Welcome to Breaking Banks, the number one global fintech radio show and podcast. I’m Brett King and I’m Jason Henricks. Every week since 2013, we explore the personalities, startups, innovators, and industry players driving disruption in financial services.
From incumbents to unicorns and from cutting edge technology to the people using it to help create a more innovative, inclusive, and healthy financial future. I’m J.P. Nichols, and this is Breaking Banks. Let’s play ball.
AI is simultaneously taking over the world and disappointing in its ability to take over the world depending on what publication or podcast you are looking at. No doubt we are in the first innings. Today, Derek Higginbotham, CEO of First Electronic Bank, Matt Bocinek, CEO of Avant, and John Sun, CEO of Spring Labs, and I talk about why everyone needs to be getting up to bat with AI, no matter the size of institution you work for.
Hey, next week, Spring Labs is hosting the AI Native Banking and Fintech Conference, and they have a powerhouse lineup. Register at conferences.springlabs.com and use codes springlabsxbreakingbanks for a discount. You can find both the link in the code in the show notes.
Gentlemen, thanks for joining today. I don’t know which is the hotter topic, the amount of regulatory consent orders and proposals dropping or AI. It feels like they’re fighting for who’s going to be center stage, but so let’s stick with AI for right now, and maybe we can get in on some of the guidance that’s come out around this.
What I want to start with is Andreessen Horowitz recently released this great graphic just showing the pace of AI and its adoption and impact far exceeds, like if you look at it, it really looks kind of like wave starting with the PC versus the internet and APIs and data. AI just has this super steep curve, and I don’t know if this is controversial to say or you’d agree with me. It also feels like we’re going through the hype cycle faster than we ever have before, like AI is going to change everything.
AI isn’t changing anything. AI is going to replace all the jobs. AI is a copilot, right? Do you have a similar sense? You’re the practitioners and the ones on the front line.
Are your organizations feeling that whiplash? Derek, why don’t we start with you? I think we hear the whiplash more than we engage in it so much internally. I think in the banks right now, we’re all feeling pretty cautious about doing dramatically new things. So everybody personally got really excited when they started seeing the AI tools coming out not too long ago.
But I don’t think there was ever really a thought, we can just go, let’s just blow everything up and just put it all on AI. But in friends that run other companies in kind of the periphery of what we do, definitely see big peaks and valleys in their energy around diving into the AI. Yeah, I would say, Jason, our baseline internally, and I think this is what applied to most, is that the technology is here to stay, right? But obviously, the technology is evolving so rapidly.
I think it’s very, very difficult for anyone to predict with a high degree of precision, kind of where that is going to go over the next 12, 24, 36 months. I’ll say, kind of given our experience, I peeled our CIO and co-founder kind of off of the day to day about 18 months ago to start to build out our internal roadmap. I think kind of where we landed really after a lot of iteration, a lot of whiteboarding is really not to the point you were making kind of using the hypercycle as an example.
It’s really, you could cast such a wide net with this technology. The key, at least from what we have observed and identified, is trying to first narrow down into one, two, or three real applicable use cases in the business, kind of where you could deploy that technology, take some learnings, iterate, use that as the foundation, which we have done to now build out our larger strategic roadmap. Because as everyone here is well aware, you can kind of like really boil the ocean with this type of thing, and you kind of walk in and there’s like this nirvana on the whiteboard where it kind of could solve every single enterprise solution.
And practically speaking, that just doesn’t work, right? So again, our experience is kind of take really kind of small incremental steps, learn, iterate, and then build thereafter. And we’re seeing some really, really early signs of success. And you really see that a lot with new tech trends.
It does seem like there is kind of a grace period at the beginning of every single new tech trend where companies are just willing to invest unlimited resources without nearly a thought for ROI or returns or anything like that. And we’re very much in that phase of AI right now. So I mean, I think a good kind of reminder for everyone out there is like this has happened probably like across the last four or five tech cycles.
