A.I. tools enable alternatives to underwriting loans

SOCIAL media behavior and online footprint can be incorporated in the underwriting process of loans extended by digital banks and mobile lending applications through artificial intelligence (AI), a Singapore-based big data firm said.

Digital banks and financial technology (fintech) platforms are generally designed to reach out to the unbanked sectors in order to promote financial inclusion. The unbanked population, however, usually do not have a credit history to support loan applications.

Thanks to technological advancements such as AI, an alternative credit profiling process is seen helping the unbanked segment apply for the muchneeded financing, Advance Intelligence (Advance.ai) Pte. Ltd. Regional Sales Director Aradhna Sharma told the Businessmirror in an interview.

How often an email address and a mobile number are being used is a potential metric for loan approvals, she explained, noting this is part of the identity verification of the borrower.

“If you are registering with an email address, there are services that can determine when the email address was created by crawling through the web,” Sharma explained.

Tracking behavior

THE activities of the email address, such as logins in websites and purchases online, can be tracked to make sure that the email is valid. If there is not much usage, she said that it is potentially a red flag.

“With Advance.ai, we’re able to check the status of telephone. Is it currently working? Is it off line? Is it switched off?” she added, explaining that checking such conditions will help the financial institutions verify whether the mobile number is fake or not.

There is also a search tool that enables the digital banks and fintech platforms to search for a person’s personal details such as name and birth of date and indication of online footprint, Sharma explained.

Tracking a borrower’s online behavior, she said, can help in determining the potential credit risk.

“What these innovative solutions are trying to do is to assist you in collating the data that is already available to help you create credit profile online that makes it easy for you to get accepted loans,” Sharma explained.

Consent is key

SHARMA explained that all the information collated by the AI tools are publicly available.

“The truth of the matter is, with so much activity happening online, your information is already being collected through apps you use, through Google, through GPS [global positioning system],” she said. “It is not that it actually isn’t collected; it’s just that it’s not currently being utilized in a way to help you.”

Still, Advance.ai stressed that the borrowers should be able to understand what they are consenting to when applying for a loan digitally. Sharma said clients must be aware that their public online data are being harnessed to build their credit profiles. The borrowers should be given the option to call upon an online service provider to explain vague items in the terms and agreements of loans before signing the deal, she said. Sharma said digital banks should “provide evidence that it [loan agreement] was communicated to the customer and the customers accepted and provided consent.”

It can be done through having the borrowers place their e-signature in each section of the borrowing agreement, acknowledging every term of the loan, she explained. Sharma said that the loan process could also be part of the electronic “know-your-customer” portion.

Digital banks, she said, must also be able to draw a line when data gathering is deemed excessive already. “There needs to be obviously some law that protects data privacy,” Sharma added.

Data management

THE company also highlighted the importance of data management strategy for the digital banks, with Sharma explaining that it is the organizations’ roadmap.

“This roadmap ensures that all the activities surrounding data management, which includes from collection to collaboration of data, are able to work together effectively and efficiently and useful as possible to the government [for audits and reporting],” she explained.

Sharma enumerated two classifications of data management strategy: defensive and offensive.

Defensive is “about minimizing the downside risks, ensuring compliance with the regulations...and building systems to avoid theft,” she explained. “Data offense focuses on supporting business objectives, such as increasing revenues, profitability, customer satisfaction.”

“We guide our clients towards adopting a framework that not only promotes efficient use of data and allocation of resources for example, but also help organizations design their data management activities to support the overall business goals,” Sharma said.

In November last year, the Bangko Sentral ng Pilipinas released the Digital Banking Framework, recognizing digital banks as a new bank category. Digital bank applicants, according to the regulator, should be able to have effective data management strategy and practices, apart from sound digital governance and secure technology infrastructure.