AI transforming the future of banking, finance

IN 3D UnionBank’s booth at the Singapore Fintech Festival 2023, attended by about 10,000 techies, features 3D walls to underscore its theme of how the bank has embraced human-centric AI. — CONTRIBUTED PHOTO

SINGAPORE—Artificial intelligence (AI), once considered the proverbial “elephant in the room” that no one talks about, is all the buzz at the Singapore Fintech Festival (SFF) 2023 this week.

And Union Bank of the Philippines—which thinks of itself as a technology company with a banking license—has seized the opportunity to make it a roadshow of how AI is transforming the future of banking and finance.

The theme of UnionBank’s booth at SFF, the world’s biggest gathering of fintechs that drew 10,000 participants this year, is how ‘human-centric AI’ has been infused into the fabric of its ecosystem. Its AI tools focus on three areas: customer experience (via hyper personalization), cybersecurity/fighting crimes and optimizing business revenue/operational efficiency.

In hyper personalization, sending the same email or SMS blast to all is a thing of the past, as AI now enables the offering of products to customers with high likelihood of availing.

Ana Aboitiz-Delgado, senior executive vice president, tells the media that since 2020, UnionBank’s relationship managers (RMs) have used an AI-powered platform called MAX (Mobile Assistant Express), which suggests which leads to pursue, and matches each new prospect to the RM with the most success in selling that particular product.

“Since we deployed the MAX platform, we’ve actually increased sales productivity by 400 percent. So this is a case where AI does not replace but enhances productivity so much,” Delgado says.

AI models are able to predict the best timing to pitch. After all, it’s not everyday that a person needs a loan to buy a house or a car.

Up to 80 percent of an organization’s data is typically unstructured and here lies a “treasure trove of insights” that the bank has started to mine.

Customer feedback via free text is one such valuable resource. Rather than sifting through thousands of texts manually, advanced natural language processing can extract topics, identify major issues and uncover “delighters” or features that customers do not expect but appreciate once given to them.

Based on UnionBank’s reckoning, such proactive approach has enabled swift intervention and enhanced efficiency by about 50 percent.

Adrienne Heinrich —Contributed photo

Bank of fintechs

Ramon Duarte, executive vice president and head of transaction banking, says that 80 percent of fintechs operating in the Philippines bank with UnionBank. “That speaks a lot about our capabilities,” he says.

In transaction banking, Duarte says AI is very relevant day to day: as customers pay their suppliers, collect money from dealers and distributors or pay other liabilities.

“Along the way, there’s a lot of data passed on and that’s a huge source of insights. Modeling things like alternative scoring opens up the opportunity for lending to other ecosystem participants like payroll or constituents, benefactors or whoever [needs it],” Duarte says.

“We take that data and put meaning into it so people can makes sense of the data,” says Erica Dizon-Go, UnionBank senior vice president and head of corporate products, open finance and digital services. Those could thus be used to solve the pain points of customers, she adds.

The bank has an AI-powered suite of solutions for corporate clients, including eDAO (ePaycard Digital Account Opening), a fully digital payroll and carding solution, which has grown payroll user count by 127 percent in the last three years. Another is UPAY, a platform that consolidates several payment methods and provides a single integration point for collections needs.

UnionBank also partners with Singapore-based affiliate Aboitiz Data Innovation (ADI) to deliver bespoke data parlays, data models and data lake housing (for business intelligence and machine learning).

Jaime Garchitorena, senior vice president and head of trade and supply chain finance, says that by using technology to deep-dive into customers’ behavior, the bank can tailor-fit product offerings. When a customer is using credit card to buy a small truck or a refrigerator, for instance, he says it suggests that this person must be opening or expanding business and in need of financing.

UnionBank’s P72-billion acquisition of Citibank’s retail and consumer banking assets in the Philippines has indeed opened up a lot of cross-selling activities, adds Delgado.

AI TALK UnionBank senior officials (fromleft) Jaime Garchitorena, Erika Dizon-Go, Ramon Duarte and Ana Aboitiz-Delgado —PHOTO BY DORIS DUMLAO-ABADILLA

Combating financial crimes

In the past, UnionBank took a reactive approach to detect mules or those who use their bank accounts to transfer dirty money, mainly through complaints from customers reporting suspicious transactions. The investigation was manual and took weeks.

Adrienne Heinrich, head of AI and innovation at ADI, explains that using the Mules Detection AI solution, the bank could now more easily identify probable mules, thus prompting proactive investigation. It shortened the detection process from more than a month to only a few days — 10 times more efficient as AI never sleeps in flagging the outliers.

Using network analysis, the bank tracks complex relationships and dynamics between account holders and an individual’s role within a network of people.

“We can also analyze evolutional emergence of fraudulent patterns within the network, which allows for earlier identification of these fraudulent patterns and connected actors,” she says.

Authentication through voice is also seen to play a key role in fraud prevention. Despite the prevalence of security questions and one-time passwords, many people are still being tricked to divulge critical information.

“Imagine your voice becoming your password—within mere seconds of conversing with a customer service agent, a matching score can verify the true account holder. For UnionBank, we care about making processes easier for our customers without impacting much of their perceived workflow,” she says.

Bank branches deal with thousands of suspicious transaction alerts but 99 percent turn out to be false positives. Using AI, alerts are sorted based on risk levels and invalid alerts are detected. Since its launch in late 2022, UnionBank has seen a 40-percent reduction in alerts referred for manual investigation.

Operational efficiency

Using historical data, UnionBank’s collection models predict which accounts are likely to miss their next payment. By knowing which to prioritize, the bank can manage its collector capacity, thus improving collections flows and reducing credit losses while preventing customers from falling further into debt through timely reminders.

The bank has also developed a cash mobility solution that digitizes the cash delivery and pick-up for corporate clients that use its cash management service. This is to optimize and automate trip schedules and routes of armored trucks, thereby reducing travel time and carbon emissions.

The algorithm enables real-time tracking of the armored trucks. The mobile app enables roving tellers to view their trip schedule for the day, input actual transactions and review them.

To date, ADI has deployed a multi-stop route optimization solution for Shiptek, a logistics company and subsidiary of UnionBank. It takes into account domestic challenges, such as road condition, traffic, type of goods and respective customer interaction and loading time.

At the end of the day, AI is only as good as its employer. Cognizant of concerns on AI, particularly the displacement of human staff and cybercriminals stepping up their game, UnionBank officials stress the imperative of shaping AI development with a focus on transparency, accountability and ethical considerations.