AI and Credit Unions: Striking a Balance Between Innovation and Regulation

In an insightful discussion on The Credit Union Connection, Sarah Snell Cooke sat down with Mitch Rutledge, CEO and Co-founder of Vertice AI, to delve into the evolving landscape of AI in the credit union movement. With over a decade of experience in AI, Rutledge offered a comprehensive perspective on how credit unions can leverage artificial intelligence for growth, efficiency, and enhanced member services, while also addressing the crucial aspect of regulatory oversight.

The conversation kicked off with a discussion about a GAO report concerning NCUA’s oversight of AI, which suggested that banking regulators were ahead of the NCUA. Rutledge emphasized the need for a balanced approach to regulation, one that is “clear but not restrictive.” He cautioned against a “one size fits all regulation path,” highlighting that AI used for member engagement differs significantly from AI used for lending decisions. This nuanced view is crucial for fostering innovation without stifling progress.

efficiencies through robotic process automation (RPA), streamlining processes like loan approvals and document collection. AI-powered credit decisioning, as seen with companies like Synapse and Zest, is also enabling alternative decisioning to offer more credit to a broader range of members. Fraud detection is another critical area where AI solutions are making a significant difference.

One of the most compelling aspects discussed was how AI can revolutionize … Watch the full interview for more!

Disclosure: Transcript below is automatically generated

Sarah Snell Cooke
There we go. Hey, ready, ready, ready. All right, hello. Welcome everybody. My name is Sarah Snell cook. I am your host here at the Credit Union Connection. I’m here today with Mitch Rutledge, welcome. Thank you for having me. And Mitch is the CEO at Vertice AI, and he has 10 years experience in AI, so I’ll let you go ahead and do a little more introduction of yourself, but great background for sure,

Mitch Rutledge
sure. Sarah, so I’m Mitch Rutledge. I’m the CEO and co founder over at Vertice AI. We launched Vertice not quite three years ago as an AI powered member growth solution for credit unions. So we’ve been excited to be bringing the power of advanced analytics and AI to the credit union movement to really focus on empowering credit unions to know, grow and measure every member as they climb their financial mountain. As you said, before launching this, I spent over a dozen years working for big data and AI platform companies, bringing the power of these models and capabilities to lots of different industries, and so that’s where I sort of learned. And my co-founder was a product manager there. I’ve been in technology my whole career, and software and delivering solutions for different industries. So there’s been an exciting journey that we’ve had in the credit union movement.

Sarah Snell Cooke
Yeah. How much must that have changed over the last 10 years, for sure? So the reason you had reached out to me was you were, you saw, you saw an article that we had done on the GAO report regarding NCO ways oversight regarding AI, it wasn’t particularly favorable, but basically saying that the banks were ahead, banking regulators were ahead of the NCUA. And I find, obviously there’s guardrails needed. We need innovation, we need guardrails, and that needs to be balanced. But I tend to find that regulators like to work toward the lowest common Dominator, kind of you know, whoever the bad actor is, that’s with who they’re regulating to when there’s 99% who aren’t doing that thing, whatever it is. So what would you have to say about NCUA approach to the AI oversight the GAO report.

Mitch Rutledge
Well, I mean, I think you stated a bit of it right, is that we have a we have a chance to strike a balanced tone and come up with something that’s clear but not restrictive. I think if we take that, like you said, the kind of lowest denominator approach that might be that’s not going to be ideal for us to continue to, you know, drive innovation in the movement. So, you know, I think borrowing from things like NIST and some of the other frameworks, but tailoring them to credit unions as a really great opportunity to, you know, keep driving innovation and serving members, but understand risk based oversight, right? I mean, that’s the key. Ai that’s used for member engagement is very different than AI that’s used for lending decisions. And if we go down the one size fits all regulation path, right? That’s not going to be that’s not going to work, and it’s not going to be ideal to, you know, continue to drive innovation.

