Scaling lending operations with AI-assisted data validation

For mortgage lenders, the underwriting process can be one of the most time-consuming and resource-intensive parts of the business. Manual document review, data validation across dozens of sources and the constant pressure to maintain loan quality while keeping pace with volume leave many teams stretched thin.

Kelly Millard, Director of Underwriting at Equity Prime Mortgage (EPM), knows this well, and has spent the past several years systematically rethinking how his team works. "Turn time, loan quality and keeping up with overall volume. Those are the things that matter," he says. "What I love about EPM is that we are focused on embracing the efficiencies that AI and automation can bring to help us deliver on those priorities.”

As one of the nation's leading wholesale mortgage lenders, EPM is focused on building a team that can scale with volume without sacrificing loan quality. Kelly is leading that charge by putting tools in place to offload the repetitive manual work that used to consume much of his underwriters' time. To do that, he has embedded the ICE Mortgage Analyzers into EPM's underwriting workflow within Encompass®. The Analyzers use AI to automate the most time-consuming aspects of validating borrower income, credit and asset data, improving the accuracy and consistency of the information used for underwriting. That means his team can spend less time on manual “stare and compare” review tasks and more time on the critical analysis that actually moves a loan forward.

Beyond managing costs, EPM leadership wanted to design a better way of working for their staff and for the broker partners they serve. Their innovation journey started with a question: what would it look like to build an underwriting operation that could scale to support growth plans, without adding additional headcount?

Answering that question started with an honest look at how EPM underwriters’ time was actually being used. A significant portion of every review was consumed by what Kelly calls 'ticky tacky' tasks: cross-referencing names, Social Security numbers, dates of birth and other data points across dozens of documents. Left uncaught, these mismatches can send a loan to AUS with incorrect borrower data and return incorrect eligibility information for Fannie Mae or Freddie Mac, impacting the loan’s salability. While this work is necessary for every loan, it didn’t represent how EPM wanted the majority of their employees’ time to be spent. The path forward became clear. If automation could handle the data validation, EPM’s underwriters could focus on the complexity and risk assessments that actually move the needle. And that’s when the leadership team decided to move forward with adopting ICE’s Mortgage Analyzers solution to help make this vision a reality.

Sustainable change doesn't happen overnight. EPM invested more than a year in workflow analysis, planning and testing before fully rolling out the Analyzers, configuring the tools to reflect EPM’s business needs and underwriting workflow goals. They then used the platform’s configuration options to define precisely what should be flagged and what should not, so that when underwriters relied on the Analyzers, they could trust the results.

Central to the success of the rollout was EPM's internal tech team. They translated the company’s vision for a new underwriting process into a tailored workflow within the Analyzers. The result was measurable operational gains with minimal disruption. "Our tech team took the recommendations that ICE provided based on the challenges we wanted to address and made the change relatively seamless for our underwriters,” Kelly explains. That partnership between underwriting knowledge and technical execution was key to making the rollout stick.

One advantage that made the transition to the new underwriting workflow easier for internal teams was the direct connection between the Analyzers and Encompass. "We explored other solutions and some worked okay, but there isn't the cohesiveness with Encompass," Kelly explains. "The fact that the Analyzers are an ICE product means the income calculations and data changes feed back into the loan file." That connection keeps underwriters focused on the review rather than the administration without ever having to leave Encompass.

With the foundation in place, the positive impact to EPM underwriters’ time was almost immediate. After adopting the ICE Mortgage Analyzers, borrower names, Social Security numbers, dates of birth and income calculations are all validated automatically, before an underwriter has even opened the file. "When I'm training someone new, I point them to the 60 rules the Analyzers have already checked and verified as soon as the loan is submitted for review," Kelly explains. "That's where the efficiency comes from. With the tool handling the busy work, your energy can go toward decisions that actually require your judgment and skill."

The Analyzers don't just handle the routine. They also add an extra layer of scrutiny that even the most experienced underwriters benefit from. In one case, the Analyzers flagged a Social Security tax withholding percentage on a paystub that exceeded the allowable tolerance, a discrepancy that revealed the document was fraudulent. It's the kind of detail that's easy to miss in a manual review and exactly the kind of catch that protects EPM, their broker partners and the borrowers they serve.

This significantly improved underwriting workflow is just the beginning. EPM is already planning to extend the technology further upstream in the lending cycle to accelerate their pre-approval process. In the next phase of their adoption plan, EPM is planning to leverage the Analyzers to automatically review new loans submitted by brokers and issue pre-approvals with minimal involvement of an underwriter. Once outstanding documents come back, the Analyzers will pick them up, feed the data directly into the loan file for the underwriter to review and issue the final approval. The EPM leadership team believes this model is particularly well suited for the company’s wholesale focus and the natural next step for leveraging technology to scale the organization.

The results EPM has achieved since fully embracing the Analyzers reflect what happens when the right technology is implemented with intention. Underwriting throughput has more than doubled, from an average of around 2 loans per underwriter per day to over 5 per day. Turn times have dropped from five to six days for initial review down to 24 hours, with resubmission and final conditions following the same trajectory.

For Kelly, those numbers are less about outpacing the competition and more about what they prove is possible. "Many of our competitors’ turn times are at two to three days for initial review," he says. "We are often turning files in 24 hours and doing it at a volume that would have been unthinkable before these tools were implemented." The competitive advantage that EPM has created is a reflection of what is possible when a team is willing to set a vision, rethink how work is done and embrace innovation to get there.

EPM’s increase in loan volume is another testament to this approach. Since adopting the Analyzers, EPM has seen record-breaking broker loan submissions and some of the biggest production months in the company's history. "What sets us apart now is that we can do more loans than we've done most months throughout the history of our company and we did it by working smarter."

For lenders curious about getting started, Kelly's advice is straightforward: test it, invest time in the training and build it out in a way that fits your specific business model. The Analyzers are configurable, and what works for EPM may differ from what another lender prioritizes. "Take the time to listen, go through the training and then build it out how it works for you," Kelly says. "There were things we liked that others might not, and we adjusted accordingly to make it work for our business model."

But behind the technology and the configuration work lies a bigger lesson. With the help of ICE Mortgage Analyzers, EPM was able to reinvent their underwriting operation. Underwriters now spend their time on work that actually requires their expertise, turn times are measured in hours rather than days and the capacity to grow is built into the workflow itself. For other lenders, this story is a reminder that the right tools, implemented with intention, can fundamentally change what a team is capable of.

EPM's story is a practical example of how lenders can use artificially intelligent technology to improve speed, quality and capacity and give their best people the space to do their best work.