Your company’s lending program is only as good as its credit scoring. More than any other non-macro factor, how — and how deeply — your company assesses prospective borrowers will determine the size and profitability of its lending program.
This is where artificial intelligence comes in as an adjunct, and a catalyst, to alternative scoring. To explain how this works, let’s step back for a moment and look at traditional scoring methods and their underpinnings.
Character versus credit score
In fact, there are at least two “traditions” in play here.
Before the 1990s, assessing a retail loan involved a bank-based loan officer determining the applicant’s creditworthiness based on inputs such as income, indebtedness, prior-credit outcomes, bank balances, work history — and, to no small extent, “character.” Here character would be determined by a range of assessments from the quasi concrete (like the applicant’s prior dealings with the lender) to the relatively ephemeral (like the “standing” in the community).
In other words, old-school inputs might suggest an applicant is good for the loan she sought in terms of her financial and employment standing, but the loan officer may not like something else about her. Maybe the trigger for rejecting her application, or imposing a higher interest rate, was a particular character fault — the applicant’s notoriety as a boozer, for example — or maybe it was something frankly discriminatory such as her race, name, neighborhood, accent, or appearance.
By the late 1980s, however, lenders had hit on the concept of credit scores as a transparent alternative to making subjective value judgments about credit applicants, both in the name of fair lending and operational efficiency. To meet this nascent demand, FICO began compiling information on the financial behavior of individual consumers as determined by credit bureaus such as Experian, EquiFax, and TransUnion, and selling these scores to lenders. In time, the big three credit bureaus rolled out rival gauges to the FICO score, but the original remains the industry standard.
A new world of credit scoring
But FICO and other scores based on credit-bureau data form an incomplete picture of a loan applicant. Borrower traits such as line-item spending habits, social-media comportment, and behavioral “tells” regarding financial obligations shed light in new corners.
Along these lines, “psychometrics,” a branch of psychology for evaluating a person’s state of mind, is a fascinating new contributor to credit scoring. Applied to lending — and as an illustration of how granular alternative scoring can get — psychometrics show up in borrower traits such as:
- Familial relations
- Money management and financial planning skills
- Risk aversion
- Organizational skills
- Mental agility
Coupled with additional insights into the applicant’s education level and personal savings and investments, these new data give lenders a sharper sense of the applicant’s ability to repay, and — writ large — a base on which to construct more durable loan portfolios.
The difference between a deluge and a credit score
Making sense of new inputs for credit scoring isn’t easy, however. Aside from raw computing power, it requires sophisticated artificial intelligence (in turn animated by dynamic machine learning) to rationalize an array of data from different sources to form a cohesive and legible internal credit-scoring system.
“With an applicant’s permission, our technology leverages information taken from bank accounts, retirement holdings and other transaction-rich sources to divulge spending habits, and monitor employment and non-employment income for a clearer picture of the would-be borrower,” says Elena Ionenko, a co-founder of lending-technology maker TurnKey Lender. “And because our perspective is global, we know that arming lenders with rationalized alternative scoring inputs equips them to thrive, even in markets where traditional scoring is either rudimentary or nonexistent.”
For instance, adds Ionenko, “Alternative scoring uncover hidden risks such as an applicant’s gambling expenditures or overdraft durations, and use them to make informed credit decisions.”
TurnKey Lender’s credit-scoring functionality, which combines traditional and alternative approaches, allows for dynamic customization. As a result, it answers the credit-assessment needs of mainstream lenders — such as banks, credit unions and finance companies — as well as new entrants. This growing list includes retailers of all sizes and types, medical practices, governmental agencies, NGOs, and invoice financiers.
Alternative credit scoring as a contributor to recovery
Now, with Covid-19 pressing on economies, the importance of credit as a tool for rebuilding has come into sharp relief — along with the need for alternative-scoring approaches for lenders.
In the US right now, lending activity is centered on businesses, largely with a view to maintaining worker headcount by means of the Small Business Administration’s “Paycheck Protection Program” and other initiatives run by or for the federal government. Uniquely, government-backed PPP lending is triggered not by creditworthiness, but a combination of need and the borrower’s ability to find a traditional lender willing to administer the low-interest loan.
Glaringly, however, there are no measures in place to speed the wheels of consumer lending — even though a full economic recovery is unlikely without such a kickstart. After all, the pandemic has eroded the credit standing of many who have lost jobs or seen wages cut. Getting money to borrowers in need these days will call for new ways of evaluating the creditworthiness of consumers.
“In these times, there is no way to exaggerate the importance of secure, reliable and intelligent scoring for lenders of all kinds,” says TurnKey Lender’s Ionenko. “The importance and utility of traditional credit-scoring is unassailable, but there is an acute need now for alternative approaches that lead to better decisions and, in turn, support more robust portfolios.”
How TurnKey Lender factors in
Since the founding moment, TurnKey Lender team understood that in order to transform the lending industry in a meaningful way, we need to apply AI and Big Data to credit processing and risk assessment. That is why the core team included a department of PhDs and Artificial Intelligence experts that have done successful scoring projects for Fortune 500 companies. And while many traditional lenders were hesitant about making the jump to the new type of intelligent credit decisioning before the 2020 crisis, now the need to adapt to the borrowers’ demands is obvious and even unavoidable if they want to stay in business.
To achieve an unmatched level of credit decisioning accuracy, we use deep neural networks with self-learning scoring models based on both traditional and alternative evaluation approaches and data sources. Working with the clients’ data, the system learns to use prediction, classification, clustering, and association to process loan applications.
For safety purposes, the system doesn’t just use the data the client is providing but also pulls the available information from the sources it’s synchronized with (like the credit bureaus, bank statement providers, borrower’s smartphone, or social media). All the data is processed by TurnKey Lender’s intelligent Decision Engine and is then presented in the form of a risk evaluation.
All in all, risk assessment, borrower evaluation, and credit decisioning take our solution as little as 30 seconds. To draw a comparison, most traditional lenders can take up to 9 business days to complete this process.
Let TurnKey Lender’s intelligent software analyze millions of data points in a matter of seconds while you grow your business.
But AI and Big Data, no matter how advanced and sophisticated, don’t bring much to the table unless they process the right borrower data. TurnKey Lender team analyzed the operations and credit decisioning flows of our clients in 50+ countries and dozens of business verticals to come up with optimal scoring models and decision rules that help weed out unreliable borrowers and pre-approve the loans that are most likely to be paid off on time. Not to mention, the dynamic selection and assignment of fitting credit terms to each application.
With all the heavy lifting done by the system, the employees’ time and energy are freed up to focus on business development and customer relations which translates into better performance.
TurnKey Lender Decision Engine comes with a proprietary AI-driven scoring model and decision rules built-in. Both are fully configurable, allowing you to set the factors you want to evaluate and assign values to them, based on what is more important for your operation. For cases when there’s no sufficient data available to evaluate borrowers, you can use a dedicated TurnKey Lender Psychometrics app that evaluates the borrower’s smartphone data and behavior patterns.
At the same time, to make sure we leave no stone unturned, TurnKey Lender provides lenders with an option to also collect and process traditional data like credit bureau reports, bank account statements, and payment data.
Data from traditional and alternative sources goes through the TurnKey Lender Decision Engine which in turn analyzes it with machine learning algorithms and deep neural networks. Pre-configured integrations with 75+ relevant data sources and technology providers allow lenders to put their entire decisioning workflow on autopilot and focus on big picture things.
Interested in learning more? Schedule a call with one of our experts today!