When billions of dollars are at stake, even a “small” or “incremental” improvement using AI and behavioral analytics can potentially save a bank or financial institutions millions in the long run. Which is  why AI application to improve loan origination process is a hot bed of innovation, from both established enterprise solutions providers like JurisTech and for up and coming tech start-ups trying to break into the financial services industry.

Let’s explore a few use-cases where AI can make a noticeable impact in loan originations:

1. Credit-worthiness

While FICO and CTOS scores are a good measure of people and organisations with significant credit history, there are also millions of people in both developing and developed parts of the world who are largely unbanked. ASEAN countries like Malaysia, Indonesia, Thailand and Philippines still have massive populations who have little or no credit history.

What they do have, is high levels of internet penetration, and social media accounts. Not to mention the wealth of data generated by their smartphones. These alternative sources of data can be eerily accurate while gauging an unbanked person’s credit-worthiness.

However, there is one big problem with this use-case. Banks and FIs are not exactly comfortable in sharing data with Facebook, or have people log-in to their apps using their Facebook accounts. In countries like China, where the government owns the dominant social media, it is viable to use social media date to determine a person’s credit-worthiness. A start-up named ZestFinance is doing just that. They are partnering with China’s biggest search engine, “Baidu” to use search and location data to judge credit-worthiness of each individual customer.

But in most other countries, gathering social media data of your customers can prove to be problematic. Banks are generally not comfortable sharing data with Facebook or using Facebook accounts as a log-in mechanism. And they have good reason to be wary, since Facebook is notorious for security breach and privacy issues.

If banks and FIs can bridge the gap between social media and their own CRM, AI can create a significant impact in the loan origination process by doing risk-assessment and judging credit-worthiness of an unbanked person before giving out loans.

2. Personalized Offerings

When marketers try to launch products in any industry, they come up with broad segments, usually clustered by age group and income level. Which is a reasonable approach because it is impossible for human beings to customize products and services to every single customer. The costs hiring an army of marketers would outweigh the benefits.

But with AI, one can tailor loan offerings to each individual customer based on their transaction behavior. In fact, if the likelihood of defaulting is extremely low, the AI can proactively reach an individual customer with a tailormade pre-approved loan offering.

A simple example: Reuben is 37-year-old in the fintech industry, which puts him at peak earning point of his life. Reuben also has a degree in Computer Science from a top-ranking US university. Based on what Reuben earns, and his credit card transaction behavior, his educational background and the places he spends money (gourmet restaurants, swanky bars, and fundraisers), our AI can safely assume that chances of Reuben defaulting is extremely low. Based on Reuben’s transaction behavior, occupation and age, the system should be able to predict that Reuben qualifies for a personal loan, a car loan, and a mortgage.

But which one is he most likely to buy?

Well, based on historic data, the system sees that married people in Reuben’s age-range will usually take a family vacation around June-July or around December. The system also realizes that people who take personal loans for vacation are usually married and are looking to spend quality time with family.

So, Reuben is automatically approached with a customized offering for a personal loan, with a headline that says something like,

“A Once in a Lifetime Opportunity to visit the Happiest Place on Earth: Take your family to Disneyland! X% interest for 12 months! This offer ONLY applies to you.”

The value of X can also be determined by the system. X will be low enough to entice Reuben, but also high enough so that it is feasible for banks. Similarly, car loans, housing loans and personal loans can be offered to people with the right message, at the right time.

3. Customer On-boarding

If you have a customer on-boarding platform like Juris Access, it can do wonders coupled with an AI. Paired with your CRM information and by talking to Google Ads via their API, it can automate the customer on-boarding process starting from the very first stage, digital marketing to the final on-boarding stage during which an AI can assist the customer every step of the way.

This use-case might sound a bit far-fetched, but we’ve tested it out in a real-life scenario.

Imagine a small business owner is trying to get a business loan. But while filling up his information he is confused. He doesn’t know which category he falls under: “Professional Bodies”, “Limited Liability Partnership”, “Sole Proprietorship”, “Partnership”.

If he waits more than the normal amount of time required to fill up the form. Or even hovers over the section “What is your company type?” for longer than necessary, a helper-bot can pop up proactively with a helpful suggestion:

“Hi there! Looks like you are having trouble with the type of company you’re applying for. Why don’t you answer a few questions for me then I can tell you what’s your company type!”

Then the helper-bot proceeds with a few simple questions and figures out that the company type actually falls under “Professional Bodies” and the applicant and move forward with his loan application. Reducing a lot of confusion, paperwork and processing time for all parties involved.