AI, Big Data as Key Drivers for State Digital Bank

KUCHING: Adopting technology such as artificial intelligence (AI) and big data analytics may be the key component towards a successful Sarawak-based digital bank.

With the right strategy and execution blueprint, Business enabler ecosystem firm Vision Group managing director Chua Zhu Lian believed that a Sarawak based digital bank can certainly become the mainstream in the state’s digital economy.

“And with the right business model and human capital to drive the bank, we strongly believe that Sarawak Digital Bank can eventually become a game changer to the banking industry, especially after the first 3 years of operations,” he said to The Borneo Post.

“One such component for a successful digital bank in Sarawak is AI. With this technology, Chua said account opening, loans processing and disbursement can be instantaneous and cost efficient while effectively filtering quality applicants.

“This would mean digital bank customers could open an account, apply loans and obtain money into the bank account within the same period of time.”

Chua explained that for account opening, artificial intelligence can remotely ascertain the identification of one person through optical character recognition (OCR).

This technology enables automated data extraction from documents or images to verify the identity of a person through the Identity Card provided and matched against a person through a live video interaction.

“It can then perform an automated search to identify if the applicants is blacklisted elsewhere due to money laundering or other malicious purposes and be selective on which ones to approve,” Chua added.

“This technology would reduce costs significantly (branch set up cost, manpower cost to process the application manually) while convenient to the customers. Anyone with a working smartphone equipped with camera function and wifi access can do that seamlessly.

“This is also a requirement by BNM whereby banks must know the customer and be able to verify the person through its biometric data (such as face, fingerprint) to prevent identity theft for malicious purposes.”

Compared to traditional banking model, Chua said it requires customers to queue up at physical branch, fill up the necessary paper works, and wait for approval in days and come back to the branch to set up the ATM card, and thereafter to set up internet banking account separately with the pin and email registration.

Meanwhile, Vision Group deputy managing director of group digital technology Albert Chee also proposed the use of big data analytics and machine learning to drive the lending model of Sarawak’s digital bank.

“Lending primarily involves credit risk, specifically the risk of non-repayment. To give some simple example, a loan aggregated up to RM100 million with a net interest margin of four per cent will only translate to RM4 million profit before taking into account of all the ancillary costs.

“Assuming each loan size is at RM1 million, this would mean that of the 100 applications (RM1 million times 100), the allowed margin of error for nonperforming loans is only at 4 applications (or a total of RM4 million) before it wipes up the entire profit.

“In real life, the size of loans differs with different risk dependent on borrower’s repayment capacity. This is where big data analytics comes in to provide an objective assessment to evaluate the risks, be selective to the customer and able to minimise the nonrepayment risks.”

Chua opined that big data analytics is especially important for a digital banking model as it mainly focuses on the digitally savvy crowd, typically millennials and Gen Z as well as the underserved segments such as micro business owners as a majority of this population are operating in the gig or informal economy.

“For gig workers who wanted a bank loan to upgrade their devices (smartphone with better camera, lighting equipment, microphones), they will be rejected by the banks if they do not have prior credit history records.

“This is because most banks have limited models to assess the repayment capacity beyond the traditional CCRIS data and salary data.”

Big data analytics driven models are also able to determine the repayment capability in the near future. This can be achieved by projecting the potential growth rate of individual earnings, analysing both credit data and complement with noncredit data such as transaction data, review ratings, followers population, growth rate, active follower engagement rate, etc.

This can be continuously enhanced with machine learning, as the more data it captures, the more predictable it becomes on its lending parameters, hence being able to contain the non-repayment risks. Nonetheless, it requires time to collect data points at the initial stage and acquire enough sample size to be statistically measurable for this to happen.

With big data analytics driven model equipped with automated data capturing process on the right data points coupled with machine learning capabilities, the time taken to achieve this will be shortened significantly.

To conclude, a resilient digital bank with sustainable earnings over the long term requires a strong tech capability with big data analytics coupled with artificial intelligence and machine learning capabilities to improve the model continuously.

And with the right business model and human capital to drive the bank, we strongly believe that Sarawak Digital Bank can eventually become a game changer to the banking industry, especially after the first 3 years of operations.