- 4 months ago
On today’s sponsored episode of Power House, Diego chats with Rafael Goldberg, the Head of Sapiens Decision, about the role of AI decision-making in all facets of housing.
Rafael talks about how AI has the power and potential to blend business rules and data analytics to make decisions faster and more accurately, all while fitting smoothly into existing systems. He also discusses how AI decisioning can transform personalized mortgage processes like product eligibility and underwriting, and how Sapiens' AI platform can be up and running in just 12 weeks.
Here’s what you’ll learn:
Why AI decisioning is a game-changer for lenders
How Sapiens' platform integrates with existing systems without massive overhauls
The role of AI decisioning in product eligibility and automated underwriting
How the platform can be live in just 12 weeks, offering cost savings and increased revenue
The benefits of AI decisioning for compliance and localized decision-making
Related to this episode:
Sapiens Decision
https://sapiensdecision.com/
Rafael Goldberg | LinkedIn
https://www.linkedin.com/in/rafaelgoldberg/
HousingWire | YouTube
https://www.youtube.com/channel/UCXDD_3y3LvU60vac7eki-6Q
The Power House podcast brings the biggest names in housing to answer hard-hitting questions about industry trends, operational and growth strategy, and leadership. Join HousingWire president Diego Sanchez every Thursday morning for candid conversations with industry leaders to learn how they’re differentiating themselves from the competition. Hosted and produced by the HousingWire Content Studio.
Rafael talks about how AI has the power and potential to blend business rules and data analytics to make decisions faster and more accurately, all while fitting smoothly into existing systems. He also discusses how AI decisioning can transform personalized mortgage processes like product eligibility and underwriting, and how Sapiens' AI platform can be up and running in just 12 weeks.
Here’s what you’ll learn:
Why AI decisioning is a game-changer for lenders
How Sapiens' platform integrates with existing systems without massive overhauls
The role of AI decisioning in product eligibility and automated underwriting
How the platform can be live in just 12 weeks, offering cost savings and increased revenue
The benefits of AI decisioning for compliance and localized decision-making
Related to this episode:
Sapiens Decision
https://sapiensdecision.com/
Rafael Goldberg | LinkedIn
https://www.linkedin.com/in/rafaelgoldberg/
HousingWire | YouTube
https://www.youtube.com/channel/UCXDD_3y3LvU60vac7eki-6Q
The Power House podcast brings the biggest names in housing to answer hard-hitting questions about industry trends, operational and growth strategy, and leadership. Join HousingWire president Diego Sanchez every Thursday morning for candid conversations with industry leaders to learn how they’re differentiating themselves from the competition. Hosted and produced by the HousingWire Content Studio.
Category
🗞
NewsTranscript
00:00You can really think about decisioning,
00:02AI decisioning and decision models,
00:04as anywhere you have a business question to answer with data.
00:11And one of our team members says,
00:14it's a bit like the matrix.
00:16Once you see it, you can't unsee it.
00:24Welcome to Powerhouse,
00:27where we interview the biggest names in
00:28housing and ask them about their strategy for growth.
00:32I'm Diego Sanchez, president of HousingWire.
00:35And my guest today is Rafael Goldberg,
00:37head of decision at Sapiens.
00:40Rafi, it's so great to have you on the show.
00:42Rafael Goldberg, it's a pleasure to be here.
00:44Thanks so much for having me.
00:45I love what you guys are doing at HousingWire.
00:49So really thrilled to be here.
00:51Yeah, I'm really looking forward to this conversation.
00:54So Sapiens has been recognized as a leader
00:58in AI decisioning.
01:00How do you define AI decisioning?
01:03And why is it an evolution beyond traditional business
01:07rules automation?
01:08Yeah.
01:09So let's jump right in.
01:11So AI decisioning, I'll start at the end,
01:13like why lenders should care, and then I'll work my way back.
01:17But lenders should care about AI decisioning because it enables them to compete on decisions.
01:25So much of what they compete on involves decisions, and AI decisioning enables them to make faster,
01:33more precise, more adaptable, and I think this piece is really key, more close to their business intent those kinds of decisions.
01:46And so it's really a game changer for lenders to harness this.
01:53And what this is, AI decisioning, you mentioned business rules, it is an evolution of it.
01:59It is a, Forrester describes it in their report as a bringing together of many disciplines that up to this point have been siloed and require different specialization.
02:11So business rules is one of them, data and analytics is another, and then good old fashioned AI, like machine learning, bringing those three things together in one platform with AI, more modern AI, gen AI, and even agentic AI layered throughout that, and you're getting this really turbo charged opportunity to make those faster, more precise, more adaptable decisions.
02:39So, so we're really thrilled that Forrester and others are talking about this space, they're writing reports about it, and to be recognized as a leader is a real, is really just a real pleasure, pleasure for us.
02:54So there's three different practices or disciplines that we're pulling together into AI decisioning.
03:01How do you apply that to some of the legacy technology and process challenges that we see in mortgage?
03:10Yeah, right.
