Q&A: Inside Greystone Labs With Director Zac Rosenberg

One of its freshest offerings—called “The Underwriter”—provides data-driven quotes in minutes, which are updated daily using an algorithm that uses over 10,000 transactions and more than 100,000 property comparables to give borrowers a swath of Fannie Mae, Freddie Mac and Federal Housing Administration financing options. According to its website, it will soon offer financing options for the commercial mortgage-backed securities space.

“When you order a Domino’s pizza you can track the [entire process] on your phone,” Rosenberg said. “You can fully track a $10 pizza but not a $10 million loan? I wanted to approach it from an engineer’s perspective to find a solution. That’s snowballed into, ‘wow, we could really change the way this industry runs.’ ”

Commercial Observer caught up with Rosenberg to get a look inside Greystone Labs and learn how its initiatives might change the way the lender does business.

Commercial Observer: Tell us a bit about Greystone Labs.

Zac Rosenberg: Greystone Labs had humble beginnings in January 2015. I happen to be a total geek at heart, the geek of the family. I saw that at Greystone there was a lot of manual work that analysts were doing, looking at deals, sometimes one at a time. I thought I could write a code loop that would do the same thing, but much faster. From a programmer’s standpoint, it was a rather simple script that netted us around $150 million in additional loans [in the few months after it was first implemented in 2015]. It was then that we took a step back and saw we had something here. Since then, our mission has really been to help deliver technology solutions to Greystone’s current lending platform and figure out what elements of lending that we’d keep, remove or do differently. We’re now trying to change what we think the lending experience ought to be.

What areas of the process are you looking to change or make simpler?

One thing we’ve tried to simplify is [correspondence]. Within our loan processes, a lot of it is sending documents to us and signing them and making sure all parties are coordinating properly. Rather than sending hundreds of emails, with weekly updates, let’s do it all online, in the cloud and with everyone on the same page. One feature is tracking where you are in the loan, what has to be done next and how it gets done; we launched [this feature] a couple of years ago, and since, the online loan process has worked through close to $2.5 billion in debt. [As a company] we’ll close $10 billion this year as far as loan originations, which means we see so many transactions.

Tell us a bit about what you’re currently working on within Greystone Labs. 

Where we’re going now is we’re really going after the underwriting process. As soon as you give us an address, I should be able to already know how all the properties in the surrounding area are performing. We have two data scientists and a large database, or dataset, of real estate figures available to us. So, how’s demand in an area? Are there more jobs, and are rents increasing? Is payroll increasing? Are prices in the area getting more expensive and do people have more disposable income? We’re trying to look at everything around an asset, as well as any information we already have, to come up with [an appraisal] today and where we think valuation is headed.

Outside of Greystone’s bread and butter in multifamily, have you shifted focus yet to underwriting other assets?

We’ve started to turn our gaze upon health care because we feel there’s great opportunity there through [agencies]. In terms of what we’re able to do, we know exactly how many facilities there are in the country as they leave a large paper trail. That means I can collect all the Medicare and Medicaid payments that I know those facilities are receiving. I can actually begin doing the initial parts of the underwriting. We can have a good idea of what every health care facility in the country is doing today and where we think it’s going. Because it’s such a large paper trail, it might actually be easier to do than multifamily.

You’re launching the underwriting pilot in Seattle this month? Why that market?

We looked at every single multifamily owner over the last 10 years in the Seattle metropolitan area—everyone who’s purchased, built and/ or managed. We looked at all the transactions, and we set all of this into a clustering algorithm. We wanted to see if different owners fit different patterns and profiles. We wanted to take into account the interests of these players based on their habits. We found owners who were more keen to do ground up construction or deep value-add and then hold the asset for three years and flip. Many real estate companies in the Seattle area are very local and live and breathe there, as opposed to, say, a New York and Miami. So, we looked at products Greystone can offer that might be really impactful to these different clusters of owners. We try to build up profiles of owners to know what would be helpful to them, specifically, and tailor the experience accordingly. So, we’ll show how a lender views an asset and the area around it. At the end, we’ve already done an initial underwriting of how we think a lender would look at a property—from FHA, to Fannie Mae and Freddie Mac.

So, how has this affected your loan processes?

Already, we’ve been able to cut the loan process time down in around 60 percent of our [loan portfolio]. As far as underwriting, we can do it a lot quicker and produce results that might have taken a team of analysts weeks to come up with.

Who are some of your direct competitors in this space?

I’m not so worried about the creation of other cloud-based underwriting models. Being that we have the power of the purse, I think we have a big advantage. As far as current lenders, I’ve chatted with many of our competitors and most of them are still trying to figure out if they want to embrace cloud document storage at all. I think many players are just struggling culturally to make the shift. Groups I’m much more scared about are those that aren’t traditionally in commercial lending but have noticed the demand; the financing might become a bottleneck and when it does, they might get impatient. 

What are some of the risks of putting so much faith in underwriting and originating online in this way?

One of the most appropriate fears is how do you make sure you’re not receiving fraudulent information to boost numbers. We’ve had this issue before, not on the multifamily side, but in our student loan business; we did stuff online and people did send fraudulent information to us.  The more you’re online, the less personalized the experience. Many feel if there’s not people speaking to you and building a relationship, there’s a fear there could be bad actors. But, [underwriting and tracking an asset’s performance] online allows us to see where the market is and where demand is. So, by using more and more data, we’re actually able to see when there are outliers to red flag. It’s similar to fraud detection today, where you’re kind of watching transactions and looking for anomalies.

What’s an example?

Electronic signatures, for example. There’s been a lot of pushback on e-signatures because many people want wet signatures [for real estate documents]. One thing that’s characteristic of these is when you sign, you’re not just signing your name, your sending your IP address, from where you signed it and from what email address you opened it; you’re sending over a ton of digital metadata. Some people were upset about this because you sometimes had correspondence where brokers signed on behalf of their clients and we caught it. You can’t fake an IP address.

We’re always looking for what produces the most value. That’s allowed a single underwriter to not just look at 10 or so deals at once; they’re now able to look at 30 to 40 deals.

In terms of working with Fannie and Freddie, how do they feel about the implementation of this kind of service into deals they’re backing?

I’m much more familiar with Fannie. I’m actually very impressed by that group. They’re very forward thinking. There seeing the same trends and having the same conversations. They want to be scalable and they have no desire to crank back and reduce their volume; they want to [underwrite and originate] in a more efficient way. I know for a fact they’re looking at every avenue, whether it’s AI or it’s RPA [Robotic Process Automation].

There are some very important details and terms within [agency lending] that borrowers should know about. We’re not just trying to show people a small subset of options; we want to show them a menu of everything.

How much scrutiny do agencies give this?

Everything still goes through the same credit process.  I think how they see it is they want their product to be ahead of the game. You know, how do we make this stronger?