Banking on the power of AI

Deloitte Philippines Risk Advisory leader Anna Pabellon writes about AI use cases for the banking sector in an effort to guide financial institutions that have yet to adopt the technology

11 September 2023

By Anna Marie Pabellon

A FEW months back during Deloitte Philippines' first CFO forum, my fellow speakers and I encouraged finance leaders from a broad range of industries to dip their toes into artificial intelligence and the metaverse if they hadn't yet, because these technologies were already having a transformative impact on the businesses that were putting them to good use. Take financial services: A 2021 Deloitte survey of business leaders with a purview into their organization's AI strategies and investments found that 86 percent of financial services AI adopters believe the technology will be very or critically important to success in the next two years.

It is now two years later. The banking industry has no doubt seen significant changes in the quality of products and services it offers after having applied AI and machine learning to different operational areas. How have they done it?

For starters, the sector has long been technology-dependent and data-intensive, so it is primed to benefit from the lightning-fast innovation AI can trigger. Looking at a sampling of AI use cases, one can see that this technology can be applied across different areas of a bank: the front office, core banking products and services, and back-office operations.

On the front office end, banks are using AI to address a key factor in nurturing customer trust: personalized services. Data-driven AI capabilities allow banks to analyze customers at a level of granularity that positions them to predict needs and behaviors more accurately. One bank that took this route used AI to predict customers' redemption preferences, which resulted in a 40-percent increase in reward program usage.

During the first year of the pandemic, meanwhile, the delinquency rate of local credit card issuers jumped to 11.5 percent, or more than two and a half times that in 2019. The health crisis triggered a wave of dire situations — sudden job losses, an exodus from cities that served as central business districts, and a global lockdown — that left credit card holders dealing with varying priorities. Applying a ""one size fits all"" strategy to reaching out to customers for collection purposes during such a time was unlikely to lead to much success.

In this regard, AI can help collection departments drive efficiencies and shift to preemptive strategies to better manage delinquencies and defaults. An AI-driven analysis of customer data can identify warning signals for those customers who are at risk of defaulting, predict why they might miss payments, and then suggest customized solutions to help them catch up.

The disruptive events of the past couple of years, including the emergence of technologies that are reshaping the way banks do business, have intensified pressure on financial institutions to comply with increasingly stringent government rules and regulations. Keeping up with compliance processes, for example, may be too much work for a purely manual approach. Banks can leverage AI to automate some of these processes, including detecting regulatory changes, to ensure they are always on the right side of regulators. When setting up regulatory reporting configurations, embedding AI with its deep learning and language processing capabilities can shorten the period between reading compliance requirements and incorporating changes to the systems that generate reports.

These use cases, which are just a small sampling of the way AI is driving efficiencies in banking operations, demonstrate the practical applications of a technology some banks still see as experimental hype. For financial institutions that want to get serious about their AI implementation journey but are still unsure about where to begin, defining a use-case-driven process may be a good starting point. Look for low-hanging fruit within the institution — i.e., diverse, achievable projects — and study how AI solutions can be applied to these projects to realize immediate successes.

With the success of some short-term use cases, move on to experimenting with prototypes that can be scaled up to the enterprise, making sure to plan for expectations around data use, timelines, business goals, and strategy. At this stage, it is crucial to give the implementing team space to explore different solutions and to fail; the failed trials generate insights that are just as important as those gleaned from successful attempts.

Another important mindset to have is a proactive focus on risks and ethics. Early in the AI implementation cycle, make sure to work only with processes and models that are ethical and regulation-compliant. Monitor AI processes to make sure, for example, that underlying data is protected and only used for agreed-upon purposes. This will help preserve trust among stakeholders, which is key in scaling AI for enterprise-wide deployment.

With the wall-to-wall media coverage AI is getting, it's easy to feel overwhelmed about its potential and even its risks. But the undeniable fact is that this technology is fast becoming mainstream. And as the banking industry shifts in response to our increasingly digital economy, AI could spell the difference between the institutions that thrive and those that are left behind.

As published in The Manila Times on 11 September 2023. Anna Marie Pabellon is Deloitte Philippines' Risk Advisory Leader.

Did you find this useful?