03:12So that can be scary, right?
03:14Because you're dealing with existing systems that have been in production for a long time, existing processes.
03:20So what's neat about this capability is it's more of a horizontal strategy and capability, meaning let's think about the system, the ecosystem as a whole.
03:35Let's not think about massive transformation and rip and replace.
03:41Let's think about the decision in place, let's provide a logic layer with this AI decisioning platform that you can work with a customer and look at that low-hanging fruit for the use case.
03:53Start there.
03:54You can extract decision logic from an existing system.
03:59Keep using that existing system for what it's good at, like system of record, but take out that complex decisioning that's probably slowing it down and making it more complex than it needs to.
04:09And then if that system is making a third-party call, which I'm sure it is in other areas, this is just another API call.
04:17And then over time, you're going to start collecting more and more use cases into that logic layer.
04:22Your systems are going to become more lightweight, easier to manage.
04:27So that's kind of the technology, the legacy technology play.
04:32And then you mentioned process.
04:34On the process side, you're also going to get a lift.
04:37Why?
04:38Because it turns out that a lot of times we're trying to manage decisions with process.
04:44What do I mean by that?
04:46I mean, if this, then that, but the other thing and maybe this.
04:50And we're trying to do that through process maps and people.
04:55But if you, again, to the legacy tech, if you extract those decisions into a purpose-built decisioning system, those processes can actually get much more simple.
05:08More simple, more stable, less people required.
05:13And now those processes just run.
05:15It's the decision logic that you're going to be changing to be more adaptable.
05:19And so this is a great strategy to address those legacy tech and process challenges.
05:25Yeah, it sounds really interesting and great in theory.
05:31But let's drill into the actual mortgage origination process.
05:36Yeah.
05:36Can you share some compelling use cases that you've seen in your work with lenders?
05:42Absolutely.
05:44You're right.
05:44Let's make it real.
05:45So I would put it into two categories, the one being kind of the larger grain, more obvious and like product eligibility, automated underwriting.
05:57And then the second group is I have a friend that says, you know, the riches are in the niches.
06:02So there's all of these niche use cases where you don't even think about them as decisions because you've never been able to automate them.
06:11You just have people with their checklists and cubicles and a lot of tribal knowledge and exception handling.
06:18So income calculation, mortgage insurance, clear to close.
06:27So but, you know, automated underwriting, that's a huge one.
06:33And you can really think about decisioning, AI decisioning and decision models as anywhere you have a business question to answer with data.
06:45And one of our team members says, it's a bit like the matrix.
06:50Once you see it, you can't unsee it.
06:52And so you start with that first use case and then you start to recognize all of these business questions that you're answering with data.
07:00And you're like, wow, I could create a decision model for that.
07:02And so so you may have that that first group of those big obvious ones, product eligibility, automated underwriting.
07:09And then you have those niche ones as well.
07:11Housing Wire is on a fairly new AI platform that is more applicable to media companies.
07:19But it took me like a month and a half to build my first assistant on the platform.
07:24But then it took me a week to build my second and third assistant on the platform.
07:28Exactly.
07:28So you get excited about it.
07:30Like I think what you're saying.
07:31Yes, 100 percent.
07:32Now, that's awesome to hear.
07:33And you're right.
07:34And also these platforms are focused on one of the things that Forrester talks about is usability.
07:43And a part of that is reusability.
07:45So you talked about, you know, going from a month and a half to a week.
07:49And you can do that with these platforms as well, because you essentially start developing these building blocks of assets.
07:55And then you start to compose them together.
07:58So you talked about the origination process, and it sounds like there are a variety of use cases that can plug in to origination.
08:06Are you also looking at the secondary process or maybe even more fruitful, the servicing process?
08:14Absolutely.
08:15So there are great decisioning use cases throughout the mortgage process.
08:20It's interesting you mentioned marketing.
08:23And so even before origination, there's decisioning in marketing, meaning, you know, which product should I market to which customers when?
08:32That's a great decisioning use case.
08:34Post-origination, secondary, that's huge.
08:37Investor eligibility decisioning.
08:39We have a customer that implemented for that type of use case, and they were able to just achieve things like, you know, 80, let's see, 60 percent.
08:53Reduced risk of repurchase, meaning, and 20 percent increase in ability to determine where they could sell loans to, which investors to sell loans to.
09:04And then servicing, very rich, some challenges there with systems, but certainly loss mitigation is an obvious one for decisioning use case and servicing.
09:18Yeah, so many areas that you can plug in, and it sounds like you're plugging into existing tech and then allowing folks to extract decisions out of there, but you don't require people to rip out their legacy tech.
09:34Exactly.
09:34And so we're not looking to be a replacement system, a verticalized business application.
09:41We're looking to be that horizontal layer, that strategy, make your existing systems lightweight and work better.
09:48We provide a simple API using OpenAPI standard.
09:51And the really nice thing that we've seen for customers is once you have that strategy in place, when you're ready to make that larger systems transformation, you're not recoding your decision logic.
