A New Intelligence Layer for Community Lenders with Mike de Vere, CEO of Zest AI

Mike de Vere runs Zest AI, a company that has been applying machine learning to credit underwriting for over two decades, starting with some of the largest banks on the planet and now serving a large share of the credit union market. Since his last appearance on the show three years ago, Zest has expanded well past underwriting into fraud detection and portfolio management, tied together by an intelligence layer and a generative AI companion called LuLu. Mike makes a specific argument in this conversation: machine learning still makes the credit decision, generative AI makes the feedback loop faster, and the real advantage available to community financial institutions is a willingness to pool what they know.

What We Covered

• Zest today, from underwriting to fraud to portfolio management
• Why the intelligence layer is what makes an ecosystem
• Starting with Discover, Citi and Freddie Mac, then moving down market
• LuLu, named after a corgi, and what she actually does
• Safety and soundness as the first use case for most institutions
• Replacing quarterly reports that used to take weeks
• Peer benchmarking versus building your own data lake
• Collective intelligence across 2,000 credit models in production
• Why generative AI has no role in making the credit decision
• Shrinking model refit cycles from 18 months to daily evaluation
• Zest customers versus non-customers on growth, delinquency and efficiency
• Cash flow underwriting, and why generic national models fail
• Zest Protect and fighting AI-powered fraud with AI
• The two objections that come up most in sales conversations
• Takeaways from the IQ AI Lending Forum in Santa Fe

Key Takeaways

• The performance gap is measurable. Comparing Zest customers to non-customers across 2024 and 2025, Mike says his customers grew 16 times faster, ran roughly 20 points lower on delinquency, and were 501 basis points better on efficiency ratio.
• Generative AI belongs around the credit decision, not inside it. Zest still uses supervised, locked-down machine learning models for underwriting, because a regulator will ask you to explain the decision. What generative AI changes is the speed of evaluation, from an 18-month refit cycle to daily.
• Comparison is where the value sits. A lender looking only at its own data lake has visibility on itself and nothing else. LuLu is built to normalize performance data across institutions so a chief lending officer’s instinct can be checked against thousands of real policy instances rather than one career’s worth of experience.
• Community lenders have a structural advantage they underuse. The credit union industry holds roughly $2.4 trillion in assets. If it acted as one institution, it would be bigger than Wells Fargo, and unlike the big banks these institutions are actually willing to share.

About Mike de Vere

Mike de Vere is the CEO of Zest AI, the AI lending technology company that has been doing machine learning in credit since well before AI became a standard fintech conference track. He came to Zest from a career in data and consumer insights, with leadership roles at J.D. Power, The Harris Poll and Nielsen. Zest now touches $5.6 trillion in assets under management, and by the end of this year expects one in three credit union members to have their consumer loans decisioned with its technology.

Cleaned Transcript

Mike (00:10): It was so important for us to be able to make our AI accessible for all. So that whether you’re talking about a small credit union in Molokai, Hawaii that does hundreds of applications, at the same time it’s accessible to Citi on the other side. And so we democratize AI, but that doesn’t mean the human is totally ready for it. And also you have to cut through the clutter of the industry, because everybody adds AI to their brand. Everybody’s now an AI expert, and so it becomes quite difficult for a financial institution to kind of weed through that and identify where the snake oil’s at.

Peter (00:46): This is the Fintech One-on-One Podcast, the show for fintech enthusiasts looking to better understand the leaders shaping fintech and banking today. My name is Peter Renton, and since 2013 I’ve been conducting in-depth interviews with fintech founders and banking executives. My guest today is Mike de Vere, the CEO of Zest AI, a company that has been at the forefront of AI-driven credit underwriting for more than two decades now.

This is Mike’s second time on the show. The previous episode was three years ago. In our conversation, we talk about how Zest has expanded beyond underwriting into fraud detection and portfolio management, and the intelligence layer that ties it all together. We spend a good deal of time on LuLu, their generative AI companion for lenders, and how it is democratizing AI for credit unions and community banks.

