Unlock the future of mortgage lending! Discover how AI decisioning is revolutionizing the industry, enhancing compliance, and creating a competitive edge.
In this insightful interview, learn about AI's role in streamlining complex decisions, from policy interpretation to improved outcomes. Explore how businesses can leverage AI to accelerate processes, reduce costs, and launch innovative products.
Find out how a decisioning layer can integrate with existing systems, empowering business leaders and IT professionals alike. Don't miss this opportunity to understand AI decisioning and its transformative potential!
00:00Live from Las Vegas, I'm Diego Sanchez, president of HousingWire, and I'm joined this morning
00:14by Raphael Goberg, who is the head of Sapiens Decisions.
00:18Raphael, thanks so much for joining me today.
00:20Thanks, Diego.
00:21It's great to be here with you.
00:22So what does that mean, head of Sapiens Decision?
00:24So we have an AI decisioning platform, and AI decisioning is really a new category,
00:32and it's bringing together three evolving disciplines into one place.
00:39And so as head of Sapiens Decision, what I do is I run the business, I focus on product strategy and go to market.
00:46And the three pillars of AI decisioning are good old-fashioned AI, like machine learning, as well as gen AI,
00:53business rules and the evolution of that, which is really decision management, and data analytics.
00:59Now, all three of those have been operating independently as silos, and an AI decisioning platform,
01:05a part of which we're bringing to market, is bringing those together so that businesses can engage with that entire ecosystem
01:12in one cohesive experience.
01:14So other industries maybe are more familiar with the Sapiens brand name.
01:21Maybe mortgage is less familiar.
01:23Tell us about the history of the company and what sectors you've been working with to date.
01:28Yeah.
01:29So we got our start about 15 years ago, and we've been running at scale in production, battle-tested.
01:34We really started coming out of the business rules space and rapid application development space.
01:41We felt like there was a huge gap in the market.
01:44There were business rules engines, technologies, but fundamentally no one had provided the business
01:50with a methodology and a model to define, to discover, and define their business rules requirements.
01:56And we said, let's start from the beginning, first principles.
02:00We came up with the decision model, and one of the GSEs actually loved what it did for their business,
02:05and they said, build us an enterprise platform for that on a condition that you'll take it to market
02:10so we can create a community practice.
02:12Okay.
02:12That's exactly what we did.
02:14And so we've had great success with larger lenders, larger banks, as well as insurance carriers and public sector.
02:21But we're really bullish and are reinvesting in mortgage roots where we started and super excited about what's to offer.
02:31Well, it sounds almost like you didn't push into mortgage.
02:34You got pulled into the mortgage by one of the GSEs.
02:37Yeah, that's a great way to put it.
02:38That's really interesting.
02:39Yeah.
02:40So walk us through what AI decisioning actually is.
02:45AI is a big buzzword.
02:47When you connect it with decisioning, what are we getting?
02:50You can think of it as, so the cornel, the core, are the decision models.
02:57These are the representations of the business policy.
03:02Now, the AI piece happens in two ways.
03:04One is generative AI to help you build those structures, but you're not going straight to production with those structures.
03:11So you have a lot of governance and human in the loop and validation and testing,
03:15but you're harnessing AI to help you interpret policy documents into these decision models.
03:21Then you're implementing those decision models.
03:24Now you're getting execution data and analytics, and now you're applying AI again to help introspect that data,
03:32those analytics, to help inform you how can you improve those decision models to improve your outcomes.
03:37And so it's really a virtuous cycle at that point.
03:39So walk me through the integration process.
03:43Am I needing to rip out systems that I already have in place?
03:48How does that integration process work?
03:51Yeah, so we, from day one, set out to be a good, a playmate in the playground, you know, in the sandbox.
04:00So API first, and really, we didn't want to replace systems.
04:05We wanted to elevate decisions as a first-class asset of the enterprise.
04:10So if you think about customer experience and data, both of those are thought of,
04:16and if you look at an architecture diagram, you can see where they are.
04:18But where are the decisions?
04:20When I ask business leaders, where are your decisions, it's hard for them to point because they're kind of everywhere.
04:26So we said, let's elevate that to a decisioning layer.
04:29That's what our platform does.
04:31And then let's make it an API economy easy to integrate with so you don't have to rip and replace those systems.
04:37Who tends to become your decision maker inside the organization,
04:42the person that you're kind of lifting up and enabling and empowering with this platform?