And a lot of people invested a lot of money and didn’t really get the yield they were looking for. I think it’s important to develop the organizational muscle because I totally agree, Matt, like the technology is here to stay in some form or other, whether it’s kind of the LLMs that we see today or some iteration of that in the future. I think it’s important to start growing that organizational muscle immediately.
But I think it’s also important to just look at the activities that you’re choosing to take on and figure out what the ROI behind those things are. And, you know, it’s one of these areas where it could be massively impactful to some activities and not at all impactful to others. And what I’m seeing kind of out in market is a lack of kind of discrimination kind of between the two possibilities.
Yeah. And maybe add one more thing, Jason, you said this on the kind of on the onset, bifurcating between kind of cost reduction and value creation, I think is really critical. And particularly in this environment, when, as we all know, capital is not free any longer.
The easiest trap is to say, hey, we have this big line item, right? We’re going to take out that wide item completely or by 75%. And candidly, like generally speaking, that’s typically the lowest hanging fruit, right? And a lot of our time and attention has been on identifying real insights that could be used kind of broadly across the organization to create value. There certainly is an efficiency gain from an expense or spend perspective.
But I think that that bifurcation is actually really, really critical. Derek, you know, banks are notorious and loving cost reduction, right? Efficiency ratio is part of our official religion within community banks in particular. And we love that certainty when we say, how much cost can we take out? How are you guys in Fib thinking about, you know, the possibilities as you’re building out a strategy? You know, when I was just saying really resonated, the first thing that we tackled was absolutely not a cost reduction thing.
It was insight creation. I think right now, like the call to action for the banks doing a lot of this FinTech work and these partnerships is much more, we need to find as many interesting insights as we can as fast as possible, much more than cost reduction. So the thing that we’ve done for starters is adding expense or adding an agent that happens to be an AI agent to go out there looking for more insights.
I think John was right too. Like we felt like we had to do something. This technology is clearly super powerful.
There’s an enormous amount it can do. We need to understand it so that we can start thinking of more use cases, but we weren’t going to start, we weren’t going to try it in a way that was really going to sub plant any work to start with. It was more, let’s bring in this more resources to go sit next to our traditional resources to find more insights.
Matt, given the pace of change, you had mentioned your six and 12 month plans. How are you baking in just the speed of change? Ah, that’s a very great question. And that is evolving by the day.
We have a small executive group that meets weekly to discuss kind of where we’re at, you know, where we are going. And again, I want to once more kind of reinforce our strategy, and it might not be applicable to everyone that we’ll be listening in. But again, we started very, very wide.
We narrowed it down to one, two, three. Fraud, a fraud area, complaints, compliance area, and kind of the operations area, right? And we’re building in such a way, again, going back to the point, and I’ll answer your question here, Jason. The speed of change and the pace in which this technology is evolving, once more, I think is very difficult to predict your 12, 18, 24 months out, right? So we are building out a flexible enough foundation, right, where we could go, you know, build and iterate over here, right, internally, where we could potentially buy off the shelf, and really ensure that the underlying data is structured in such a way where it’s exceptionally fungible, right? We believe that the data is the most important thing.
We’ve been making a ton of investments there. If you have a very sound sort of infrastructure, kind of all-encompassing around data, we think we could keep up with that, you know, speed and or pace of change over time. That’s a great point, because I think that’s really what AI adoption, transformation, and exploration does for you today, right? Like it gets your team aware of this now completely different architecture that needs to, at some point in the future, fit into what you guys are doing.
I think it opens a lot of doors just in terms of, even at a very basic level, even just doing a couple AI projects, exposes your team to this architecture, and makes sure that as they’re architecting the platform for the next, you know, two, three, five, ten years, they’re not locking themselves into a pattern that doesn’t allow for kind of the future incorporation of AI. Yeah, and the AI development internally, and John and Derek, you know this full well, was fascinating because we’ve been doing this for almost 11 years, right? Almost 12 years, actually. And if you think about the way this business has evolved, and we built some technology, we iterated on a process, we built some more technology.