We also discuss why machine learning still makes the credit decision while generative AI speeds up the feedback loop, the state of the American consumer, cash flow underwriting, the rise of AI-powered fraud, and his recent IQ AI Lending Forum in Santa Fe. Now let’s get on with the show.

Peter (02:11): Welcome back to the podcast, Mike.

Mike (02:13): Good to see you, Peter.

Peter (02:14): Good to see you. So I was looking back, it’s been over three years since we last had you on the show, and a lot has happened since then at Zest and also in all of fintech, which we’re going to get to. But first, why don’t we just get started by how you describe Zest today?

Mike (02:31): So listen, Zest, we started mastering the use of AI over two decades ago, and we were solving an important problem around credit underwriting. Our customers have asked us to modernize their lending ecosystem, and it’s in new areas. And so it’s not only looking at underwriting alone, whether it be personal loans, credit cards, HELOCs, auto, et cetera, but it’s also detecting fraud. And with fraudsters becoming more and more sophisticated, our customers are needing better tools.

And then the final area is around portfolio management, which is an area we’re going very deep in. And so if we think of the application and mastery of AI in underwriting, in fraud detection and portfolio management, well, actually, to create an ecosystem, you actually have to have a feedback loop. So that’s where our intelligence layer comes in. And so those are the ways that we’re supporting both large financial institutions as well as some of the smallest ones here in the US.

Peter (03:39): Right, so are you focusing on credit unions, community banks, or is it just across the board?

Mike (03:44): Yes. So in short, we actually started our journey with some of the largest banks on the planet. Super proud of it. So Discover Card, Citibank, GSEs, little GSEs like Freddie Mac. And geez, those took years to win and years to implement, but they really hardened us and hardened our technology, and really pushed us from a compliance and regulatory perspective. But you know what? There’s not 5,000 of them sitting around. And so we had to figure out how to make AI accessible for all. And so we migrated from some of the largest banks, which we’re still servicing, and mid-sized banks, and really took off in the credit union space in particular, and the community financial institutions.

Peter (04:31): So let’s get into the platform, and I want to dive straight into your product called LuLu. I’ve seen you describe it as a purpose-built generative AI platform for lenders, essentially a ChatGPT interface pointed at a loan portfolio. Maybe describe the platform, and what can a lender do with this platform that they couldn’t do before?

Mike (04:55): So I would describe LuLu, which in every interview, every time I say LuLu, it just makes me smile. So LuLu was actually named after a corgi that runs through our office.

Peter (05:05): Oh, that’s great.

Mike (05:07): So it doesn’t matter the client. Yesterday I had an opportunity to meet with Donna Bland up at Golden 1, and I’m going through LuLu capabilities. She stops me, she’s all, I love this name, I love the name.

So LuLu is a lending intelligence companion. And what she does is she takes publicly available data, imagine NCUA call reports, bank call reports, information from the Fed like inflation rates, et cetera, things of that nature, HMDA data. But she also then brings it together with a financial institution’s performance data. That may be data from their daily applications, their loan applications they’re doing, or their core banking data. That then gets married together with Zest’s normative data set.

So we are sitting on a lot of data. Peter, in the last three years, it might surprise you to know that right now we’re tracking at $5.6 trillion of assets under management where our technology touches. By the end of this year, it will be one in three credit union members will have their consumer loans decisioned with our technology. And why that matters for LuLu is it’s more information. She’s just becoming smarter.

And so LuLu sits at the epicenter of this industry, bringing together performance data, but normalizing it such that that then financial institution can compare their performance, evaluate their strategies, assess the health of their portfolio, not blind and narrow to just what their financial institution does, but having a broader view.

Peter (06:46): So then once a financial institution implements LuLu, what is the typical first use case that they really want to get into?

Mike (06:54): I’d say the first use case is around their financial health, so safety and soundness. And so currently most financial institutions will buy a report using old technology or dashboards from the 90s, and it’ll usually be months and months late after the quarter is done. What we’ve done with LuLu is we’ve created direct APIs into the various publicly available data sets. So in the case of credit unions, it would be the NCUA call report.

