For years, banks and other traditional lenders have used credit scoring systems from the likes of Vantage and FICO to determine the creditworthiness of borrowers, usually with varying levels of success. While they’re often effective at weeding out potential defaulters, they are also notorious for locking out those who don’t use traditional banking services and about a quarter of the American population without a FICO score, even when such borrowers would have made good clients.
As a result, many fintech startups have come up to help bridge the gap. Many of these lenders use artificially intelligent tools to analyse troves of alternative data to help make credit decisions, which usually turn out to be more helpful for the unbanked compared with traditional scores.
AI and big data analytics have been disrupting industries across different niches for years now. From creating preventative maintenance plans in manufacturing plants to plotting market trends and customer preferences in e-commerce “ there are virtually no limits to the applicability of big data.
By capitalising on big data and AI-powered tools, alternative lenders have become significant players in the lending industry. A report by BI Intelligence estimates that alternative lenders will control 20.7 percent of the total small business lending market by 2020, up from the 4.3 percent share they had in 2015. This speaks volumes about the potential for lenders in this space, which will be interesting to see as techniques in big data analytics continue to evolve.
How Lending Startups are Leveraging Big Data and AI
Upstart is one of the many fintech companies that have specialised in improving the loan approval process using big data and AI. Upstart uses a peer-to-peer online lending system to make underwriting decisions for loan amounts between $1,000 and $50,000 within minutes. The company uses machine learning to analyse your employment history, the area of study in college, educational background, and other different types of alternative data to make credit decisions.
For people with limited credit histories, low-income earners, and young borrowers who are often hit with high-interest rates, Upstart offers one of the most beneficial, data-driven platforms for would-be borrowers.
SoFi, short for Social Finance, is another fintech startup that is using AI and big data to disrupt lending. The startup, which was recently valued at over $4 billion, unleashes AI-powered algorithms on alternative data to help lenders find patterns that would otherwise be overlooked.
Some of the data that SoFi uses includes educational and professional history, cash flow, and the borrower’s history of timely bill payments. The lender uses an aggregate of this data to disburse personal loans, mortgages, student loans, and mortgage refinancing in addition to traditional credit scores, which often results in a better scoring model.
Another fintech startup, Avant, is using machine learning and artificially intelligent algorithms to analyse thousands of data points to help borrowers with subprime credit scores access unsecured loans of between $1,000 and $35,000. Avant’s AI-powered algorithms help the startup to analyse over 10,000 data points associated with potential borrowers with low FICO scores, which helps identify people who would be good borrowers but are locked out using the traditional scoring system.
Avant is also going spreading its AI-powered tentacles further into fraud detection. The lender uses machine learning to compare customer data and behaviour against a baseline, which helps identify patterns and single out outliers. For instance, some of the algorithms look at how long a customer spends reading through contracts, application questions, and even time spent on pricing options to determine a normal curve.
Avant is also among a growing list of alternative lenders that are looking to integrate their predictive and data analytics systems with traditional banks. They are looking to join the likes of Juvo, First Access, and eCredable that are already working with banks, mobile network operators, and other third-parties to provide alternative credit rating systems for the unbanked.
Challenges
Even with all the advances in AI and big data analytics, there’s still a long way to go before machines can dispense loans without human intervention. There are tons of security and confidentiality issues to address before many of us can be comfortable with the technology.
As an example, many lenders require you to download and install apps that collect personal data and mobile usage habits on your mobile device. Too much data and information on one platform can be risky for privacy – if the recent Equifax breach is anything to go by. Additionally, information security is further complicated by advances in other technologies, including trends in VPN technology and the Internet of Things.
There’s also the real possibility of bias when algorithms are allowed to make credit decisions, which can complicate alternative credit scoring platforms.
Still, pundits and observers in this area are convinced big data and AI will play a vital role in lending for both digital and brick-and-mortar lenders over the next decade. With the current rate of tech advances, it’s not difficult to see why.