Generally speaking, the underlying data infrastructures have been relatively constant, but, you know, we had assumed as we kicked off this project that we could kind of just bolt on some AI technology on our existing process, and it would yield significant efficiency gain. But the reality is you also find yourself having to rethink your existing processes, right, that compliment the AI technology. And that forced us to entirely take a major step back at some point during this journey to, again, kind of pivot to where we’re at today.
Yeah, that’s totally been our experience, I think. And it’s been really productive internally to drive more and more of the staff and the stakeholders internally to think about, how am I going to solve problems with data? And then how am I going to have to make sure that the data exists here so that I’ll be able to solve those problems? Most of the work that we’ve done is thinking about how to align people around the data acquisition and aggregation, and how are you then on the line going to be solving your problem with this data? Oh, Derek, you have no idea what you just stepped in as one of the newer members of Alloy Labs. Wait till you get into the Data and Analytics Center of Excellence.
Like, this is the hot debate. So I want to debate part of it here is, with that group, we find there are two struggles. One struggle is cultural, which is how do we change a culture to be data-driven? Because historically, we don’t know what that is.
We think of the world as reports, not hypothesis-driven. And I think AI actually magnifies that problem, right? It’s a new way of thinking. It’s not just writing a query in a search engine.
And then the second piece of this debate I want to talk about is always around how do we even get the data? Like, the data are hard and they’re messy and siloed, particularly in financial services. Why don’t we start first with how is it we need to change our cultures to not just be data-driven, but the impact of how we need to change our culture in light of AI being a co-pilot, non-branded co-pilot there. Sorry, Microsoft.
Yeah, well, this is something we’ve been working on for three years. I think it was clear in the way that our business model works that many problems need to be solved by tapping into data and letting the data tell you where something needs attention, as opposed to kind of just a human kind of moving through a process and procedure, getting to kind of a point of action. I’m actually really grateful for this project of implementing this AI tool to have people sitting around and actually thinking, okay, this is how I move through my day normally, and now I have this new tool.
Like, how am I going to shift? I think we just have to, for us, a lot of it is asking people questions in a way that will drive them to have to go query data, as opposed to just tick through their procedure. Instead of asking, okay, did you finish the procedure? It’s kind of asking, okay, what are the five insights that you got out of the data? So I think just putting people on the problem, putting people kind of on the task of, okay, go figure out from this data what the insight is that we need for your job to have been accomplished well. That’s super interesting, Derek.
We have a different problem statement, actually, Jason, as you might imagine, right? So sort of being pioneers in the category, we’ve been data first since day one. It’s really core to how we solve problems and how we solve problems over a decade. And so it’s less to do with adoption of data and analytics and using that to drive the business.
It’s actually, we have to rethink the way we do work internally, right? As you kind of use some of the AI tools, we no longer, and we’re still building towards this to be clear, but at some point, I think in the very, very near future, we’ll no longer need an analyst to build a query and pull the reports and segment the data for us, right? The way that we’re constructing kind of our infrastructure will just kind of prompt, right, this tool with all of the underlying data pointed to it. It then becomes a way of how you contextualize as a user of that tool to drive towards the right set of questions, right? Versus the sort of binary logic within the SQL code itself, right? And it’s pretty amazing what these tools can do. And again, it’s going to completely reimagine the way that we work internally.
And frankly, probably ultimately the skill in which we bring it to the business over time. You know, it’s kind of an interesting point. And it kind of just like reminded me of kind of a tangential learning as we were working with these AI tools is that sometimes it’s hard to predict ahead of time what the AI agents are actually good at and what they’re not good at.
There’s been several things where we were like, oh, this is a slam dunk. Just drop some AI in here and you’ll get like 80% productivity gains. We bang our head against the wall for like four months and just don’t get anywhere with it.
And then there’s other use cases like Matt, we were describing like, hey, can like, you know, can LLMs and these AI agents write queries? They’re surprisingly good at it to a point where I looked at it, I thought it was like, there’s no way. Like these things are not going to be able to figure out like how to construct queries in kind of a detailed way and pull like cohort analysis and like segment data and all of that. It was actually surprisingly good at it.