04:47Yeah, that's a great question.
04:48And it really gets to the heart of a lot of the conversations we've been having here at the conference.
04:53It's all about operating model.
04:55You know, how are you going to integrate this, not technologically, but from the operating model perspective?
05:02And from that perspective, it's a marriage between business and IT.
05:06And that's been our mission from day one.
05:07So on the one hand, we want to enable business leaders, functional leaders who are producing policy and they need to have a way to give that to IT to implement.
05:20And we kind of bridge that.
05:22And from the IT side, it's really that enterprise architecture group and developers that don't want to be interpreting ambiguous requirements.
05:29They want to do hard engineering work.
05:32And so we're bridging that gap between the two.
05:34So what's the main priority?
05:37Are you helping people reduce costs?
05:40Are you helping people speed up?
05:42Is there other parts of the organization that you're integrating into and helping with?
05:49What's the main focus?
05:50Yeah.
05:50So definitely speed, time to market.
05:53So we've had a lot of customers that we've been talking to this year.
05:56They want to release new products, complex products, for example, non-QM products.
06:02That's hard for them to do with existing systems.
06:05This allows them to get to market quickly with those.
06:09So speed to market.
06:10And then cost.
06:11If you can't automate decisions that otherwise you're really handling by exception and with manual policies and check the checker, you're going to reduce costs.
06:19So time and cost are both there.
06:22So ultimately, lenders need to differentiate on something.
06:27You know, some folks, it's speed to close.
06:30Other folks have great technology.
06:32But a lot of folks really try to build trust with the borrower.
06:36Yeah.
06:37And really be there for the full customer lifecycle for that borrower.
06:40How do you help lenders build that trust with their borrowers?
06:44Yeah.
06:44That's a great question.
06:45And I think it's really important in the age of AI.
06:48And, you know, our category has been called AI decisioning.
06:52And so we have to be cognizant of that.
06:55And the beauty here is we're using AI to help speed up the process, but not to actually be the process.
07:03And so the decision model and the analytics it produces, it's very much a clear box.
07:10And so what lenders love about using our platform is when the auditor comes, they can go to the reports, push a button, and they can show exactly why each decision was made.
07:21And so that consistency is going to build that trust with their borrowers.
07:27So I think some of the nervousness at a lot of lenders about AI stems from, how can I do quality control?
07:35This is kind of a black box.
07:36How do you ensure that there is quality control and compliance as lenders integrate AI into their decisioning process?
07:45Yeah.
07:46So it's key.
07:47And, you know, back to our roots, starting with the GSE compliance, especially post-financial crisis.
07:52I mean, that's what it was all about.
07:54Okay.
07:54And so it's super rich governance.
07:56Every asset has an approval lifecycle.
08:01It's customizable.
08:03And so any kind of AI that you're bringing into the system to help you inform those decision models is being checked, reviewed, validated, tested, regression tested, and then finally deployed.
08:17And so you're really pulling back the curtain on anything that would otherwise be a black box.
08:23So I consider you an AI guru.
08:26Take out your crystal ball.
08:28Where do we go from here in terms of AI decisioning and you growing your business at Sapiens?
08:35Yeah, that's a great question, Diego.
08:37So, you know, I feel like in some ways there's a lot of excitement, but we're also maybe a little bit in the trough of disillusionment with AI.
08:47Nonetheless, despite that, I really do believe in the agentic era.
08:54I think it's upon us, this idea of a semi-autonomous enterprise.
08:59And so we're doing POCs internally and with customers to drive out how can we apply agentic AI to the decisioning process.
09:10And I have to say it's really exciting.
09:12So I'll give you an example.
09:15One of the POCs we've done is, okay, we've talked about, we've used Gen AI to go from policy to decision models.
09:22And now we've deployed those.
09:23Now we're getting analytics.
09:24Can we apply natural language KPI to those analytics and have the agents tell us, based on the decision models and based on the KPIs, are we achieving the KPIs?
09:42And if not, if there's drift, what suggestions should be made to those models?
09:47And so that's, I mean, that's game changing.
09:49Totally.
09:50I mean, not just for mortgage lenders, but, I mean, most organizations want to hit their KPIs and understand what they need to be doing if they're not hitting their KPIs.
10:02Absolutely.
10:02Really interesting.
10:03Yeah.
10:03Well, Rafi, it's always great to catch up with you.
Be the first to comment