Sarah Snell Cooke
Yeah, as you mentioned, there’s a ton of different ways that credit unions can use AI, a lot of people, because chat GPT is popularity, and the rest of them Claude blanking on the rest of anyway, they’re all around and so that’s a lot of what people think about when they think about AI. But within credit unions, there’s a lot of different use cases, and can you just kind of highlight a couple that really makes sense for credit unions?

Mitch Rutledge
Right? Well, I think you called out something important when we talk about these new generative AI models, right? The chat gpts and the quads and those sorts of solutions. Those are new, but we’ve been using AI for, you know, many, many years, right? FICO scores are built on machine learning models, which is a form of AI, and we’ve had those for how many, right, 40 plus years. So this is not new, if you’re doing, you know, card swipe, fraud detection, those have been using neural network models for dozens of years. So these have been in the movement for a long time. Just seems it’s become much more of a topic, as we’ve seen this the rise in the generative AI capabilities in the last couple of years. So, you know, I think the use cases are far and wide, but those are the ones that have been around for a long time. At Vertice, we’re very focused on leveraging the power of AI to distill data that credit unions have to better, you know, like we say no, grow and measure every member. So that’s about understanding needs and behaviors of members so that we can offer them more personalized product and service recommendations. But we know about the credit decisioning. There’s lots of great companies that are, you know, taking, you know, bringing. Those kind of capabilities to do alternative decisioning for offering more credit to more members, which is a big, important part of what we’re doing. We’re starting to see some of the chat bots, both for internal use cases around policies and procedures, as well as some people experimenting with that for member facing functions. I think that’s the one that’s people are discussing more of like, are we ready for that? But there are many use cases, you know, even some of the, you know, the RPA that’s been around for many years, robotic process automation is leveraging some AI capabilities around, how do we make loan approvals more efficient? So I think there are many, many use cases that have this as underlying capabilities. I think the biggest part is that all of the providers that credit unions work with are likely trying to find ways to embed AI capabilities into their offerings today. And so just like you know, the FICO scores and credit decisioning have been doing it for many, many years. It was embedded. You might not have known it. Or again, card fraud detection has been doing it for many, many years. It’s embedded into it. More and more of that’s going to be coming, right? CO pilot’s another great example where we’re seeing it embedded in all of Microsoft offers, right, right? And again, we think about autocorrect is just another that’s an AI model, right? That’s doing types of predictions. So not my favorite, either. But

Sarah Snell Cooke
so, yeah. I mean, I think, you know, the chat gpts of the world just being in your face kind of has made it more obvious, whereas before, it was behind the scenes, and we didn’t know. And but there are amazing things like robotic process automation that can, like, speed up your business lending, speed up your, you know, just by collecting simple documents. And that’s a waste of time for a human to do, and not to mention the the chance for human error as well. Whereas you know this form field isn’t filled out, you’re definitely not getting through right, right with the AI. So you talked, I think we hit on my next question a little bit already. But how are, how are the most advanced credit unions really employing AI and getting the best results as far as efficiencies, or, you know, whatever the benefit

Mitch Rutledge
is, you know, I from what I see when I go and talk to credit unions, I think those examples around driving back office efficiency as the number one use case, as well, as, you know, We’ve seen the rise of the AI powered credit decisioning right the Synaptics and the zests and those folks that are, you know, I think have done some great work in the industry as how do we take a different way to determining credit decisions? So I think those are really powerful ones that we’ve been seeing. I think that we’re also starting to see it in in the areas of marketing and personalization, and that’s where we’re, credit unions we talk to love the vision that we’re laying out around, how do we distill data that we already have about members to better understand and personalize their experience? And you know, frankly, that’s what the members want, right? I mean, we say we live in the Amazon, Spotify Netflix era. We expect personalization from everywhere, including from our credit unions. And so they have to be thinking about how they’re going to do that. So I think that’s another great use case. And as you said, the RPA and kind of doing automation of back office functions. And then the last one, again, that we’re seeing, you know, we don’t see it as overtly, but it is around fraud detection. Lots of great solutions on all kind of all of the threat, threat vectors, lots of great solutions out of there, out there in the marketplace, leveraging it, you know, to help credit unions. Yeah,

Sarah Snell Cooke
so, and you brought up marketing, obviously, that’s what you all are into. And you know, one of the things that’s happened when there’s a poor economy or what have you, is that marketing is usually the first cut, and with the data driven, not only can you better serve the members with the data that you have, but also better reach the members that maybe aren’t using this product or that product. Are you seeing credit unions, one ad members, you know, leveraging leveraging your tool to add members, but also to deepen the relationship you have with the members.