And we’re actually, you know, for Matt and Derek, I know you guys are early supporters and kind of adopters of Spring Labs, you know, tech. So I appreciate that. But also, you know, just coming down the pipe, we’re actually rolling out a feature that lets your analysts do exactly that within the platform instead of having to go hunt down, you know, specific types of complaints and specific types of queries.
You can just ask an AI co-pilot essentially at the unbranded co-pilot, sorry, Microsoft, on, you know, what the changing trends within your complaints are and to pull data and structure data around that. You know what season it is? Well, it’s not just pumpkin spice season, it’s conference season. And there’s a new show to add to your list.
The AI Native Banking and FinTech Conference is happening October 7th in Salt Lake City. And Breaking Banks is proud to be a media partner. Our listeners get a 25% discount when they register at conferences.springlabs.com and use the code SpringLabsXBreakingBanks.
Yeah, I didn’t write that code. The link and the code are in the show notes though. And this is an outstanding conference with an outstanding set of speakers.
It should not be missed. Now, complaints are a good segue into, we’re seeing a lot of AI applications within regulatory compliance, right? Complaints being, you know, one aspect of that. I’m curious for each of you, you know, how are you thinking about the regulatory impacts? I know in our conversations with regulators, there’s a bit of hesitation in terms of, and I think that’s natural with regulators, right? Is hesitation, but they also see some of the benefits.
Internal, you know, Derek, why don’t we start with you? And then Matt, like internal, how are your risk people thinking about your regulatory compliance and risk management of such a new technology? And then John, let’s switch to you about, you know, with your clients, what are the number one questions they’re asking and the misconceptions? Yeah, I think two directions to answer that. I think that clearly there are things around data security and discrimination that we all have to be careful about when we’re using these tools. I think it’s probably less interesting and novel question right now.
I think really what we’re using this tool for is to look for, we’re using a complaints tool. We’re using it to look for any potential regulatory failures. I don’t think that the regulators will accept that tool replacing standard, like typical consumer protection risk management structures for quite a long time until we prove it to them that this works a lot better.
And not only we prove it to them that we’ve, they have tools, whether we help them develop it or they do it themselves, that let them audit us constantly to make sure that the AI version is working better than the traditional version. I don’t think any of, the way we’re approaching this is, we’re not gonna turn these agents loose to replace existing risk management structures until the whole ecosystem has evolved enough that the regulators can run their exams knowing that they’re getting better, more reliable, predictable, consistent results than if we did it in a traditional way. So for now we’re thinking, okay, we’re deploying these tools as kind of sitting next to the traditional agents, to the human agents to find more risk.
But not trusting them to be the replacement for now. Yeah, interesting. We’re thinking about it actually very similarly.
I guess first, I think this is an important point, right? Each of us are reading some sort of headline weekly, that speaks to something to the effect of, I replaced my entire call center or I replaced my entire kind of SaaS tech stack. And in my view, I think that’s just attention grabbing more than anything else, right? The fact of the matter is, we operate in a very, very highly regulated environment. And so we have to be exceptionally thoughtful to your point around how we deploy new technology, whether it’s AI or anything else within the business.
We at Avant work very, very closely with our bank sponsor as we’ve been deploying of this technology internally and very similar to Derek, our approach has been the tool right now is a supplement to the agents kind of working various cues, whether it’s in complaints and fraud, right? And it enables that agent to go deeper and wider with the actual respective complaint and drive insights that we’ve actually started to filter into the business versus peer replacement. And again, I think how this evolves over the next, six, 12, 18 months from a regulatory perspective, certainly will dictate how we think about that longer term. Yeah, I mean, I think heavily regulated industry like the one we’re in, that’s a very sensible approach to take.
I think this has to be kind of a capability augmenter instead of creating different capabilities in the sense that like, if this is strictly better, like you’re still doing everything you’re doing today and you’re layering AI on top and now your agents are able to find more insights, get more efficient by targeting their day and targeting their activities better. And net is kind of better for your customers, better from regulatory risk perspective. I think that’s going to be great.
