightened risk of fraud, which potentially costs billions annually, is a significant concern for SWIFT’s clients, necessitating collective action.

The Process: Federated learning & Azure confidential computing

To combat the rising security threats, SWIFT, under its comprehensive innovation agenda, collaborates with Microsoft. The partners employ federated learning techniques with Azure Machine Learning, Azure confidential computing, Microsoft Purview, and a Zero Trust-based policy framework. This partnership aims to construct an accurate anomaly detection model for financial transactional data—without the need to copy or move data from secure locations, ensuring the confidentiality of participants’ data.

After: A new standard in financial security

While the full potential of the model is yet to be realized, the shared vision indicates it will become a powerful tool in mitigating financial crime while maintaining high security, privacy, and cost efficiency levels. SWIFT plans to share the model with its banking partners to allow them to contribute their data to further enhance the model’s accuracy and effectiveness. This creates a cycle of continuous learning, dramatically improving the rate of fraudulent transaction detection worldwide.

Conclusion: Harnessing collective efforts for greater security

For SWIFT and Microsoft, this initiative goes beyond providing a technological solution—it’s a collective mission to enhance the global financial ecosystem’s security. As SWIFT’s Chief Innovation Officer, Tom Zschach, states, Microsoft’s strategic partnership is invaluable in this venture, given their shared core values and capabilities. The ongoing collaboration promises a future where a safer, more secure financial transaction environment prevails, benefitting financial institutions and their customers globally.