What they would be able to do is to evaluate themselves in context of their peer sets, in context of how they’ve been performing over time, at a much faster pace. But what’s really cool is that it doesn’t take a business analyst. Peter, if you and I can have a conversation about we’re wondering about the credit quality of our organization and how it’s been trending over time, you just ask LuLu. That doesn’t require any skills in Python or any other programming language. It’s just natural language questions.

There’s a very large credit union here, a top five credit union in the US in Southern California, where the CEO, it used to take them weeks to get reports on their safety and soundness. He was actually able to sit down with his business support person, and in 10 minutes they were able to recreate those reports that are automatically generated every quarter close for them.

And so it’s a massive efficiency gain. It helps them to be far more agile as an organization. And I think from my perspective, it helps them stay ahead of what’s next. Now there’s hundreds of other applications, but that’s usually where a financial institution starts.

Peter (08:36): I believe it was May 2025, right, when you launched LuLu?

Mike (08:40): That was actually one of the modules, LuLu Strategy. She was launched the year before that. LuLu has been out in the marketplace. I’d expect by the end of this year that we’ll have near a thousand credit unions using LuLu. And so we’re really excited about her traction within the marketplace.

Peter (08:57): So then now that a credit union or a bank has been using LuLu, might have been a year, eighteen months now, some of your early adopters, how is that changing their approach to not just monitoring their existing portfolios, but how they think about new business?

Mike (09:14): Well, it allows them to be more proactive. And so instead of waiting and looking in a rearview mirror that’s six to nine months behind, they’re now able to understand adjustments that they need to make in the marketplace to be that much more competitive. They can then lean in further and start identifying growth targets. And so they could be from a member perspective, but also it could be things like, who should I be looking at as far as a merger and acquisition target that’s a good fit with our financial profile?

And so she really helps you lean in ahead. And I think from my perspective, we’re trying to democratize the use of AI across the lending ecosystem. It should not be held for just a large financial institution that have hundreds of resources assigned to this specific topic.

Peter (10:05): So it’s interesting, you’ve got sort of this economies of scale benefit too. I’m curious about, it’s been, I said 20 years of data now, and I imagine just that alone is enough to really make it beneficial for a lender of any kind to kind of compare themselves. How do you sort of mix that, the soup of all the data, external and internal data? If someone wants to compare themselves, I imagine you’re going out with your own data and external data, right?

Mike (10:36): Yeah. And so comparing is where a significant amount of value comes. So I think that the mistake that many organizations, even today, are making is that they’re trying to go it alone and just look at how they’re performing. And so they’ve developed their data lake, and you know what they have? They have a view only on themselves. The real value is you don’t live in a universe by yourself. You live with others and you compete with others. And so you have to have visibility on what’s going on in the marketplace.

And so LuLu sitting at the center, really our drive there is to support collective intelligence. In particular, community financial institutions are great collaborators and they’re often willing to share. And so if LuLu can sit at the center and drive this collective intelligence, you can understand something as basic as, well, delinquencies are going up right now, do I need to make a credit policy adjustment? Let’s look at debt to income.

Now, what would happen in the past is the brilliant chief lending officer who has 30 years of experience would say, I know what to do, because back in 1998 this is exactly what I did. And it’s only from their experience set, their universe of the number of financial institutions they work at, the number of experiences. What LuLu does is it opens it up. And in the case of credit policy, we’re talking about, we’ve got near, what is it, near 2,000 credit models in production. We have over 2,000 credit policy documents.

So LuLu’s ability to do scenario building, having seen over 2,000 instances of this application of maybe this particular DTI policy, it’s on steroids. And so LuLu’s not replacing the chief lending officer or the VP of lending. She’s enhancing their human capacity. So humans stay in the loop. It’s not about her replacing someone.

Peter (12:30): And is that what you’ve seen in practice, like at the banks that are implementing this, they’re not sort of laying off their junior analytics people?

