Innovation in Lending

AI Lending: What to Know Before Buying

Ken Garcia

April 26, 2021

As Ken Meyer, chief information officer at Truist Financial in Charlotte, N.C. highlighted in a recent interview: Banks can be ‘fast followers’ of AI. And as many recent reports indicate, AI adoption in financial services is accelerating. Viewed as the key to building a competitive advantage, growing portfolios safely post-COVID, and solving industry challenges like more inclusive lending, banks and credit unions are considering AI in credit risk underwriting. With the business case justified, leaders are focused on picking the right approach. 

Financial institutions recognize the advantages of buying an AI solution for lending, including faster time to market and lower operational costs. The big “a-ha” financial institutions discover when pursuing the buy route is that the marketplace offers two very distinct options: software vendors that offer pre-built models that can’t be customized and “software partner” vendors that offer tailored off-the shelf models and help you through the model lifecycle. So, what should you consider before buying an AI lending solution? Let’s dig in. 

Leveraging Domain Expertise to Ensure Success 

A key advantage with the software partner approach is that you get to leverage their domain expertise to avoid the model validation missteps and streamline model risk management (MRM) processes to address the specific needs of stakeholders (i.e. risk, data science, regulatory, IT, legal, and business functions). With an industry partner, you avoid talent shortage issues, can more easily upskill your workforce, and have access to industry-first features. 

One of the most overlooked factors lenders miss is the expertise and resources needed to get a model through compliance and regulatory review, integrate it with existing LOS, monitor it for ongoing optimization, and plan for continuous development to keep up with the breakneck pace of innovation. These overlooked factors contribute to the high percentage of AI project failure rates. A recent Gartner report revealed that companies with artificial intelligence experience moved just 53% of their AI proof of concepts into production.

Key Consideration: Pre-Built Model Performance Gap

As you weigh the pros and cons for a buy approach, another major consideration is the performance difference between pre-built and “software partner” model like Zest. A generic model ignores the specifics of your organization's customer base, economy, and products. As a result, they typically underperform tailored models. For example, our model helped a lender reduce risk by 24.9% compared with their pre-built models, resulting in economic savings of $40M over the pre-built model.  

Download the Guide

There are many factors to consider in this process but accounting for the blindspots and putting a dollar value to the costs of getting it wrong or right will help provide the clarity needed to make the best decision for your organization.

In most situations, working with a reputable and trusted partner to deploy a proven ml model will allow your organization to capture the ROI more quickly and generally ensures that the model will have more comprehensive support throughout its lifecycle.
To get the full list of pros/cons and major considerations for each approach download our latest guide: AI Approaches to Lending: Should you Build, Buy or Rent?

Read the Guide
Thank you for subscribing!
Something went wrong while submitting the form.