Malta: Strengthening Your Data Management

Strong data management is a foundational requirement if financial institutions are to meet regulatory requirements around security, privacy and trust. This is in many ways the first base for data management. Regulators are increasingly focusing on these issues, enshrined in legislation such as GDPR and in Australia, Consumer Data Rights (CDR).

It is also necessitated through moves to create Open Banking ecosystems, requiring financial institutions to share customer data with other organizations on request to improve competition and choice for customers. In many jurisdictions, Open Banking systems are already advanced. In Australia, while it has been somewhat delayed by the COVID-19 pandemic, it is in active development. In order to comply with the demands, organizations need to ensure that their data is accurate, of high quality and that they can retrieve the relevant information quickly to meet the required timescales. But it is not just a defensive compliance issue – Open Banking is also an offensive opportunity, to gain new customers by making quick decisions based on the data provided.

Another area where data management is critical is in lending. More precisely, it is fundamental to the lending transformation work that has been made necessary through COVID-19. Regulators and governments have been looking to financial organizations to make financial support rapidly available to customers and businesses who have been hit by the effects of the pandemic. Across unsecured and secured lending, this means being able to quickly access high fidelity, high integrity data to make the necessary credit decisions and authorizations.

However, many lending processes remain heavily manual. Some mortgage lending, for example, can consist of over 100 separate steps. It has therefore become necessary to review and update core lending processes, including quite dramatic changes to how data is used. It is a combination of overhauling processes and introducing greater levels of automation and AI-driven decision flows into models – and this requires fast access to reliable data. It is an area where KPMG has been working closely with many financial organizations in the wake of the pandemic.

More broadly, data management is a vital element to delivering the enhanced customer experience that organizations are striving to create in the digital age. Nearly every financial institution is (or should be) investing in developing improved digital channels that make managing financial affairs easy, simple and relevant for customers. 

Increasingly, this includes broadening services out in highly personalized ways – to bring customers opportunities and offers in relevant adjacent services such as shopping and retail, energy and utilities, or payments services. It is about understanding a customer's needs and wants, and providing relevant services to them as a result. But you can't have a personalization capability at scale without strong data management. As an example, Commonwealth Bank of Australia (Commbank) Customer Engagement Engine runs 400 machine learning models and ingests a staggering 157 billion data points in real time to deliver highly relevant, contextual messages to customers.

Indeed, data has become more than just an asset. It has become the driver and facilitator of new business models and innovation. In the digital, connected world – and the Open Banking era – there is an opportunity to bring in data from third party providers and offer more targeted services. We see this with real estate data providers – enabling banks and lenders to apply it to mortgage offerings, and in Buy Now Pay Later businesses (such as Afterpay or Klarna) where data sets can be analyzed for insights into credit profiles and purchasing trends within sectors and categories.

Major banks and other institutions may have several hundred highly skilled staff working on these data and personalization models. There is an impressive amount of horsepower and intellect being applied. But now it is moving beyond internally focused work to look outwards and create data partnerships with an ever more diversified set of external players.

For the more traditional and established institutions, one of the key hurdles to achieving strong data management is that age-old issue – legacy systems. Achieving a single view of the customer from multiple and disparate sources is an ongoing challenge, especially as more data is generated in different formats all the time.

For fintechs and other digital native organizations, whilst they are naturally in a good position to embed AI, machine learning and automation to drive personalization and a good customer experience, they have perhaps not focused as much as they could have on data management aspects. In a sense, they haven't needed to – all their data is typically already well integrated from the outset. But as their data volumes grow and they become more complex as organizations, and as regulators continue to raise requirements around how data is stored and secured, it is an area that they are likely to need to put more resource behind.

Another growing and potentially significant issue is achieving 'responsible AI'. Clearly, as data is leveraged more and more through AI models, it is key that customers, regulators and shareholders have confidence in the technology. The explainability and transparency of AI models, including how data is managed through the process, is paramount. And yet a recent survey by KPMG in Australia in conjunction with the University of Queensland found that only a third of Australians are willing to trust AI systems. This is indicative of a wider issue internationally. It is imperative that financial institutions bring customers and stakeholders with them on the journey as they embed AI and emerging technologies ever more deeply.

No financial institution can win in the market without leveraging data – and to do that they have to manage it appropriately. Data management is integral to an effective data strategy and is becoming an increasingly critical determinant of success.