Nationwide unlocks rapid innovation with synthetic data

The financial sector is more dynamic than ever. Rapidly evolving digital services, disruptive challenger brands and the ever-escalating expectations of customers are pushing financial institutions to innovate faster.

Challenge

Recognising the opportunity of bringing in new technology in to the business, Nationwide is increasingly looking to work with Fintechs. The partnerships team need to run trials to test the capability of the technology. Many of these tools are data driven, so the team often face the challenge of giving the Fintech access to their core data assets.

Six month data provisioning processes

It could take six months to provision a data set to a thrid party in a secure environment and go through all the security and sign off processes.

Limitations of masking and anonymisation

The team often had to resort to using heavily redacted and masked data to get security sign off. This limited the scope of the trials the team could run as the data was not fully representative.

Complex transaction data

Time series transaction data is highly personal and contains very sensitive information. Banking transactions are notoriously difficult to anonymise effectively using traditional techniques.

Solution

Nationwide faced a common challenge – and synthetic data was the answer. By creating synthetic data, Hazy enabled Nationwide to unlock value, share data quickly and securely with third parties, and accelerate its innovation programmes.

Using a generative model, Hazy extracted the statistical information and complex relationships contained within Nationwide’s transaction data, creating a synthetic data set that no longer contained any of the original information.

This process eliminated any possibility that the data could be traced back to the source, providing total protection to Nationwide’s customers, and ensuring regulatory compliance.

To deliver the synthetic data efficiently and securely to collaborators, Nationwide built a secure sandbox environment on Azure to host the synthetic data.

Hazy's generator model can be deployed quickly to new sandbox environments and used to generate safe synthetic data the third party can access.

Results

Nationwide reduced the time it takes to provsion data to third parties from six months to three days. The synthetic data solution that Hazy delivered has drastically reduced the time it takes Nationwide to create safe, shareable data.

3 days

3 days to provsion data to third parties

0

No real customer data exposed to third party

100 ×

Potential to test 100x new vendors each year

Alex Bannister
We work with Hazy to unlock data for innovation in a way that puts our members and their data privacy first
Alex Bannister
Director of Strategic Partnerships, Nationwide Building Society
Next step

Want to do more with your data?

Hazy unlocks the real value of your data. Without compromising privacy.

Contact us