How Cenlar's Josh Reicher built a career in digital servicing

February 20, 2024 1:16 PM

Mortgage servicing technology is coming into its own, according to Josh Reicher, chief digital officer of Cenlar.

As a technology expert who crossed over into mortgages early in his career and works for a subservicer, Reicher has a unique view of how the industry's automation compares to the broader landscape.

Servicing has lagged in digital development compared to the origination side of the business, but as adoption of technology has broadly accelerated, it's catching up, he said.

In the interview that follows, Reicher offers details on how he started his career and the broader landscape he works within, which is being transformed by developments like advances in artificial intelligence, vendor consolidation and cyber security.

Reicher's comments on these topics below are in response to questions from this publication, both of which have been edited for clarity and length.

How did you come to be chief digital officer at Cenlar?

The role was created around something I was already doing for Cenlar. It started with the creation of our first overarching digital strategy. Before that, we had digital services, but we really hadn't consolidated them into a long-term, forward-looking strategy for our borrowers, clients and employees. So the idea was to create an overall strategy that would help us to understand the value of build versus buy, prioritize digital solutions and come up with an overall roadmap and timeline. The successful completion of that led to creating a chief digital officer role and naming yours truly to be in the spot. What I was most excited about with that, other than the fact that I got the role, was that Cenlar felt that that title was important enough to tell the industry that it was a digitally-focused servicing organization.

The biggest part of what it is that we needed to put in place at the time was improved borrower experience. We did not have a mobile app. We did not have a chatbot service that would connect to a live chat. We didn't have an optimal email and text solution. So there were a lot of pieces of the borrower experience that were kind of almost like a Frankenstein's monster set of solutions that had been put together organically over time. It was no one's fault. It resulted from a "let's add in live chat, let's go find a provider, and oh, let's create a website" type of approach. It worked but wasn't scalable and it wasn't getting modernized quickly enough.

Was your background more in mortgages or technology?

On the mortgage side, I started off at Bank of America, working on loan origination systems. I spent 11 years at Morgan Stanley and one of the things I did was to manage their home loans IT team. I had another 10 years of IT experience outside mortgages before all that. I've been at Cenlar the last eight years.

What brought you into the mortgage industry?

I don't think you could ask many people in the mortgage industry that question and find they planned on it. I was in manufacturing. I managed an IT team for a large-scale print company and I knew that I wanted to get out of manufacturing and into financial services. This was in the early 2000s. I went back to school, got my MBA, and leveraged that in order to transition into financial services. That's how I found Bank of America needed someone to learn their LOS for an implementation. I just learned and absorbed.I became an expert in everything: products, pricing, lock policy, fees, point of sale, secondary marketing. Anything I could absorb I just spent time becoming an expert on it.

I had to go out and find people and really pin them down and ask them questions. A lot of it was working through the app, talking with the developers about why they designed it that way, going back to the requirements, and then going back to the business owners that created them. It took me a good six months to pick up on all the different functions. I was very focused on what it was I needed to do to really get into the industry and expand my knowledge. So every conversation I had was about how do I learn more?

The mortgage industry has a reputation for not being technology-forward. Is it fair to say?

It's very fair to say. I'm in servicing so I feel I can say this: it's even more true, generally, in servicing than in origination. Origination is maybe five years behind and catching up fast. Servicing is maybe five to 10 years behind origination.

Were you able to trade technology expertise for mortgage knowledge when starting your career?

I was getting the better part of the deal early on because I had to learn how their business works and to be able to speak the language. There was no expectation that a person in the mortgage business knew much more than how to open up Excel and I had to become knowledgeable about the mortgage business first before I could do my job. That's why I went back to get my MBA, because I needed to be able to understand both sides of the fence. But as times change, everyone has become more technologically savvy. Now everyone knows what ChatGPT is even if you're not in technology. It was adopted quickly because people are getting more comfortable with technology.

How common is the chief digital officer position in mortgage servicing now?

It is not as common as I would have thought. I see it in many other industries. Where I see it being different is how we leverage it at Cenlar is often you have the chief digital officer and the chief information officer as the same level roles. It's almost like having two cooks in the kitchen, I believe. I don't agree with having multiple technology officers at the same level because I think it creates confusion, silos and other issues. A lot of folks that are in positions like mine are called CTOs. Those whose positions are a little more flexible are called chief innovation officers. Chief digital officer means a lot of different things to a lot of different people and in some cases even goes as far as being on the marketing side, which I am not. I see my role as driving innovation, finding the newest opportunities, as well as understanding the technology well enough to be able to apply it to solve business problems. We're a subservicing company, so our business is not delivering software, we're delivering operations as a service. So my existence at this company is around perfecting the operations process, making our company as effective as possible with AI, data integration and any technology that helps us to be better as an operational group.

What’s the latest wave of technology you’re dealing with in your daily work?

Generative AI for clients, although we're not there yet. We talked about potentially actually creating a chatbot for our clients, because in that case, it's not borrower facing, so there's less risk. But from the feedback that we've gotten and how we see them interacting with us, it wasn't going to drive additional value to them. It'd be more like a bell and whistle and we're very focused on value right now.

Robotic process automation really hit its peak across all industries earlier, it's only now hitting servicing. So we're actually not that far behind to have spun up a pilot in 2021 and really gone full force last year. Because most technologists have kept up with what's going on from an AI perspective, most are finding ways to get some generative AI and other AI capabilities into their company, not all, but at least conceptually most have. I'm actually happy to say we're I feel we're a little bit ahead. Most companies our size, and we're a decently-sized midsize company, they're still coming up with use cases and are accepting and receptive to AI but haven't yet figured out how to make it happen. We managed to figure it out.

