BPI taps AI, digital cash for economic growth

The Bank of the Philippine Islands (BPI) is harnessing artificial intelligence or AI to make more Filipinos, especially in the provinces, financially free and cyber-secure.

“The goal is we want better financial inclusion in the country. This country is severely backward because of a lack of economic opportunities for many people,” BPI president and chief executive officer Jose Teodoro “TG” Limcaoco said during the weekly Rotary Club of Makati meeting in which he was the guest of honor and speaker last Tuesday, 2 April.

Limcaoco said BPI’s AI-powered mobile app featuring basic banking services and digital cash allows money to circulate within lower-income communities, boosting the finances of their individual households and businesses.

“What happens is the physical cash there gets sucked up as it goes to a sari-sari store and a sari-sari store then goes to a major grocery store in the city, and eventually goes back to Metro Manila,” Limcaoco explained.

“A community cannot flourish without cash because commerce cannot take place,” he stressed.

Win-win

Limcaoco said digital cash and technology enable a win-win situation as money can be exchanged digitally at low costs to the bank and its customers.

“We’re a country of too many islands to move cash everywhere. We cannot have cash centers because it’s expensive to build and operate them,” Limcaoco said.

Digital cash is the customers’ money in the bank and used through a mobile app instead of pulling it out from a physical wallet, while AI is used by machines to collect and analyze customers’ data to help the bank provide them proper products and speed up its services.

By 2026, BPI expects 50 million Filipinos benefiting from its digital system. Limcaoco shared that the system now serves 12 million.

“It’s a long process but we need people to get used to the idea that cash can be transferred digitally,” he said.

That process, Limcaoco said, begins with retail stores under its “agency banking” strategy for opening BPI accounts.

These stores include Robinsons-owned stores, Ayala malls, Uncle John’s, Southstar drugstores and Seaoil gas stations.

“We’re working with retail stores where we’ll put our technology and hopefully customers will walk into these stores and be onboarded into the BPI’s system,” he explained.

BPI now has 5,000 partner-stores, Limcaoco added.

Power of AI

Limcaoco said AI allows wealth-building and protection through its multiple functions.

First, its huge data-gathering provides BPI with various types of data that describe clients’ behaviors, including those of entrepreneurs.

“We take economic data and profitability of stores. For example, we can now tell where we can open a branch that serves micro, small and medium enterprises (MSMEs),” Limcaoco said.

Based on its AI-generated report, he shared that many MSME clients and business transactions spur in third-class municipalities where there are “lousy” governments.

“This is very important because when the local government is lousy the entrepreneurs take over. They don’t have the support of the government so they do their own business and they need banking,” Limcaoco observed.

“So we’re looking for lousy local governments to open microfinance branches,” he said.

Machine-learning capability

Second, AI’s machine-learning capability allows the bank to predict whether its clients will continue good or bad and usual or abnormal behaviors.

“What machine learning does is to take a large amount of data and we train the machine to predict what happens next given similar data,” Limcaoco said.

He said starting MSMEs can avail of higher BPI loans once they have remained the bank’s clients for six months.

“Once we see the first six months of data on where the cash goes in and out, we can tell if you have good business,” Limcaoco said.

Maximizing the use of AI, he added that BPI is training and testing its machine to analyze financial reports of companies to see the “weak spots” as a risk protection to all types of clients.

“Lending to MSMEs has always been challenging because of BPI’s conservative credit culture. That’s how we want to operate the bank because I’m not lending my money; we’re lending your money,” Limcaoco stressed.

BPI is also feeding its machine the bank’s product, procedural and policy data to ensure consistent and accurate customer service through AI’s language tool called ChatGPT.

It allows customers to type their questions to chat robots online or chatbots and receive answers in a human-like way.

“We will be launching our internal ChatGPT. We have 20,000 people and some have been in the bank for many years and some have only come, so you have different levels of knowledge,” Limcaoco said.

He said machine learning is also useful for detecting activities linked to money laundering, tax evasion, and gambling, among others.

“Machines need to learn behaviors so we can just flag them and report them to the central bank and other government authorities,” Limcaoco stressed.

He said machine learning also alerts the bank on fake clients attempting to steal money or complete other crimes.

“When you open accounts online, you’re asked to take a picture. That picture runs through a machine learning system to make sure that it’s a real person and sometimes it will ask you to move,” Limcaoco explained.

He recalled learning about a $25-million theft in a Hong Kong-based firm using AI.

“Generative AI can mimic your voice. The firm’s officer was asked to transfer the money but he was fooled that he was in a Zoom meeting with its chief executive officer and the chief financial officer,” Limcaoco shared.

Human touch

“AI will never have all the answers because it uses only what it has been taught,” Limcaoco cautioned the public amid fears that human workers will soon be replaced by the technology.

For example, Limcaoco said investment and loan services to newly banked clients and non-financial industry professionals will always be human-driven.

“We’re finance people and go to Google and read about investments and invest because we can understand. For most people, they can also read about investments but they won’t have that level of comfort,” he said.

Limcaoco said BPI continues to redesign its branches to encourage financial literacy among Filipinos.

“We’re also working on the concept that our branches are places where you can learn, where people can go and ask questions rather than do transactions,” he said.

“There are very transactional services like fund transfers or payments and we want people to do that on their phones,” Limcaoco added.

What’s the future?

Limcaoco deems the journey of BPI with AI and other technologies to move relatively slowly but surely.

“Today you can use generative AI in many ways. It’s quite generous because you teach the machine, but it reflects the prejudice or bias of the teaching,” he said.

“At BPI, we use emerging AI tools internally where only employees have access to them and use it before communicating with the clients,” Limcaoco stressed.

Moving forward, he sees smarter AI tools that can aid humans in discussing investment products while the bank’s employees also deepens clients’ trust in BPI.

“I foresee that one day customers can talk with the tool about their investment portfolios, investment alternatives, or what they should be doing about wealth planning and insurance,” Limcaoco said.

Instead of using keypads to accomplish financial transactions, he said customers might be able to simply verbalize their instructions.

“I see the day when you can use your voice, tell the app to transfer funds from my account to my wife’s account, for example,” Limcaoco said.