Mike (12:37): Yeah. It’s really about enabling them to do work faster, to be able to do more with less, because community financial institutions, many of them aren’t growing their workforce by twenty percent, but yet they’re growing. And so LuLu comes in and helps them do that. I mean, heck, even here at Zest, we’ve been able to flatten out our resources while at the same time having explosive growth. And that’s just being thoughtful with the applications of tools like LuLu.

Peter (13:06): Interesting. I was also thinking as you were talking that when you’re launching a new product, for example, say someone wants to launch a commercial credit card or some sort of SBA loan, or any kind of personal loan that they might not have done before, that I imagine you can not only have your own projections, but you can see, well, I’m launching a new personal loan program, for example, other banks that have launched or credit unions that have launched this program, what were their metrics like in their first three months? And so you’re not just comparing, I could see it becoming so much more effective for new products, right?

Mike (13:40): Have you been sneaking around the Zest labs? We’re not public about any of that stuff, Peter, come on.

Peter (13:47): I promise I haven’t.

Mike (13:52): Yeah, no, what I would say to you, that is a brilliant idea. And if an organization comes up with a product like that, man, is it going to take off.

Peter (14:00): Okay. Look forward to that. Look forward to that. Okay, so let’s talk about underwriting specifically, because that’s sort of been your bread and butter since the founding of the company. I’m just wondering, how has your underwriting changed over the last five years, shall we say?

Mike (14:16): Well, let’s first start with, the argument has been won. Machine learning underwriting is far more accurate than traditional credit scores or industry scores that are out there. Consumes more data, applies better math. And when we talk about outcomes, when I asked LuLu a question before this interview, I said, hey, compare all Zest customers to all Zest non-customers, 2024 versus 2025 outcomes. And we grew 16 times faster than non-AI adopters. We had roughly 20 points lower in delinquency rates. And the efficiency ratio was 501 points higher. Efficiency ratio, so for every penny that you’re spending, what are you getting in return on revenue? 501 basis points. And that’s automation. And so that argument has been won.

Now, specifically about generative AI. From my perspective, and this is why we talk about this loop in this lending ecosystem, is that it helps you adapt faster. So when our first model launched six or seven years ago, I think it was with Discover, we would launch a model and we would evaluate it, and it took us 12 to 18 months to deploy the next model, the next iteration, and try to learn from it. With generative AI, we’re actually able to shorten that cycle down significantly. So we’re evaluating models on a daily basis and looking at, if there’s been a shift in the marketplace, do we need to make an adjustment? Do we need to have a refit for that model and put it back out into production?

And so from my perspective, generative AI does not have a role in actually making the credit underwriting decision. You have to have supervised, locked-down machine learning. But where generative AI can be is it can be that entity that is evaluating your performance, so that you can quickly adjust and adapt to what’s changing in the marketplace.

Peter (16:10): Right, right. So the models are still there, there hasn’t been a massive shift, but it’s the reaction to the models that’s changing.

Mike (16:16): You’ll get shut down, Peter. You go ahead and deploy a Claude decision engine, and then when the regulator says come and explain it, and you’re all, I got no idea. So no, we do supervised machine learning models, but there is a role for generative AI, and we’re happy to have led that effort with community financial institutions. We’re excited with our engagement with both the regulators as well as the government on this topic.

Peter (16:43): Okay, so I’ve seen you talk about this concept, and you sort of touched on it, you called it the thriving lending ecosystem. So let’s just unpack that for a little bit. What does it mean for a lender’s credit infrastructure to be thriving?

Mike (17:00): So when we think about a thriving lending ecosystem, it’s understanding that there’s connectivity between each of these touch points. Whether you’re talking about member acquisition, maybe you’re pre-screening or pre-approving, the handoff to actually underwriting a loan, while at the same time you’re trying to detect fraud, do an estimation of income, screen to make sure this person is who they say they are.

And then once you actually issue a loan, it gets put together into this overall portfolio. And you actually have to evaluate the health. And then you need to maybe do credit line adjustments on those as well. And you know what, you don’t want your members or customers to end up in collections. So you want to also evaluate your portfolio to try to find individual members or customers that may be at risk. It’s understanding that all of these things need to be connected.