We have a use case where an agent has a live chat with a borrower. At the end of that conversation, they have to be able to summarize in notes they put on the loan file. They have to summarize the conversation in terms of whether it was an escrow call or something else. We take that information, we clean it up to make sure there's no NPI and then we run it against a generative AI API, to say summarize the conversation in 200 characters or less, because that's what the field supports, and give us a recommended category code. It pops up a modal on the screen for the live chat agent to see and says, this is what we're going to post. Do you want to change anything? They can go in and make changes and when they're done, they submit. We still have the human in the middle, which is really important, because, especially with generative AI, it could hallucinate, in other words, it could get it wrong. Agents have to review it, but say 93% of the time, they're not having to change anything, so we're shaving 93% of that process off. That means they can spend more time helping borrowers and less time doing administrative work. It's a simple use case,

We've been measuring it now for, I think, several months. Every month, we watch the performance. We have many controls built in to make sure that our agents are actually looking at it, they're not rubber stamping it, and that the results are getting better, not worse. We really thought it through all the way from a governance process down to function and data controls to make sure at every stage of our first implementation, if, say, you were the OCC and you were asking me questions I could be comfortable answering.

To what degree is your technology proprietary vs. off the shelf? Do you have a mix? What category is that function you just described in?

That one we did have to build. We have a vendor partner that we work with for some of the builds, but not all of them. Especially when you're getting into a category like generative AI, there aren't a lot of off-the-shelf products out there. But, to answer your question, whether we buy or we build is not a matter of all or nothing, it really comes down to competitive advantage. So borrowers are relatively captive in servicing provided we can provide a satisfactory solution. Our solution is more than satisfactory. Borrowers can be able to make their payments, get their documents, ask questions, etc. There's a limit to how much functionality you really need in a borrower-facing website for it to be effective on the servicing side.

Origination is a different story. We didn't need as much flexibility on the borrower side but we did want to keep our systems up to date, and if we go outside to buy then we know that every quarter we get updates. But if you look at the build side, in B2B where we're working with our clients, the banks, the credit unions, the investors, etc., then it starts to make sense to provide a proprietary solution. We want to make sure that their experience is something that is better than what our peers can offer.

When it comes to B2B services, clients want to be able to come in and interact with us, get their data, see their most recent policies, and everything that they need in order to be able to manage their portfolio. Some prefer to be able to self-serve as well. So we consider questions like: Can they make their own request? Can they interact with their agent or their client manager? Can they listen to call recordings? They could call in and say, "Hey, can I get a call recording for loan 1234?" We also could build something where they can go online and access a recording. I have seen some other subservicers now having access to that but we're revamping the overall portal so that we have things like real-time dashboard data. That's going to be something coming out soon, to show our clients that it's not just data extracts that we're providing to them, but click-through dashboards they can use to manage the portfolio.

We have a newer request system out there we call the "client case management tool" that they can use to put in requests. That one has been very effective because it gives them transparency over the status of their requests.

Those are the ones that come to mind, off the top of my head: a request management system, the upcoming dashboard solution and the overall revamp of the portal that we're building.

Compliance for AI is evolving. Is that a factor in that decision?

You're spot on. In fact, the CFPB recently came out with a couple articles in the last, I'd say, six months or so on the concerns they have around the proliferation of AI, and the impact that that could have on a borrower. We looked at that and said, at no time soon are we going to be using AI to put results in front of a borrower because if you got it wrong, and it's in writing, we get ourselves in trouble. Even worse, we could be misleading the borrower. We want their experience to be good. So with that, we decided for the short term, we're only going to use it for back office automation.

And you have human review coming into play there?

You've got it, although we do have the chatbot, which we've had for years. But the responses the chatbot can provide or curated responses. It goes through marketing, business, legal compliance and that list of responses is then locked. So there is a bot out there, but the most the bot is doing is natural language processing like Google search, right? The responses are already there. It just tries to read your question to figure out which responses to put out.

How is ICE’s Black Knight acquisition changing the servicing technology landscape?

We're a Black Knight customer, now an ICE customer. What we've seen so far is there were additional solutions that we started to hear about and we're considering now that weren't part of Black Knight. So we now have more opportunities to expand our technology where we still want to do buy over build. Document management is a good example. We have a really strong document management custom integrated with the Black Knight platform, but now we could buy something that is already part of that platform. That's convenient. Another example is with something like a DocuSign process, the ICE team has some technology that will allow us to be able to further integrate that type of process into our default operation space.

As we explain more to the broader ICE team what our business goals are, I get the feeling that more products are going to expand since this is really part of ICE's strategy. ICE purchased Black Knight because they wanted that end-to-end experience. They're looking for folks that have appreciation for the servicing side. They want that feedback. It was definitely a strategic play, to bring in Black Knight to understand what servicing customers want. ICE has the Ellie Mae Encompass front end, that's their big LOS product. So being able to have some end-to-end integrations from the LOS to the servicing side, Encompass clients are going to appreciate what comes out of that.

Security breaches and related reporting are big topics now. What are you seeing there?

Our focus on cybersecurity has never been higher. Obviously, we had already changed out our entire cybersecurity strategy. We have a CISO who joined us about a year and a half ago. He has created a whole new strategy for Cenlar that has intensely focused on the cybersecurity threats that are out there. It's hit our industry specifically, so we know that we have to be on our game.