Mitch Rutledge
So that’s exactly what we are empowering credit unions to do. And to your point about budgets, a big part of what we’re trying to do is help them be more efficient in that process, right? Again, one of the things, when you think about new member acquisition, that’s a costly process, right? I think the stat says it’s roughly $500 to add a new member. And so at Vertice, we are trying to bring a more data driven approach to, as we say narrow the funnel, right? So the old ways of we’ll go blank in a neighborhood with mailers or whatever it might be, you know, that’s expensive, and the conversion rate is typically low. And so if we can take a more data driven approach to really finding those prospective members that you know are going to be those, one. Long term, believe in the values of the credit union members. Let’s be a little bit more targeted in who we are. You know, paying to do outreach to right direct mailers are expensive, so if we can focus on not just, you know, the the old spray and pray and get a little bit more precise in our engagement, we will, you know, we think that’s a much more efficient use of the marketing dollars that to your point are, we have to be there are precious. So I think that’s one, you know, key around what we’re trying to drive. The other piece is while, while marketing is a primary use case. And I did say that we do believe that, you know, credit union growth and Member Services, enterprise wide, right? So that’s part of our messages. While, you know, marketing, obviously, is a primary use case, we see other functions as well, right retail teams, you know, lending teams, strategy, do we have the right products for the members? So using that data to just, you know, understand the memberships needs and wants, and do they align with what we’re offering, and where are the gaps? Is part of that, what we say growth strategy. So, you know, growth strategy is enterprise wide, and growth is around, you know, serving the members we have. Because, you know, a big part of growth is protecting the memberships we have. So can we reduce churn? You know, expanding the relationship we have with existing members, as well as going and finding the next generation of members, which is a critical part of of that. So I think those are all three legs of the growth school,

Sarah Snell Cooke
right? Absolutely. Marketing,

Mitch Rutledge
well, and I would say it’s not just marketing’s job. It is the, as you said, is the whole, you know, everyone in the credit union needs to be thinking about, how do we serve and grow relationships in the membership?

Sarah Snell Cooke
Yeah, obviously at that enterprise level, you know, and and down to the individual member themselves and their needs. So what are the character, general, more general characteristics of the cranes that are doing the most with AI and getting the most benefit?

Mitch Rutledge
We see, it doesn’t relate to asset size. It’s about progressive leadership that wants to be a little bit experimental. You know that they are willing to have sort of a test and learn mentality to find, you know, opportunities to do things differently, to serve the members. I think there is mostly a belief that we want to focus and not true member facing as first tests, right? So how do we test this with driving efficiency, with our team members, you know, kind of that back office efficiency as the you know, the starting points. But I it is absolutely not asset size, right? We have credit unions as small as, you know, 30 million in assets that that see the value in this. And it’s about leaders that say, we want to find new ways to innovate, to serve the members, as you say, doing more with less. And you know, want to, want to, you know, be innovative. I think that does come down to making sure that the boards of those credit unions are understanding of that innovation desire. And, you know, let’s be honest, not all boards are are like that. So I would say, as an as an movement, we need to do more education around what this is and what it’s not. And you know, like we did a session at GAC that was, you know, you’ve been doing AI for, you know, dozens and dozens of years through credit decisioning and FICO scores and, you know, fraud detection on this. So this isn’t new. This is just an another version of something we’ve already been doing. And so I think continued education is key.

Sarah Snell Cooke
Mm, hmm, absolutely. So what sort of results are you seeing? Can you put numbers to it?

