Tag: credit union

Whitney McDonald 0:02
Hello and welcome to the buzz of bank automation news podcast. My name is Whitney McDonald and I’m the editor of bank automation News. Today is January 4 2024. Happy New Year. Joining me is Sai CEO Mike de Vere. He is here to discuss how AI is improving the decisioning and lending experience for financial institutions, as financial institutions look to serve their entire communities and lean on automation to make smart lending decisions. Prior to this day, Mike spent more than four years at Nielsen served on Google Surveys advisory board, and was the CFO at radius financial join me in welcoming Mike.

Mike de Vere 0:38
Well, thanks for having me, Whitney, super excited to be here. So Mike, de Vere CEO of zest AI, I have been, you know, perfecting the translation of data into insights over the last few decades. I’m here at CES AI, we’re our technology deal delivers and automates underwriting with more accurate and inclusive lending insights. And so just to unpack that, there’s a lot of discussion around automation, certainly with this economy around efficiency. So definitely topical, but foundational to automate your underwriting is you have to have better insights going into the system. And that’s where this more accurate approach to assessing credit comes in. That’s applying better math called AI. And so you can’t just have this more accurate inclusive lending insight. Because it also you have to make sure that you’re serving all your members and all your customer customers. And that’s where inclusivity comes in. And so we have been solving that problem for the last roughly a decade and a half. And excited to share more about the adventure that we’ve been on. It’s just

Whitney McDonald 1:42
great. Well, we are definitely excited to hear more. So thanks again for being here. Let’s start here with kind of a market update tell us about the current lending market. And then we can kind of get into how credit unions can really navigate this space as we close out 2023 and get into 2024.

Mike de Vere 2:01
Well, if I think about the last 100 or so conversations I’ve had with credit union executives a consistent theme surrounds were lent out, you know, in this economy with rising interest rates demand going down because of these rising interest rates. And so many credit unions find themselves in a position where they have very little to support their communities. And what they’re faced with is because of the tools that exist today, there, they’re inaccurate. Their face was really only lending to a small segment of the population, you’re a tear paper. And so, you know, from an economy perspective, certainly there’s a lot of focus in on lending. Really what people are asking us for help with is around decreasing charge offs, improving yield, being able to serve your entire community, not just those at the top socio economic bracket.

Whitney McDonald 2:57
Now, when it comes to being able to accomplish exactly what you were just saying, let’s kind of get into how technology fits into this. And more specifically, we can’t really have conversations right now with talking through AI. So how can credit unions really optimize look to technology, technology, optimize automation, improve underwriting using AI right now?

Mike de Vere 3:20
Well, I think that there’s three pillars that that we work with credit unions on smart, inclusive and efficient. And so smart is, as it says, which is, the current credit system is failing America, whether you’re talking about a good a good economy or a struggling economy, it’s failing America, because it’s only serving parts of it, if you’d segments of the population are left out whether they be thin file, there’s significant segments of the population that are where there’s bias and discrimination in the end. And so, this idea of smart means, we’re appending to the current credit system, which uses roughly 20 variables to assess if we should give a person alone, the current industry scores that are out there, and it tries to boil an individual down to a number. But what we know is that members and customers are more than a number. And so you’d have to open up the aperture and consume more information. And that’s where AI comes in and enables a credit union or a bank of any size to accurately and smartly assess if they should issue that loan. The second pillar that was around inclusion, that’s really where purpose comes in. Because it’s one thing to be more accurate and drive your balance sheet but it’s the second is fulfilling your mission and being able to serve your entire community that you’re within. And that’s why being purposeful about the models that you built to ensure that they’re inclusive and then finally, around automation. Listen, there is such a huge business case right now, for this third pillar on efficiency, where you’re taking this more accurate inclusive Linda inside, but now you’re looking at the the human policies that get overlaid on top and the manual review that gets overlaid on top. So let me give you an example. The average credit union automates their decisions roughly 20% of the time. Now, the challenge with that wouldn’t be is that the average credit union number one, eight out of 10, roughly one a decision in less than a second. And so four out of five are getting kicked out for manual review. You’re really dissatisfying, your customer, that’s a problem. And so really being thoughtful not only about the technology, but around your policies and overlays, is really important. And so what we find is that the normal credit union might have 20 policy overlays, on top of this industry score, which you know, for me is really like duct tape and spit and chewing gum and in dirt, but you’re just trying to put on top of this failing industry score. Well, when you use AI that’s more accurate and more inclusive, you actually have to address those policies, what you find is that up to roughly 20 to 25, probably 10 of them, you don’t even need, because the signals that you’re trying to measure are already within the model itself. So you can dump those out, that manual step is gone. The second bucket is around, well, there’s a lot of policies that frankly, have no signal whatsoever. You know, it’s I love hearing, we’ve had that in place for the last 50 years, the old clo Chief Lending Officer has had that in place. And I frankly, don’t know why it’s there. And so we kick those out. And then there’s this this last bucket around really optimizing policies, so you end up with four or five. And the net result, if you do that implement AI driven underwriting is you should be able to audit a decision 80 to 90%, for those loan applications that come across your desk, which is what customers want. And from an efficiency perspective, dear gosh, probably our poster child in efficiency was able to eliminate two thirds of the resources for underwriting through automation. That’s a heck of an ROI.