And so in the beginning we had this Ferrari that we created that was AI-driven underwriting, and it was highly accurate and highly performant. But what we realized is then it went to that next step and it ran into old tools. So the old set of tools in the fraud space, or the old set of tools in portfolio management, or loss reserve calculators, or credit line management. But AI has an application and is really part of the core infrastructure of this entire lending ecosystem. But for it to be thriving, you have to identify the connection and feedback points between each of these. That is how you modernize your lending ecosystem.

Peter (18:35): Okay. So I want to talk about the state of the consumer, because you’ve got a great window here. You’ve got all this data that’s coming in. What are you seeing in the data? What are your customers telling you about the state of their portfolio, and therefore the American consumer?

Mike (18:49): Well, I mean, the old approach to credit risk and evaluating with industry scores has failed to keep up with the changes in the US. I mean, heck, if you think of, what are they on, FICO nine or ten, something like that? There have been more releases of Fast and Furious movies. And I’ve got to think, I mean, well, who doesn’t love Vin Diesel racing around? I do, I mean, that’s great. But it’s not as important as providing access to responsible credit for American consumers so they can fulfill their American dream. So it’s failed America, hands down.

Now, when we think about it from a consumer perspective, I think why this modern lending ecosystem is so critical is that things are rapidly changing. You could have, regardless of the administration, make a comment about increasing tariffs on auto, and you immediately start depressing demand for cars. Interest rates staying high. You’re changing complete balance sheets for all the financial institutions out there by one comment.

And so from my perspective, I think it is not a nice to have. It is an imperative that financial institutions do modernize, because the rate of consumer change, behavior, opinions, it’s almost like that 24/7 news cycle. It’s like we’re now in that place with the financial industry where we need to just have our constant pulse on what’s going on with the average consumer so we can support them.

Peter (20:24): Okay, so on that, cash flow underwriting’s having a moment. It’s been going for a long time, but it’s become more popular these days because of that. People want to get a real time look into the bank transaction data of their potential borrowers. So what is Zest’s view on cash flow, and where do you kind of sit in relation to this movement that more lenders are looking to incorporate that type of data?

Mike (20:48): So we tackled this issue many, many years ago. And we do believe that cash flow underwriting does have signal for certain product lines. But I actually would take a step back and have more of a philosophical discussion on it, which is, is it like Burger King where you can have it your way? And so when you build an AI model, they shouldn’t all be the same, because you’re not solving the same business issue.

And so if I’m discussing with a financial institution here in Southern California, the data assets that I’m going to bring to bear when I build that model are going to be different than the ones, let’s say, at the financial institution based in northern Mississippi. I’m going to bring different data assets in. It might be just credit data, it might be credit plus cash flow data, it might be other alternative data sources.

And so the cool part about our tech is we don’t come in with this generic national model, which some other providers try to do. They try to sell you generic AI, because they believe that New Yorkers behave the same as people from Alabama, behave the same as those from Kansas City. Well, that’s a bunch of hooey. We know that the data doesn’t support that. And so you actually have to have a thoughtful exploratory data analysis to evaluate what model to put in. And so many of our models today will include cash flow data because it does add signal, in particular for things like credit cards. And so we do add it, but we don’t add it just for the sake of saying we have it because it’s the cause célèbre of the day.

Peter (22:23): So I want to switch gears and talk about fraud, because there is more fraud being committed now than ever before with the rise of generative AI. I just wrote an article about it, talked about how it’s gone from being this craft to a factory operation, where it’s just done at scale. And so it’s much more important to have your fraud fighting capabilities at an optimum level, because there’s just so much volume going on. So tell us a little bit about what kind of fraud patterns you’re seeing today, and how does your AI-native approach help the outcomes for lenders?