Mitch Rutledge
Yeah, from our perspective, you know, our credit union partners, when they take a more targeted, precise approach to member engagement, we see anywhere from two to five times higher conversion rates, right? I mean, it’s, it’s significant, and that, you know, comes from we’re talking to the right members about the right products, services, programs that they want to hear about. And when you do that, you get better conversion rates. You know, I tell this story about my credit union that I’ve been a member of for over 30 years, right lots of products and services, and they are not taking a data driven, targeted approach. And I get an email every week, and it’s, you know, sign up for the scholarship. And, you know, none of that applies to me, right? In my state of life, right? Do you need, you know, just not knowing the members, then you’re not getting those engagement and so when we talk to them, when we see credit unions engage around the right products and services, we see higher conversion rates, higher balances. You know, we believe that’ll drive NPS scores up all of the things that we want to have and, you know, ultimately help them grow and thrive in an efficient way.

Sarah Snell Cooke
Yeah. I mean, it’s really just taking that. You know, credit unions are always prided themselves in knowing the members and in the branches, but not right? This is what, this is the digital equivalent.

Mitch Rutledge
That’s right, right? So we’re 100% and, you know, it’s funny, you use that we, my co founder did a session, and it was from smiling faces to AI interfaces, right? So in the old days, it was that branch to branch interaction. I was literally in a branch yesterday. I’m out in California at a conference. I went into a branch yesterday, and I was waiting for my meeting, and I saw the interactions of the you know person. It was two generations of members having a conversation in the branch with the branch staff. They knew them, and it was that they know the members. But the reality is that there is much more digital interaction, and how do we keep that personal experience in the digital era, and that’s about we have the data we need to think about, how do we personalize in a digital way? And that’s, you know, very much, what the power of AI will be able to do. And if we bring this back to, you know, where we started with around the NCUA, I think they need to recognize that as they consider, you know, what are we going to do? Right? If they come down heavy handed, and they, you know, either put, put something that, I’ll say, scares the movement into, well, we’re going to have to put a pause on all of the things we’re trying to do around leveraging AI to serve the members. That will not be, I mean, that will just stifle innovation, and, you know, stall anybody from doing anything. And that would be just, you know, not a great outcome for the movement,

Sarah Snell Cooke
absolutely. Yes. We’ll see. We’ll see what. What happens at NCUA? Only one board member right now, they can’t do much of anything. Allegedly, there’s, of course, debate on that, so I’m going

Mitch Rutledge
to leave that for a different person to comment on. I’m going to stick

Sarah Snell Cooke
to this. That was my comment. Yeah, you don’t have to respond. So for those credit unions that haven’t started yet, where? Where should they start?

Mitch Rutledge
I would say, start somewhere, right. Go. You have lots of partners and providers and leagues. There are lots of solution providers out there. Find where you think the biggest challenge or opportunity is to serve your members right? What’s causing you heartache, either operationally or, you know, from a member perspective, and, you know, go find a way to leverage the power of AI to really, I would say, start with helping your associates right, be more efficient. You know, I’ve given this talk a few times, you know, I think we should stop talking about artificial intelligence and we should be talking about augmented or assisted intelligence, right? My perfect metaphor, you know, this shows my age, is, I think about Knight, rider and kit, right? Kit was the perfect AI assistant, right? That’s what we want, we want. Kit was a, you know, he was a great co pilot, and he was an agent for, you know, for Michael Knight. And that’s what we want, and we want to give our associates kit to help them be more efficient and more effective in understanding the members and serving them. And that’s those are the kind of projects that I would say every credit union should be finding, whether it’s around, you know, more efficient loan approval, more personalized member engagement messaging for frontline staff, or in marketing. All of those are opportunities for us to leverage this, you know, this capability,

Sarah Snell Cooke
yeah, and I love the reference, by the way, thank you for that. Just had the was a Trans Am in my

That’s right, that’s right.

