Incorporating AI — common sense updates to credit policy

How to navigate onboarding new technology and updating the policies your institution uses

‍

Having a solid credit policy in place at a financial institution often determines the organization’s financial success or failure.

A credit policy protects a lender’s business and oftentimes protects their borrowers, too. By utilizing a clear and consistent policy, lenders can ethically navigate loan decisions and even use their resources to help those high-risk borrowers develop better financial habits if they wish.

But what happens on the policy level when a lender onboards new technology, like AI-automated underwriting, and begins using it to help navigate lending decisions? Many lenders will hold fast to their credit policies or even add more policies on top of the new technology. However, research shows that this is rarely helpful, and in fact, harmful more often than not, particularly for protected-class borrowers.

Still, the question remains: Is there a perfect balance of policy and technology that can be established? And if so, how do we get there as an industry and as individual lenders with unique borrowers? Let’s look at an example. 
‍

AI helps make decisioning less binary and more insightful‍

Say you have a loan application come in from someone who just started a new job. At first glance, this may appear to be a higher-risk borrower. But using AI lets us take a step back, take in more information to guide our decision-making, and better predict risk.

With AI, we might see that despite this person changing jobs — or even job-hopping, which is more common among younger generations — they have a strong repayment history and steady income. 

While these aren’t the only data points that factor into a credit decision, it is an excellent example of how AI helps take decisions from a single variable to a multifaceted decision that mimics a quality argument for why this individual should get a loan. This is something that mechanically-applied credit policies and legacy scoring systems aren’t capable of.

If that same person were to come into a financial institution with a credit policy stating loans of a certain amount wouldn’t be granted to a person with less than a year’s employment at their current job, their application would be denied, and that financial institution would miss out on someone who likely could have been a great applicant.

Credit policies were put into place to protect lenders from making risky investments, and we cannot forget that history. We’re fortunate to have new technology and tools at our disposal that enhance a financial institution’s ability to predict and prevent risk. However, credit policies that prohibit lenders from acting on what this new technology is finding means that lenders are closing their eyes to what they’ve invested their resources in, which doesn’t move our industry forward and hinders better, fairer, and safer credit decisions.  

‍

The best time to plant a tree is 20 years ago; the second best time is today‍

Implementing AI doesn’t have to be an all-or-nothing approach. In fact, going all in on new technology can be risky in and of itself. This is why lenders should work with their technology partners to incrementally incorporate the technology into their institutions. You have to learn to crawl before you walk and walk before you run.

When clients integrate with Zest AI, we recommend a couple of updates for credit policies that will help enhance decisioning and increase the efficiency in their lending processes:

• Remove redundant policies – Identify and eliminate credit policies that overlap with features or concepts already factored by Zest AI’s model. 
• Relax thresholds (within reason) – After deciding the desired approvable borrower range with Zest AI’s technology, consider relaxing certain credit policy thresholds. This can increase instant approvals and improve overall efficiency at a financial institution.
• Adjust credit policy fields – Evaluate the impact of different policy threshold ranges on portfolio risk and use this data to adjust and optimize your lending strategy. With the insights provided by Zest AI’s technology, lenders can potentially increase automation in credit approvals without adding undue risk.
  

Using AI as a tool to enhance your lending practices can be both efficient and secure. Your technology partner should always provide your institution with an in-depth analysis of the work you’re doing together, and support changes that best align with your risk tolerance and lending goals. And if you’re not quite ready to move to the next step — whether that be walking, running, or even a sprint — a true technology partnership can be there to provide top-of-the-line service wherever you’re most comfortable.