10:06Just because your system's changing doesn't mean your decisions are changing.
10:10So you just start with that new system, but your decisions are there waiting for you.
10:14So you get this massive acceleration and kind of future-proofing.
10:19So let's walk through a couple of your clients and what you've seen in the way decisions get made, the shifts that you've seen in the way decisions get made when you put some of these tools in their hands.
10:34Yeah, that's a really interesting question, like kind of on the process side and operating model, I'll call it.
10:41One of the interesting things that happens is, I mentioned this at the beginning, this idea of making a lender's decision more closely to the original intent, the business policy.
10:57If you think about traditional methods like business rules, you have the policy created, written down, maybe a business analyst gets involved, a configurator then takes it or a developer.
11:10And then you're dealing with some pretty significant constraints of the target technology.
11:18It's like a game of telephone and compromises are made.
11:23And what's running in production may be a version of policy.
11:29Now, with these tools, they are no-code, visual, drag-and-drop interfaces, but they're not shadow IT.
11:39And so business loves them because they can see what's running in production.
11:47They get a lot of confidence.
11:48IT loves them because they're not interpreting written requirements.
11:53They're just getting an API to interact with.
11:55And their constrained resources are freed up to work on fun and hard engineering problems instead of interpreting requirements.
12:04And so you really get this great trust back between those two parts of the organization, business and IT.
12:15And so it's like the shift left of implementation.
12:20You get much more robust testing up front, and your outcomes improve, and your cycles get faster.
12:29So that's how I would answer how I see our customers adapt when they bring on this type of technology.
12:36So fingers crossed that this is changing, but we've been in a margin compression environment for several years now where every basis point matters, right?
12:48Yeah.
12:48And so I would imagine that part of your pitch when it comes to bringing AI decisioning into organizations is, well, you can also save some money here.
13:00100%.
13:01You know, it's obviously top of mind for everyone.
13:05I think what's great about these platforms for lenders is by implementing them, first of all, it's not a huge lift.
13:15You're not talking about a years-long implementation.
13:18Our clients can get live within 12 weeks because of some of the things that I've talked about.
13:22But you're really able to increase revenue more profitably.
13:28And so what do I mean?
13:28For example, if you're able to implement better, more precise, more adaptable, more granular decisions, you're probably going to be able to have your loan officers get up higher pull-through rate, more files per day.
13:44If you're able to automate some of that checking the checker niche decisioning, then you're going to have less touches per loan.
13:52And so you can see increased revenue more profitably.
13:57And so it's really a great strategy to deal with exactly what you described, which is these really tough, compressed margins.
14:05So earlier in the conversation, you talked about compliance and policy and then implementation of policy out in the field.
14:19And in distributed retail, you can have different implementations of the policy in different branches, right?
14:26So how does sapiens and AI decisioning help bring some transparency and help with compliance while also allowing some of that entrepreneurial spirit that lenders have who are moving faster and with more confidence?
14:45Yeah, no, it's a really important question, especially because AI is involved and I don't want people to walk away with the wrong impression that an AI is making the decision.
14:57And so what's neat about this, we talked in the beginning about this combination of disciplines, you're getting the advantages of AI to speed up some of your processes of creating these implementations, but you're combining that with a highly, highly declarative visual model.
15:19So you and I, the business, IT, anyone can look at that model and understand exactly how the decision is being made.
15:28And when it's pushed into production and executed, you're able to trace back from the actual decision that's made in production all the way back to the original change request.
15:38And so you get this extremely clear path on your compliance side.
15:45And then on the entrepreneurial side, what's neat is the platform provides you with a lot of opportunity for creating what we call views.
15:56And so the Southeast may have their view of a certain product and the Northeast may have their view.
16:02And that's available for the system to understand when to call which.
16:08So you really get the best of both, that combination of those metrics.
16:13So I'm hearing that you can almost have a local flavor, right?
16:17There's a business-wide policy, but there's a local flavor that can be executed on by a branch or a region.
16:26That's a really good way to put it, yeah.
16:29Well, this is really interesting stuff.
16:32Our audience is really engaged with AI as a topic right now.
16:36If they want to learn more about AI decisioning and your business, what should they do?
16:44Thanks, Diego.
16:45I mean, it'd be great.
16:46You can go to our website, sapiensdecision.com.
16:50I mentioned the Forrester reports there available to read through.
16:53It's a great kind of primer or primer on the space, and it highlights a lot of great vendors.
17:00We have demos on our website.
17:04We did a really fun demo day with HousingWire with you guys, and we have that on the website,
17:10and they should check it out at HousingWire, too.
17:11And I think that demo really goes through in 10, 15 minutes the core value proposition,
17:18so that's a great way.
17:19And then, of course, reach out to us, and we're really happy to have a conversation,
17:24understand what you, as a lender, are trying to accomplish and how we might be able to help.
17:28Well, Rafi, again, this is a great conversation.
17:33Thank you so much for joining me on Powerhouse today.
17:36Oh, it's my pleasure.
17:37Thanks so much for having us.
Be the first to comment