Sarah Snell Cooke
Oh, the 80s. So how can small credit unions, seems like they could take the most advantage, they could get the most benefit out of it. You talk a little bit about that,

Mitch Rutledge
yes, and we have, we have lots of smaller, you know, sub $500 million credit unions, like I said, that are seeing the value of this. So it is absolutely a way to take, I have a one or two person marketing team, and give them this powerful tool to, how do we do more efficient and, you know, effective campaigns, right? And scale that from doing one or two campaigns a quarter to, you know, can we now do many more campaigns and personalizing the messaging and content that they have? So I think there’s absolutely an opportunity for smaller credit unions to do it. You know, in some respects, they’re a little bit more nimble, right? That they can, they can, you know, they can take these test and learn opportunities to do that more efficiently. They all have the same kinds of data. And, you know, the cores, for the most part, are pretty similar. To be able to get the data that that they can start to to distill this into into action. So I don’t think size of assets is any kind of way that can impede this. So I think it’s about the desire of leadership and educating the board to say, we want to take some steps to, you know, again, test and learn and find opportunities to serve the members in a more effective and efficient way.

Sarah Snell Cooke
Right? Yep. And so we’ve started this conversation with talking about the oversight of the NCUA. How are credits self regulating? It with AI,

Mitch Rutledge
from what, from what we see they are, you know, they are taking steps to make sure that they understand again, when they think of. Providers and partners that are providing solutions and services to them, I think making sure that those providers have transparent, you know, model outputs. I think, as we say, this idea that there’s always a human in the loop, right? It’s not that this is truly artificial. It’s doing things on your behalf, for the most part. It’s how are we making the associates more efficient and effective? So it’s this idea of serving up decisions more efficiently for your associates to make those quicker and more effectively. So I think that’s those are two big keys, transparency, as well as this human and loop concept are really the critical ways to do it. And I think education right continuing to make sure that from the top the organization, the board level, down to the front line, that they understand, what are these capabilities? What are these tools? You know, that’s just, I think as a society, we’re continuing to do that. I think about my my kid, and what he’s starting to see in school, and how do we educate them, not specifically on a tool, but on this kind of types of tools. And what are the things, you know, the guardrails that we should be thinking about. So I think continuous education will be key as well. Yeah,

Sarah Snell Cooke
yeah. Because one day there will be, and one day not too far off a job title of, you know, chat, GPT, prompt, writer, something like that. Because it does. I mean, a lot of people think they can, you know, say, make me an ad or whatever, but you got to be very you got to already have the knowledge to begin with. And I, couple of summers ago, I busted an intern who I knew was using chat GPT to write blogs and stuff. And stuff, and then it was just it, there’s, there’s, I mean, even now, not just a year and a half ago, there’s a big difference there. There is big difference. So there definitely needs to be that, at least on the marketing side, from my experience, it needs to be a lot of human intervention. And so, okay, we always wrap up with final thoughts. You’ve been here before. So what say you What do you want to leave our credit union executive office?

Mitch Rutledge
I think, you know, leave it where we sort of started, which is, I think it’s up to the movement to really guide the NCUA and make sure that we don’t get sort of a one size fits all, something that’s either too nebulous or too restrictive, that, that there’s thoughtful discussion about what are the AI use cases and how they should be regulated and governed. So I think that’s a key thing that the movement as a whole should be doing. So all of the people and the key stakeholders should be thinking about that and and leveraging providers like us. I mean, I’m not volunteering, I’m I’m happy to be in one of the inputs. But there’s lots of great voices around this topic that I would hope that, you know, the NCUA and the leagues and all of the you know, the key stakeholders will consider if, as we consider this, and then I think for credit unions, as we said, you should be taking steps. And I see that the movement is doing this. There are lots of tests and programs that credit unions are doing, not just tests, right, that people are really rolling this out in meaningful ways to think about serving the members. And I think that we need to continue to do that. And I think the last part is like, this is the differentiator of the credit union movement, is you? We need to be sharing, right? So think about the work that we’re doing together and and sharing the use cases so the movement can can win together. And that’s, I think, a really powerful piece of this as well, is leveraging the power of the collective to move us all

Sarah Snell Cooke
forward. Absolutely awesome. Thank you so much for your time today. Mitch, appreciate it