Assisted and Automated Secured and Unsecured Lending

Project Reference :

AISG- 2018-100E-001

Institution :

National University of Singapore (NUS)

Principal Investigator :

Dr Phan Tuan Quang

Technology Readiness :

7 (System prototype demonstration in operational environment)

Technology Categories :

AI- Machine Learning

Background/Problem Statement

The global credit rating software market size was valued at $0.42 billion in 2020, and is projected to reach $1.92 billion by 2030, growing at a CAGR of 16.5% from 2021 to 2030. Credit rating involves validating the credibility of individuals and companies based on their prior transactional and credit behaviour.

A credit score is a number used by lenders as an indicator of how likely an individual is to repay his debts and the probability of going into default. It is an independent assessment of the individual’s risk as a credit applicant. A secured loan, also called a collateral loan, is a form of credit that is backed by an asset. By putting forward an asset to act as collateral, it implies that the lender is able to cut down the risk involved in lending cash to a borrower.

Credit scoring for secured and unsecured lending is a highly labor-intensive process with many points for human errors.  These errors can be highly costly and result in inefficiencies in the lending industry.  As a result, it is a highly risky business with high-interest rates.


The solution is a credit scoring method for heterogeneous data sources and Android mobile app for secured and unsecured lending. It works as follows:

  1. The borrower has to download the app to his phone and accept the requested permissions to his/her mobile phone data.  
  2. The borrower is then asked to verify his/her identity and provide other relevant information.  Combined with traditional credit scoring data from existing sources, the borrowers’ phone data including call logs, SMS, web browsing behaviour, geolocations, and other data from the phone will be sent to a cloud service to perform advanced machine learning to predict the borrower’s creditworthiness.  
  3. Within a short period of time, the borrower will be offered a product with the requested loan amount, interest rates, and payment schedule.  
  4. The borrower will then be directed to a money lending outlet to process the loan and collect the cash.


The solution aims to improve efficiency in the ecosystem by providing higher accuracy to the lenders, reduce the barriers to applying for a loan for borrowers, reduce risk from lending institutions by using artificial intelligence to perform better credit scoring based on mobile data, provide more touchpoints and engagements to help customers manage their personal finances, and thus reduce costly errors in the process.

The solution helps to: 

  • make it easier for customers to apply for a loan
  • enhance productivity and reduce manpower requirements for the current labor-intensive and manual credit scoring process
  • reduce costly and error-prone manual credit scoring by providing better credit scoring algorithms and methods
  • increase regulatory compliance

Additional revisions may be required in order to make the solution compatible and compliant with any change in regulations imposed by the government before it is deployed.

Potential Application(s)

The AI-assisted credit scoring software helps banks & financial service providers assess credit risks precisely and make informed decisions in credit origination and loan monitoring. The credit scoring software helps in faster approval of loans for consumers and provides instant results about the credit score of a consumer.

We welcome interest from the industry for collaboration/ co-development / customisation of the technology into a new product or service. If you have any enquiries or are keen to collaborate, please contact us.