Whitney McDonald 7:08
Yeah, I’d say that’s huge. And throughout the year, it’s been a consistent theme across the industry where we’re focused on efficiency, we’re pulling back on costs, where can we automate? Where can we invest in technology? So that leads me to the next question, I know you talked about the three pillars where technology can fit in kind of throughout the institution? How do we really approach this technology strategy? If you’re a credit union? How do you how do you prioritize those must haves? Where do you start, we

Mike de Vere 7:36
did a study of credit union executives and eight out of 10, asked for, and they said that they’d like to deploy AI with an underwriting because the impact that it could have on their balance sheet as well as their members. That to me, is a good starting point. And why do I say that? Because if you think about what a credit union or bank does, at its very core, it’s lending money. And so that foundation, if you get that, right, that cascades out to all of the other technology, things you may want to do as a business. But you got to get that right first. Imagine if you’re overlaying technology on a broken system, it’s a wasted effort, you have to start with a smarter brain at the core of the credit union or bank.

Whitney McDonald 8:20
Now, when it comes to innovation within ZX sai we can kind of get into your technology a bit here. What really are your credit union clients asking for I know that you just mentioned the survey that they’re asking for more AI within the decisioning. What is really driving that innovation within this AI, maybe a few things that you’ve you’ve got in the works or some products that you’ve got working on.

Mike de Vere 8:41
So we actually started solving the most difficult problem, which is how to safely and soundly underwrite a loan. So that’s the core. But now you can move up the customer journey and talk about pre screening or pre approvals, you could actually go down the customer journey and say, Now once I have an individual loan, well, now let me look at the health of the portfolio itself. And understand things like credit migration, you know, 18% of your portfolio was a paper, it’s migrating now it’s 22%. So you’re now skewing more towards higher paper. Within the analytics, you’re able to look at numbers that may be in distress, that are moving from an ATR all the way down to a C tier, and there’s an opportunity to engage them before they end up in collections. And so, from a technology perspective, when you’re asking that question of assessing credit, that’s where our technology really shines. And so pre screen pre approval, we look at the underwriting question itself, as well as portfolio management. Now, I would be remiss if I didn’t talk about some of the significant innovations that we’ve had around fraud and detecting fraud. And so it always starts first with us understand that every customer has their own unique set of issues and so one fraud solution doesn’t fit all. And so for example, our partner So Equifax have a phenomenal fraud solution. But sometimes that might not be the right fit. And it might be that you could use AI. So zeste uses AI to detect fraud, and identify early default and things of that nature. And so it really depends on the individual credit union and their needs and the type of fraud that they’re experiencing. And so I think if I were to say a very, a very consistent theme across each of our offerings, is that we tailor them specifically thoughtfully to that credit union or bank understanding that one size doesn’t fit all.

Whitney McDonald 10:36
Now, speaking of that, one size doesn’t fit all approach. I’m gonna go off script a little bit here. But when it does, when you do get approached by a credit union, or a credit union is interested in Sai, what are those conversations usually looks like? What are they asking for? What are what are you really solving for? I mean, other than the obvious, but what are those questions kind of look like when you’re in those early stages?

