New ways that loan providers can improve their underwriting

The process of underwriting refers specifically to when a finance or loan provider assesses an application and decides whether or not to approve the applicant for a loan. An individual, or a team of underwriters will assess eligibility criteria by looking at behavior and statistical analysis, as well as other kinds of checks in order to ensure the customer is able to afford to the financial product they have applied for.

Underwriting is key to FCA regulation, to ensure that lenders are upholding all the important checks prior to funding. But are there new ways in which loan providers can improve their underwriting process? We take a look.

Use of alternative data

Social networks

A small number of loan providers in the UK have started using other forms of alternative data when it comes to the underwriting process. One provider has started underwriting loans by looking at applicants social network associations.

For example, Oakam, the lender, will take a look at current borrowers networking data (such as referrals, as well as their geographic location) and analyze these against a new applicant as a way of seeing if they can identify patterns between the two that suggest if the new applicant shows creditworthiness, or a higher potential for fraud. It’s research has shown that this has helped the loan company to better determine the level of risk for someone with no credit, or poor credit history, as well as helping to provide customers with much better, and more personalized products overall.

Online chat conversations

Most loan providers these days have a website where applicants are able to pose questions live and directly to their customer service against prior to making an application, or during the process of doing so. Some loan providers have started to use these online chat conversations between agents and potential applicants as part of the underwriting process such as My Jar.

This is achieved by using natural language processing (NLP) as well as machine learning. The technologies can help to speed up the underwriting process dramatically, as they can speedy go through conversations and detect algorithms that may give a firm indication that some applicants may have a higher risk of defaulting at a later date, or are likely applying for a loan in order to commit fraud.


As part of the underwriting process, some are using gamification to enhance their accuracy of the behaviour of their customers, as well as influencing it too. For example, some provide customers with the option of earning ‘points’ on an app, due to things such as referrals or paying promptly, and then receiving real-time benefits for doing so, such as lower interest rates on loan. This behavioral data is used for its underwriting methods.

It may seem controversial to some, but these techniques are a much more robust and innovative way of underwriting that better serves applicants today. For example, with many traditional consumer lenders the underwriting process tends to look at data that is limited and difficult to analyze in real time, and have often only tended to serve segments of the public that they are able to conduct these risk assessments on.