Smart Data Foundry
Igniting Innovation through Synthetic Data

In the rapidly evolving fintech world, decision-making and innovation hinges on having access to data. However, getting relevant and unbiased data while navigating privacy and regulatory constraints such as the General Data Protection Regulation (GDPR) or California Consumer Privacy Act (CCPA) is not an easy task.

How do businesses overcome this predicament?

Enter Smart Data Foundry, a specialist in crafting high-quality synthetic data for financial institutions and fledgling fintech enterprises. It provides a powerful boost to their innovation efforts while ensuring efficiency and security. This synthetic data, devoid of any real-world data concerns sidesteps biases and sensitivities by controlling information such as demographics and other characteristics. Furthermore, it contains no sensitive or personal data, thereby upholding stringent privacy regulations.

With privacy concerns diligently addressed, synthetic datasets can circulate more freely among researchers, developers, and organisations, fostering collaboration without compromising data security and privacy. This positions Smart Data's synthetic data as a catalytic force for innovation, whether it's for software development, machine learning model training, business model creation, or generative AI solutions.

Smart Data Foundry's clientele spans advisory firms, regulatory bodies, and various organisations within the financial spectrum.

“We are committed to unleashing the potential of data to drive innovation” says Bryn Coulthard, chief product and technology officer, Smart Data Foundry.

At the heart of Smart Data Foundry's capabilities lies its synthetic data engine, ‘aizle.’

This innovative engine employs agent-based modelling to generate data sets tailored to the requirements of organisations, spanning banks, established financial authorities like the UK’s largest regulatory body, the Financial Conduct Authority (FCA) to emerging fintech startups. What sets ‘aizle’ apart is its ability to generate data independently of real-world inputs, making it an invaluable tool for igniting the development of groundbreaking products.

aizle addresses many challenges faced by clients, such as fraud detection, anti-money laundering, or know your customer (KYC) processes. It achieves this by populating its datasets with instances mirroring the problems clients aim to resolve. This ensures that clients access reliable and labelled training data, essentially providing them with a clear roadmap for their data requirements.

Nonetheless, even within the realm of synthetic data, ethics remain paramount. Smart Data Foundry provides ‘explainable’ or ‘training’ data, offering clients insights into the data's origins and composition. This commitment to transparency prevents biases related to demographics, income, race, or other factors from infiltrating the data. Such liquidity is quintessential for developing trustworthy, generative and predictive AI solutions.

A distinctive facet of Smart Data Foundry's service is ‘signal amplification.’ For instance, in the context of fraud detection, realworld fraudulent activities may constitute only a minuscule fraction of all transactions. To aid developers in identifying patterns and techniques effectively, aizle can amplify fraudulent instances in the dataset, creating a denser landscape. Once AI models are honed with this amplified data, it can gradually be reduced to mirror real-world proportions.
Clients can choose from an extensive array of ready-to-use datasets, encompassing open banking and fraud detection, to SME and identity data. However, for clients with specific requisites not covered by existing datasets, Smart Data Foundry's aizle engine can tailor bespoke datasets on demand through a continuous subscription, affording clients the autonomy they need.

A remarkable success story revolves around the FCA, the UK's banking regulator. The FCA grappled with the menace of consumer payment fraud, specifically impersonation fraud, incurring annual losses worth hundreds of millions of pounds in the UK. The amalgamation of authentic data from the financial and telecommunications sectors to combat this issue proved daunting. Nevertheless, after forging a partnership with Smart Data Foundry, it received a comprehensive consumer fraud dataset encompassing four banks, two telecom companies, two years of transactions, and five fraud topologies—comprising a staggering 500 million rows of data.

This dataset engendered a groundbreaking collaboration across industries on a unified platform. It empowered the FCA to formulate effective approaches to combat fraud and potentially influence new legislation.

The key to Smart Data Foundry's success lies in its commitment to crafting sophisticated, integrated synthetic datasets that expedite innovation. With Smart Data Foundry at the helm, the fintech industry can envision a future where data challenges cease to stifle innovation and instead propel it forward.