Fannie Mae‘s latest Mortgage Lender Sentiment Survey points to a growing trend of lenders that want to make the mortgage process and its various components as efficient as possible. In 2025, lenders ranked streamlined business processes, reduced costs and consumer-facing technology as their leading priorities.
Every touch point in a mortgage’s life cycle — from sifting through years of servicing notes or thousands of emails to correcting missing or inaccurate data — creates inefficiencies and errors that could potentially inflate costs, heighten risks and erode investor yields.
To shed light on the importance of faster business processes — particularly in the face of rising costs and regulatory pressures — HousingWire spoke to Cade Thompson, chief growth officer for Rocktop Technologies, about how innovations in data management and AI-driven automation are helping lenders boost productivity.
This interview has been edited for length and clarity.
Sarah Wolak: One of the things that Fannie Mae’s recent survey talked about was streamlined business processes. Can you share your thoughts on why this is such a priority for lenders right now?
Cade Thompson: We cover three segments of our business: originations and servicing, the default sector and capital markets. Our perspective is that the holders of risk are the beneficiaries of lower costs and reduced risk, but it takes the whole ecosystem to participate.
One thing about the Fannie Mae survey is the shift in business priorities from cost cutting to streamlining. Cost cutting and streamlining are related, but cost cutting carries a negative connotation — does it mean cutting service, quality or jobs?
For years, there was resistance and even regulatory hesitation. But today, with higher inflation, rising rates and record-high costs, there’s pressure to deliver services more efficiently. Streamlining is the better moniker. Our perspective is that you can retain your labor force and vendors while making them more efficient.
From our experience, there’s just an allergy to the theme of ‘AI is going to take my job and rule the world.’ But if you didn’t have to copy and paste into spreadsheets all day, or fill out the same reports constantly, you’d be more fulfilled and we could deliver more efficiently. That applies across origination, underwriting, processing and servicing.
SW: Are the processes being sacrificed once the cost is cut, and is the quality being sacrificed? If so, how do you think that lenders can mitigate that?
CT: I think everything is around evaluation. For example, say we have 12 tasks, and we have 15 subtasks for a process, and we are relying on a human to do these things. If you really were to break that down, some of it’s binary; it’s objective.
In order to show how you are streamlining processes, you look at outcomes. What is the outcome that we’re measuring? Whether it’s cycle times, reduction in errors, better experience for your borrowers or your customers, those outcomes are really what should be the benchmark for whether it’s working or not working, whether it’s better or worse.
Cost is not always the right benchmark, and I think that’s what the industry is trying to finally figure out. You can go cut costs, lay off people, terminate vendors and spend less money, but that may be tripping over quarters to get to nickels.
SW: Jumping back to earlier, you talked about the ‘AI allergy,’ referring to people being nervous that streamlined business processes with the help of AI is going to get rid of their jobs. How do you assuage these fears as you actively implement automation?
CT: think it starts top down. There’s a culture and subculture in every organization, and it starts with executive leadership committing to the idea that ‘business as usual is no longer acceptable. We must transform. We must evolve or die. And in order to do that, we need buy-in from the whole organization.’ So that’s step one.
Step two would then be empowering your people, empowering the leaders of your organization to be co-authors in how this is done, versus saying, ‘Hey, procurement just found you a new tool. You’re going to implement it. Therefore, you have to cut 22 people.’
If the end result is ultimately an attrition of the workforce, then that might be necessary in cases, but for most companies that we deal with, it is a top-down approach where they’re saying, ‘We are not here to eliminate your job. We’re here to help you be more productive and have a more fulfilled work life to create work-life balance.’
SW: How do you manage expectations when you introduce a product to scale up productivity? Specifically, how are you managing the implementation and the expectations of how productive a worker should be when using a new tool?
CT: It first starts with unit economics. You have to know what your real measurement is today. What does it cost you to underwrite a loan? What is the real time it takes to underwrite a loan? How much error rate do you have in underwriting? If you don’t have a benchmark to start with, then everything just becomes relative and almost anecdotal or hyperbolic, if you will.
No. 2 is it becomes formulaic: If one person could do this task in this amount of time, and the task or the subtasks of what they’re doing are taking X amount of minutes, and there’s this error rate … then all of a sudden you can measure what ultimately would be some projection, or some forecast, of how to get to what is a condition of success.
We encourage all of our customers to start by asking, ‘Do you have measurements in place and KPIs?’ Take into consideration all the factors — not just payroll — to set the groundwork. And then we work backward from there to measure what’s an acceptable level of improvement.
SW: Can you discuss Rocktop’s role in introducing AI-enabled due-diligence and productivity tools?
CT: We’re evolving as market adoption comes along. Sometimes you can be right, or you can be dead right — which means you’re so right, but you’re so far ahead of market adoption that you’re just sitting there not able to do anything or create value.
We’re working with our customers on a journey. Most customers have a two- to three-year journey to transform their business. When you think about AI as one of the technologies around that, every single one of our customers has a different set of issues, even though most of them are on the same loan origination system or the same servicing system or the same accounting system. But they have implemented it differently or they use it differently.
Going in with the one-size-fits-all approach of deploying solutions is near impossible at this moment in time. It is not a commoditized widget yet. It will be at some point, but it’s not today.
And so what we have built is a tool belt — or as some of our folks call it, building blocks — of different AI tools. Those blocks include being able to do doc intelligence, index, bookmark, understand what a PDF is, extract data from that PDF, look at a photo to determine damage analysis, read unstructured text.
SW: What are common inefficiencies in the mortgage life cycle that Rocktop is helping to target? Are there specific offerings for specific pain points?
CT: On the origination side, the big pain points are obviously underwriting
processing, pre-closing QC and post-closing QC. Those are the three main areas where costs leak out of that transaction.
Then you go into servicing, and the servicing transfer process is a very challenging process, and it sets up the success of that loan or that portfolio for the remainder of its life. If it’s boarded wrong onto a platform, it’s got problems forever. It’s very hard to correct once you have borrowers online making payments and getting statements, and if the data is not right.
Anything in the default world is an issue. And then I would say anything around REO management is still very complex. So those are the core issues, thematically, that we see over and over again.
It’s easy to think about, well, why don’t we just build this product, bring it to the market and just sign up a ton of customers? It just doesn’t work like that. Where we can come to the table is [we can] integrate and ingratiate ourselves into their current processes without completely turning over and having a full transition to something completely new.
English