Long reads

Artificial intelligence, data and technology adoption – a balancing act

Dominique Dierks

Dominique Dierks

Content Manager, Finextra

This is an excerpt from The Future of Digital Banking in North America 2024 report.

Safe and accurate data is a foundation for financial companies. Enriching and authenticating data between parties will enable institutions to use their data more intelligently, make better decisions, address real customer needs and offer embedded finance. Yet in real world terms, data fragmentation is still one of the biggest challenges for banks and financial institutions.

While having a trove of data at their disposal, Coveo’s 2023 Workplace Relevance Report found that 16% of bank employees have to search more than seven sources to be able to execute their responsibilities. Data silos within banking environments as well as unreliable information from their institution’s intranet doesn’t only slow down employees, it also leads to flawed decision making.

ISO 20022 adoption will be one standard that will greatly improve the ways data is read and understood. The historic introduction of ISO 20022 to the Swift network sets a precedent for how data is secured, stored, and distributed. This data standardisation will allow financial institutions to boost operational efficiency as well as enable innovative new services. While the transition period to ISO 20022 is scheduled to end in November 2025, the question remains whether banks will be able sufficiently prepare their systems in time. A Celent survey of 211 global banks and corporates found that only 56% of North American banks believe the industry will be ready. 2024 will prove an important year for financial institutions to invest in digital transformation and ensure their systems are ISO 20022-capable.

Another way to combat banking fragmentation, especially as we’re moving towards open and embedded finance, are data clean rooms. Data clean rooms are used by financial institutions to analyse confidential transactional data while ensuring customer privacy. Yet another challenge remains: as we’re facing a rapid growth in the amount of data that is generated, it’s becoming impossible for financial institutions to absorb, interpret and make meaningful decisions based on the data available to them. This is where AI comes into play.

We need to talk about AI

Finding the balance between tackling new risks, adhering to regulation, protecting data and embracing new technologies is one of the biggest challenges financial institutions currently face. According to a Temenos survey, 63% of banking officials expect emerging technologies (such as generative AI, blockchain or quantum computing) to have the biggest impact on banks in the next five years.

While generative AI will not be ready for consumer facing applications in financial services by 2024, it can help institutions make sense of their large amounts of data. Its ability to sift through unstructured forms of data and summarise information found can enable teams to improve efficiency and better scale. In a real-world use-case, Morgan Stanley this year announced that their wealth management division is leveraging Open AI’s generative AI capabilities for the deployment of an employee-facing chatbot. Designed specifically with appropriate controls for their financial advisors, the teams will be able to “ask questions and contemplate large amounts of content and data, with answers delivered in an easily digestible format generated exclusively from MSWM content and with links to the source documents.”

Aside from data management, AI and machine learning also offers vast benefits for fraud detection and mitigation. Combining deep learning algorithms with machine learning-enabled tools allows financial institutions to improve their data analysis and detect anomalies that might otherwise be missed. Large datasets help AI make better predictions and intuitively set fraud transaction monitoring thresholds.

A Forbes Advisor study from this year found that 51% of business owners already use AI for cybersecurity and fraud management. “When a new fraud pattern occurs in one region, AI can help identify if fraud-detection models in other regions are prepared for this same type of incident. This can be applied to transaction approvals as well,” comments Ed McLaughlin, CTO at Mastercard. “Over time, this will result in reduced fraud rates and increased approvals on legitimate transactions, both of which are key to financial institutions and their end customers.”

When it comes to managing risks while embracing new technologies, Nikhil Lele, EY Global, emphasises the need to prioritise secure adoption before fast adoption. “Adoption of Generative AI, embedded finance, and data privacy mandates will continue to accelerate. As a result, safety, soundness and security will remain top of mind priorities for bank leaders to instil trust among customers and the financial system overall.

“Technological capabilities will continue to outpace banks’ ability to adopt these innovations responsibly. This is due to the heightened need for banks to ensure continued trust in the financial system as well as compliance with constantly evolving legal and regulatory mandates. Not receiving a delivery within two days carries a different impact than a payment not being processed on-time, or an adverse lending decision made based on opaque models. Banks will always need to take the safer route when adopting these technologies, doing so in a way that improves financial access, trust and customer outcomes.”

Comments: (0)