The Wilqo Way

Machine Learning Business Case: Fallout Probability

Written by Wilqo | Oct 3, 2024 6:15:54 PM

Is AI + Data the solution to all your problems? Can we all retire now?

Unfortunately, not yet.

But Wilqo does have an AI + Data solution that will help your operations… today.

We want to help you be smarter about your book of business. As a lender you must balance intuition, probability, insight, and ultimately action. In this post, we focus on how Charlie, our Production Optimization Platform, uses Machine Learning (a subset of AI) to be safely and securely put to work for you.

Why Probability Matters

Lending is inherently a game of probability, and there is a cost associated.

In an ideal world every legitimate loan would close. When a loan enters your pipeline, even one that should get approved, there are multiple factors that could still lead to that loan falling out. Different factors will influence the cost of that fallout.

The borrower could change their mind, market conditions shift, or unexpected underwriting issues arise. Predicting this fallout is critical for lenders who want to keep their pipeline healthy, reduce operational inefficiencies, and ethically maximize pricing strategy.

If you knew halfway through the processing of a loan that it was 85% likely to fallout, would you double-down on your work effort to keep it alive, or find a graceful way to exit the transaction and save your time and money?

Machine learning (ML) gives us a way to evaluate and assign probabilities to various outcomes. With Charlie, we move beyond mere data collection to actionable predictions. This helps lenders foresee which loans are most at risk of falling out before it happens. By knowing the probability of fallout, lenders can prioritize loans, refocus resources, and make informed decisions about which loans need more attention.

 

Better Data, Better Predictions

Not all data is created equal, and that’s where Charlie shines. Current origination systems aren’t focused on the granular nuances of each loan. Charlie’s model, however, creates granular, detailed data unique to each loan in your pipeline. This level of specificity allows our Machine Learning algorithms to deliver more accurate predictions.

For example: Charlie’s borrower level activities.

In Charlie, activities can be assigned to borrowers to provide information, fulfill a condition, etc. Examples include:

  • Provide two most recent paystubs
  • Acknowledge Intent to Proceed
  • Provide a valid photo ID

Does your current system track how long it takes for these items to be provided?

Charlie’s model tracks all borrower activities to expose a level of never seen before insight. How long is it taking to receive paystubs? How long has it been since the Initial Loan Estimate was sent and you are waiting on Intent to Proceed? Perhaps time to provide a valid photo ID is a leading indicator of fallout.

Charlie’s unique method of breaking down the loan process into activities unlocks this opportunity.

 

A Solution for All Lenders: Whether You're a Data Pro or New to the Game

Some lenders already operate with sophisticated tools and strategies in place. For these power users, Charlie’s detailed data and predictions will only make you smarter.

For others who may be newer to leveraging data insights, Charlie simplifies the process. Our model assigns loan level probabilities as well as a prediction of your outstanding pipeline. As loans progress through their lifecycle, and more data is obtained, probabilities will automatically adjust.

The more loans entered into Charlie, the smarter your model becomes, and the more informed you become.

 

AI vs. Machine Learning: Understanding the Difference

Before we wrap up, it’s important to clarify the distinction between AI and Machine Learning. AI is the broader concept of machines mimicking human intelligence, while Machine Learning is a specific subset of AI that allows systems to learn from data and improve over time using algorithms and statistical models.

Charlie leverages Machine Learning to get smarter with every data point and every loan that moves through the system. Over time, the model adapts, continually refining its predictions for fallout and other pipeline risks.

 

Ready to Harness the Power of Charlie?

AI and Machine Learning won’t solve all your mortgage data and hedging problems, but it can certainly give you a clearer view of your pipeline and help you take proactive steps to mitigate the costs of fallout. Charlie’s unique combination of granular data and powerful predictive models is designed to help lenders stay ahead of the curve.

Curious about what else Charlie can do for your lending operation? Book a demo at Wilqo.com/MoreInfo