THE USE of artificial intelligence (AI) in lending processes could accelerate Philippine banks’ credit growth to 15-18%, according to financial technology (fintech) company Intellect Design Arena Ltd.
Bank lending growth could pick up as banks address bottlenecks in the lending process with the use of AI, Intellect Design Arena Executive Vice President Lending Head Brajesh Khandelwal said in a group interview on Thursday.
“See, the way it would happen is the market as a whole might take some time, but the banks who adopt the AI-based lending practices, they reimagine the lending with artificial intelligence, with the enterprise-grade AI platform, and carefully design their workflows using AI. They would be able to command the growth and others will follow,” he said.
“As a market, there is a potential. It all depends upon how banks are adopting the AI, how they are choosing the right platforms, how they are managing the change.”
As of September, bank lending grew by 10.5% to P13.704 trillion from P12.401 trillion a year earlier, preliminary Bangko Sentral ng Pilipinas (BSP) data showed.
This eased from the 11.2% annual increase in August and was the slowest in 14 months or since the 10.4% seen in July 2024.
With this current trend, Mr. Khandelwal said there “huge potential” for growth in the credit gap seen for micro, small, and medium enterprises (MSMEs), which is largely due to slow loan processing and management.
“The time of getting the loan goes for days and sometimes several weeks as well. So, we do see a huge potential for improvement in the productivity, in the efficiency and making the banks more competitive in the region.”
Banks can also use AI to manage loan repayments, he added.
“I think that governance is very important because growth without that guardrail is not going to work at all. It will not sustain. For the growth to sustain, nonperforming loans have to be in the manageable range.”
As banks adopt AI, guidelines that focus on personal data protection and transparency should be implemented as Philippine institutions remain in the experimental stage compared to their peers in the region, he said.
“There is an expectation of non-bias when you implement an AI-based model. It should not be biased. Then there is always the expectation of a human in-look. While AI can do a job to an extent, then there has to be a right strategy to decide where the human will be brought in the loop.” — Aaron Michael C. Sy