Mike (23:00): Yeah. I mean, it’s a really difficult time to be a lender with the advent of generative AI and its application towards fraud. And you’re spot on, Peter. There’s not only industries, there’s towns that are set up with the specific thing to actually attack financial institutions here in the US. And the advantage that they have is there’s no regulatory, there’s no compliance. Their ability and willingness to take risks, to try new different approaches, and to apply the latest forms of math, their ability to share information across this town or this collective in the dark web. So they are staying way ahead of financial institutions.

That is why, in our client advisory board, this was the top thing they wanted us to tackle last year, was to provide a more advanced tool to fight AI with AI. And that was the advent of Zest Protect, which has three different layers. So first off is you’re evaluating every loan application that comes in to understand if there’s fraud, is there a potential for fraud. And it’s not a binary yes-no fraud, it’s probability that it’s fraud. But at the same time, you’re evaluating income estimations, you’re screening for income, you’re screening for ID. You need to have those three together.

And so from our perspective, that’s a nice start. But what I can tell you is that my income model, I actually have 10x the data in a 60 day period that I’m training my income model on. It is getting smarter every month that I go, because as we scale our organization and grow as a company, I can bring in more information. I can bring in more actual estimations of income, or actual experiences of fraud, and train this model. So maybe the average fraud provider has a generic fraud solution that they’ve launched two or three versions of. We’re constantly looking at evaluating our model and retraining it to make it smarter and smarter and smarter, so our customers can stay ahead of what’s next in fraud. And so this is their first line of defense.

Peter (25:16): So I’m curious about your business and when you’re going out and pitching to new credit unions, community banks, regional banks, because it feels like your scale kind of sells itself, as far as the fact that you’ve got so much more data than anybody else. But what’s the objection that you hear most often when someone is evaluating Zest, and what does it take to overcome it?

Mike (25:42): Yeah, exactly. Not now. It’s the competing priorities. I need to implement a new LOS. I’m in the middle of a core conversion. Many of these CFIs are looking just to update the core infrastructure, and so they don’t want to put a Zest into an old infrastructure. They want to wait for that to happen. And so that would be a big objection that we run into. But we do work hard to earn as many deals as we can and try to meet customers where they’re at. But I would say the main objection is that.

And then the second one is fear of change itself. And so it’s a human component, is that you have an individual, Peter, that has been doing it the same way for decades. And so to get them over the hump to say, you know, there is a better way, that these 25 credit policies that you’ve had, well, 15 of those policies are actually answered within the AI model, so you only need a few left. And it’s a big change management process from a human perspective.

We can show them all the outcomes. I shared with you the overall outcomes for Zest compared to non-Zest customers. I mean, the numbers prove out. The juice is worth the squeeze. But sometimes the humans aren’t ready for the change. The worry for them, it will be eat or be eaten.

Peter (27:05): The distance between the laggards and the innovators is just widening every day, is how it feels like.

Mike (27:11): But that’s why it was so important for us to be able to make our AI accessible for all. So that whether you’re talking about a small credit union in Molokai, Hawaii that does hundreds of applications, at the same time it’s accessible to Citi on the other side. And so we democratize AI, but that doesn’t mean the human is totally ready for it. And also you have to cut through the clutter of the industry, because everybody adds AI to their brand. Everybody’s now an AI expert. And so it becomes quite difficult for a financial institution to kind of weed through that and identify where the snake oil’s at. And choosing the wrong provider could be deleterious to their overall financial performance.

Peter (27:54): So I want to talk about your recent IQ AI Lending Forum in Santa Fe. You were kind enough to invite me to MC, which I very much appreciated, and it was a really fascinating couple of days down there. Now that you’ve had some time, it was about three months ago now, reflecting on the event, what were some of the key takeaways, would you say?

Mike (28:14): Boy, there are a lot. I think the main theme of collective intelligence and the need for community financial institutions to come together to be competitive, to stay relevant with consumers, is critical. And so if we agree that CFIs, one of their greatest strengths is collaboration, having an event like that that was curated specifically for some large financial institutions, mid-sized and even small, to come together to share information and to see products that were set up to bring them together so that they could work as a collective.

I mean, the example that I talked about was, if the credit union industry, I think it’s what, $2.4 trillion, something like that. I mean, if it behaved as a financial institution, it would be larger than Wells Fargo. And so that collective intelligence, I think, is critical.

