Smart & safe synthetic data for sharing transaction intelligence in financial services

SWIFT has been providing financial messaging services worldwide since the 1970s. Today, SWIFT still enables users across the world to communicate with one another safely and securely at the same time as innovating to create value and deliver efficiencies. SWIFT is now playing a central role in bridging the old world with the new, by establishing a community of banks and ecosystem partners to support the deployment of artificial intelligence (AI) at scale.

The goal of this community is to enhance the effectiveness and efficiency of shared services such as anomaly detection through the deployment of AI at scale. SWIFT’s unique role at the center of the world’s global transactions holds the key to unlocking the truly transformative power of AI, and they have just added MOSTLY AI, the world’s leading synthetic data company to their diverse partner ecosystem. 

Deepak Janardhanan, SWIFT’s AI Platform Lead described the cooperation as: "SWIFT's goal is to deliver privacy-preserving AI at scale for the Payments and Securities industry. Therefore, synthetic data will play a crucial role in machine learning exploration. Deploying and training AI models on MOSTLY AI's synthetic data generation capability is a foundational step towards our goal."

The data challenge in financial services

According to research by Gartner, 80% of legacy financial institutions that don’t embrace digital transformation are predicted to fail by 2030. To avoid this, institutions will need to adapt their mindsets, processes and structures. The rich insights contained in transaction data can fuel digital transformation, helping banks to realise the potential of AI. However, data sharing in a highly regulated and sensitive environment is a hard, slow, and often painful process. Legal and regulatory pressures make it impossible to collaborate efficiently with industry partners.

Synthetic transaction data is the way forward

To overcome this challenge, SWIFT has partnered with MOSTLY AI to generate high-quality synthetic transaction data that looks, feels, and means the same as the original transaction data. Synthetic data is a perfect proxy for the original since it contains the same insights and correlations – without any of the privacy concerns.  

Good quality synthetic datasets can be used in a variety of tasks in a privacy-safe, agile manner. None of the synthetic data points can be traced back to a data point in the original set. This makes synthetic data privacy safe and compliant with the strictest privacy laws and regulations. What’s more, the process of synthesization can be used to augment the original data: fix embedded biases, up-sample rare events or generate edge cases that are not present in the original.

To kick-start our synthetic data partnership, MOSTLY AI has recently completed a synthetic data superuser training with SWIFT’s AI Centre of Excellence. The intensive 5-day training was delivered by a cross-functional team of MOSTLY AI’s synthetic data experts and thought leaders. The training guides new superusers through the core concepts and principles of synthetic data, the end-to-end generation process, and how to assess, interpret and communicate the privacy and accuracy of synthetic datasets.

According to Russell Wolz, Data Scientist at SWIFT, one of the participants of the synthetic data superuser training: "The MOSTLY AI super-user training was very well put together and structured in a great way that introduced important, interesting teaching materials while giving plenty of hands-on experience to learn the ins and outs of the application. This allowed us, users, to gain complete knowledge of the application, and it was a great resource to have many MOSTLY AI team members for training discussions and questions. The training has provided a great foundation to start us on our synthetic data generation journey at SWIFT."

Developing in-house synthetic data capabilities is one of the cornerstones of building robust and scalable AI. SWIFT is at the forefront of the AI-first trend in payments and financial services. No two organizations are the same stage of this transformation and the importance of receiving tailored training cannot be overstated. The modular design of MOSTLY AI’s superuser training is perfect for this.

This flexibility was well received. Ryan Byrnes, Data Scientist said that "MOSTLY AI adapted their super-user training to our needs and did a great job balancing concepts with technical details."

Considered the gold standard for reliability and security in financial transactions, SWIFT is used by over 11,000+ financial institutions across 200+ countries. The partnership with SWIFT is yet another strong validation of MOSTLY AI as an emerging player and trusted partner to the world’s largest and most innovative financial services firms. MOSTLY AI is proud to partner with SWIFT, and the team is excited to play an important part in shaping the future of AI in financial services.