Mike de Vere 10:58
Sure. So it depends on what’s going on with the economy. So today, it’s really leading with things like my charge offs are starting to drift up, can you help. And what we know at best is that we can reduce charge offs, roughly 32%, everyone’s across, if you look at the NCAA findings, they’re all going up across the board. And so imagine if you could bend that curve down. The second area is around yield. And so most credit unions are focusing in right now only on their a paper, but there’s almost no yield there. And so what better way to generate capital than having increasing your yield. And then there’s the topic of inclusion, I want to make sure that I’m assuming serving all of my members and 10s of millions of Americans are left out of the current credit system, because of the bias that’s associated within the system. And so there’s a significant opportunity there. And then finally, it’s really around efficiency is that weren’t tough economic times right now, where we’re going to invest is where it makes us stronger and smarter with our lending. And so it really comes down to efficiency.

Whitney McDonald 12:08
And I’m guessing those those topics that you just disclosed that were that were the questions that come about is that kind of helping set up your your plan or your roadmap for 2024, and what your focus is.

Mike de Vere 12:21
So our current product offering actually addresses that. So where we’re expanding in 24, is, first off looking at additional consumer verticals, additional, commercial, vertical, so we’re addressing different types of loans. We’re also going within the customer journey and automating various steps in the process. And so imagine if you’re a large credit union on the West Coast, and you have this great technology company called SSDI, that you work with, and it automates the credit decision in less than a second. But then the underwriter has to manually turn around and do a fraud check. And it takes five minutes while automation falls apart. And so we’ve launched a product called zest connect, where we work with credit unions, and their ability to not only from an underwriting perspective, but identify those other manual steps in the process that can be automated, whether it be through native integration, robotic process automation, what have you, we’re really trying to streamline that customer journey.

Whitney McDonald 13:24
Yeah, that definitely makes sense. And thanks for kind of giving us a look ahead into the next year. Now, as we, as we kind of wrap things up a little bit. What would one piece of advice be for credit unions that are implementing technology that are looking to automate these processes? I know that you just kind of gave that that great example of automate the whole process don’t get stuck after the first piece of the automation puzzle. But what would you give? What advice would you give when implementing this technology kind of getting into the next year? I mean, cost, of course, is one, one area that has to be considered but but what’s the what’s one piece that you would give to a credit union that’s looking into these automation and AI technologies?

Mike de Vere 14:08
Well, so for me, it’s always is the juice worth the squeeze? So there are many executives I run across that have just fallen in love with the technology. I get it. We’re all emotional buyers. But there could be this rational component. And if you have a technology provider, like SAS AI, whose suggestion you can have a 10 times return on your investment within the first year. That’s going to be a pretty smart bet. And so I would encourage people when assessing what technology to prioritize is to ask yourself, the question is the juice worth the squeeze? The second piece is really the people component is that I see whether I was at sastra. And in my past life technology initiatives will fall apart because they forget change management in the human component, that this is a big change you’ll have if you’re talking about underwriting And you’ll have people who’ve been underwriting the same way for three, four decades. And so their willingness to change is not quite there. And so it’s really going to be important for an organization when implementing technology that they understand the role of change management. But they also understand there’s a human impact. And so there needs to be that software approach going forward.

Whitney McDonald 15:25
Now, lastly, as we look into the new year, What trends are you following for 2024?

Mike de Vere 15:31
If I look at 2024, and ahead, I think, you know, one of the big trends that I want to call out is certainly technologies is going to play a big role, and day to day business, but technology and the intersection between that and purpose is going to become even more important as we look ahead. And so purpose is being mindful about when I implement a technology, what outcome am I expecting? And so when I build an AI underwriting model, what outcome Am I looking for? Am I looking for better economics? Well, that’s that certainly is purposeful and how you build it. But there could also be a secondary thing on we also have a mission to serve our community. And so certainly with a credit union, that is core to who they are. And so the question is, are you being purposeful about how you’re building the model to make sure that men and women get a fair shot. Different ethnic groups get a fair shot. And so you’ve got to be thoughtful about how you build the model. It is not just something that happens. It’s having technology and IP, around D biasing the model, and so that you’re able to fulfill your mission. In really lean

Whitney McDonald 16:46
You’ve been listening to The Buzz, a Bank Automation News podcast. Please follow us on LinkedIn. And as a reminder, you can rate this podcast on your platform of choice. Thank you for your time, and be sure to visit us at Bank automation news.com For more automation news,

Transcribed by https://otter.ai