I think the second piece is human in the loop. So I am that AI CEO that is not going to sit here and tell you AI solves everything. That’s not the case at all. Because you have to have a human in the loop, whether you’re talking about the model build itself, but most importantly, setting up the purpose and intention when you’re going to deploy AI, whether it’s a chatbot on the front end or you’re trying to evaluate your collections team using AI. What’s your intention? What’s your intention? Is your intention to serve the community and drive business outcomes? You can have both. Or is it just to drive simple balance sheet returns? And so I think that purpose-built, human in the loop, I think was another one.

I’ll tell you just a random one, but it scared the living heck out of me. So Dr. Omar, he was the chief AI officer of NASA. It was so, it rated at the top. I mean, it was like, you know.

Peter (30:04): I loved his presentation.

Mike (30:09): I wasn’t even on the page. He was like so far out there. But what an insightful leader. And he talked about the risk with generative AI and deep learning of humans relying on it too much, and the cognitive decline that can follow within a decade to the entire human race. That was terrifying to me. And as a father of five kids, it really made me think about my responsibility to sit down with my kids and talk about human in the loop still, in that when you have an assignment, it’s critical that you’re using generative AI, so a tool like Claude or Gemini or ChatGPT or Copilot, as an editor, as a partner, but not as something that just creates content for you, so you don’t actually have to do the thinking yourself. And so that for me as a dad was a huge takeaway.

Peter (31:02): And the thing that struck me being there is the collaboration you talked about, was everyone was so willing to share and be open. And I felt like I didn’t think I’d been to an event quite like it, where there was so much openness and sharing involved. So I think that was one of my key takeaways. I thought it was amazing.

Mike (31:19): Thanks for it, it must have been the MC. I mean, to curate it down from the hundreds of people that wanted to attend to the hundred or so that did attend, we were intentional. We didn’t want it to become our entire client roster showing up and we have five hundred, six hundred credit unions and banks there and things like that. Those are great, they have a place. But this one we really did want it to be a forum where people could have discourse and share and help.

Peter (31:46): I think you certainly achieved that. So anyway, last question. What’s next for Zest? What are you working on? Where will you be in three years’ time?

Mike (31:56): Where will I be in three years’ time? So my great hope for Zest is that we continue to find points within the lending ecosystem that can be modernized. And so what’s coming up next beyond fraud that launched this year is really in the area of how one manages your portfolio. And so it’s the portfolio health. It’s evaluating the health of your members, identifying members that may be at risk to run into collections. It’s looking at credit line management, whether you’re talking about credit line increase or credit line decrease. And then loss reserves, the mother of all equations. Like, how do we do our CECL calculation and have it be completely transparent, that’s tailored specifically for our institution?

And so those are problems that we currently have out in the marketplace today, or solutions that we currently have out in the marketplace. But we do a V1, we do a soft launch, we learn from our customers, we co-create with our customers. So expect that to be coming next. There are going to be so many new LuLu modules. I’m just thrilled with next year. We have a module specifically around the member itself, the member’s health, the share of wallet, opportunities for cross-selling and marketing. The applications for LuLu are truly endless.

Peter (33:15): That’s a great place to leave it, Mike. Always great to chat with you. Best of luck, and thanks again for coming on the show.

Mike (33:22): Thanks, Peter.

Peter (33:29): The idea that keeps rattling around in my head after this conversation is around collective intelligence. Mike pointed out that if the credit union industry, around $2.4 trillion in assets, behaved as a single institution, it would be larger than Wells Fargo. Community lenders have always competed against scale, and his argument is that their real edge is a willingness to share and collaborate in a way the big banks never will. If a platform like LuLu can sit at the center and turn that openness into a shared data advantage, it could genuinely reshape how smaller financial institutions compete. I find that a hopeful thought. Anyway, that’s it for today’s show. If you enjoy these episodes, please go ahead and subscribe, tell a friend, or leave a review. And thanks so much for listening.