Data enrichment is the secret sauce in digital lending

In describing financial technology, it’s easy to forget that “what it does”  is more important than “how it works” for most people. With that in mind, we can promise that this will be an easy read to help you understand what data and data enrichment do in digital lending, and how they make life easier for you, the owner or executive of a business that wants to extend credit to its customers as efficiently as possible.

At TurnKey Lender,  we start with an interface that’s actually a pleasure to work within. You immediately notice a layout that makes sense, with things exactly where they should be. In this environment, you feel supported, a sensation that quickly translates into confidence and better performance. That’s the work of the industry’s best designers, analysts, and developers, all focused on delivering a superior experience to the business, its employees, and its customers. 

Your data needs enrichment 

“Out of view, beneath user-friendly UX, run rivers of data, performing complicated credit calculations, weighing industry-standard ‘if/then’ scenarios, and carrying accumulated insight from automated task to automated task,” says Elena Ionenko, COO and co-founder of TurnKey Lender. “But the data can’t do its job without enrichment.”

Data enrichment — sometimes data “augmentation” — is the “process of enhancing existing information by supplementing missing or incomplete data,” according to ScienceDirect. Typically, this enrichment is achieved by using external data sources and internal artificial intelligence and machine learning to make (and learn to make better) decisions using available information and a growing array of best options.

Machine learning is a subset of AI, a device or process that takes into account aspects of its environment to make decisions or predictions, mimicking human cognition. Machine learning supports AI by using algorithms and statistical models to perform many specific if-this-then-that type tasks at once, drawing on patterns and inferences rather than explicit case-by-case instructions. That is, machine learning takes “training data” and constructs mathematical models that bring “thinking” to nuanced processes requiring multiple inputs from vast datasets such as determining a would-be borrower’s suitability for a loan of a specific size, type, and duration.

A use case (and more definitions)

Another term that crops up in conversations about AI is “deep neural network.” This sits at the intersection of multiple incoming datasets and the decisions or inferences to apply the most relevant formula for the task at hand.

Imagine, for example, that the input in question is a gallery of photographs depicting every tree species known to man — around 73,000, as it happens, an easy lift for TurnKey Lender’s AI. Now let’s suppose the task is to grade timber for its suitability in making a particular type of handcrafted furniture, which calls for hardness, high polishability, and a dark natural color. The neural network might start by distinguishing hardwood from softwood trees (which may be eliminated from consideration) before sorting for other suitable characteristics. That’s what makes these brainy systems so “deep” in the first place, and makes them so valuable as an aid to accuracy.

So it’s no surprise that the need for data enrichment is going through the roof. The volume of “big” data in data-center storage went from 25 exabytes in 2015 to 403 exabytes in 2021. One exabyte is equal to a million trillion — 1,000,000,000,000,000,000 — bytes.

45 million Americans have little or no credit history

This sort of volume starts to make sense when you understand that determining creditworthiness is a complicated process calling for millions of data points to be assembled, sifted, read, understood, and double-checked to make any one of the myriad decisions required to measure creditworthiness. And these are all tasks that used to be attempted manually, keeping teams of underwriters, loan managers, reporting officers, unit managers, and business owners extremely busy. Now, most credit checks are performed automatically, requiring only high-level supervision.

As a result of automated AI, organizations that deploy data enrichment can:

• Reduce risks
• Eliminate human error
• Lower operational costs 
• Grow customer lifetime value 

Meanwhile, every business has a significant data layer that is rarely used to its full potential, an oversight that can block their access to some 45 million US adult consumers with “thin or nonexistent credit history,” according to the New York Times.

New ways to view loan applicants

Meanwhile, your business throws off reams of proprietary information — typically transactional data, telecom data, customer histories, etc. — that can be used in credit scoring, woven into application forms, or used to measure customer responsiveness and gauge customer preferences. In point of fact, many businesses overlook the fact that data on the habits and behaviors of existing customers can provide insight into a loan applicant’s creditworthiness, augmenting the data a traditional bank can get hold of. 

Tip of the iceberg 

More concretely, traditional data points include:

• Credit scores
• Biographical data
• Fraud prevention
• Financials
• Firmographics (employer status)
• Behavioral traits 
• Transaction history
• Collection records

Alternative scoring data, meanwhile, uncover:

• Mobile device usage and mobile network
• E-commerce purchases, trends
• Social networks 
• Psychometrics
• Spending and bill-pay habits
• Skip Tracing
• Mileage/driving habits
• Education (down to grades, majors, degrees, and of course institutions)

Any data that can shed light on a loan applicant’s ability to repay is fair game if it can be digitalized. Even when so formatted, however, data in such quantities can’t be handled by any one human — or even a whole squad of them. That’s where automated AI-driven scoring and decisioning come in to gather, sort, and manage large data flows for use in guiding credit decisions and other functionality rather than letting them stew unused in vast, opaque databases.

“Without enrichment capabilities, the vast arrays of data we can access nowadays might as well be buried in the desert,” says TurnKey Lender’s Ionenko. “Our deep background in AI and data enrichment gives lenders a better picture of loan applicants than they’ve ever had before, and the means to service accounts and manage whole loan portfolios in ways that make low-risk lending more accessible to consumers than ever.”