But if you’re kind of saying like, hey, I’m replacing some existing capabilities and this is just differently capable, like it’s better at some things, it’s worse at some things. I think that’s going to be probably from day one, tougher bill for a lot of the regulators and even internal compliance to swallow. So I think, like Derek said, the goal is really, let’s kind of use this, create additional capabilities for humans.
Let’s kind of deploy this in a human first model, if you will, to support and augment humans rather than replace them from day one. I want to push back slightly on that, John, because one of the things, it was a subtle comment made earlier by both Matt and Derek, but the ability to use this as a chance to rethink processes, even without the improvements of AI, we too often don’t go back and rethink processes because they’ve made it through an exam previously. Let’s not go through the headache of retesting and then convincing examiners of it.
But we end up with, we talk about tech debt all the time. I think we end up with process debt as a result of this, that maybe one of the biggest impacts of AI is it forces us to go back and do some rethinking. I’ll take that because that’s a great question.
It’s something that we talk about rather regularly. So I think what we want to prevent when you’re investing in any new technology, to your point, Jason, where you sort of get to an MVP state. MVP state creates some marginal benefit to the business, but the actual value creation over time is lost for whatever reason.
You move on to a secondary, tertiary MVP or what have you. The sort of objective of us in this sort of re-imagination in terms of the way we work internally is really simple. Does that MVP solution, right, to John’s point, that sort of complementing the current way that we work, ladder into this ideal state, right? And does it take us from kind of V1 to V2 to V3? What does that need to look like? What is the investment? What’s the implied return? And I think that, ideally, now there’s going to be some cases where there’s certainly this isn’t true, but hopefully that mitigates that issue where there’s just a significant sum cost or debt that we’re building into the business.
And you got the right people at the right time kind of zooming out, right? And rethinking the way that we work. And that’s the way we’re approaching it. Yeah, I wouldn’t disagree with anything anybody here said.
I think that John said earlier that sometimes we’re getting surprised by what the tools are good at or what they’re not good at. I think it’s dynamic enough that it’s going to be driving, pulling our attention, forcing us to engage with it for quite a while and think about how our people are going to work around it as we play with it. We’re definitely seeing certain kinds of insights popping up that we wouldn’t really have expected.
And we’re seeing situations where we thought all the insight creation would have been done and it wasn’t done, so. We could talk about this particular topic for hours. And John and I have had this debate and or discussion because I think Jason, it’s just not sort of thinking or rethinking the way that we will operate as a business, the way that our employees work internally.
You sort of have to zoom out and ultimately how will the consumer behavior shift over time, right? Will we be able to acquire consumers through the same channels? Yes or no. How will they behave with our tools, right? The sort of concept of kind of going into a Google search, right, is quickly kind of going away. And how will that sort of evolve over time? So again, I think it’s very, very difficult for any of us.
We’re sort of making a bet in a couple of different areas that this is the investment that we need to make. But I think this is something that we as leaders in the category need to continually just push and push ourselves around kind of where is this headed next and how can we put the business and our products, our solutions in the right sort of place or state to win and to ultimately have the consumer win at the end of the day as well. Like what Matt said about how we’ll be learning from how the consumer behavior changes, we’re also gonna be learning from the regulators, right? I think the regulators are watching this happen.
And we all have been in this phase of the last year and a half or two years where there’s just been enormous amount of regulatory pressure to improve how risk management and consumer protection happens in these programs. I don’t think that pressure is coming off anytime soon. So I think we’re gonna be sprinting for a very long time to kind of every quarter have to figure out new ways to find better insights to drive kind of more compliance, better consumer outcomes, and get more predictability, more safety and soundness in the system.
So I think we’re gonna be chasing that for quite a while. I don’t see that backing off. I mean, Matt, you as a category leader can push and say, hey, we need to invest and keep up.
John, I’m curious the FIs who don’t consider them the leaders, right? Or the service providers to the FIs, are you feeling a lot of the, we just wanna wait and see until some of this, like both the technology settles down and the regulatory compliance settles down and to Derek’s point, it’s like, hey, if I don’t do it, I don’t have to prove consumer protection in these other things, right? Like it’s a little, you know, see no evil, hear no evil aspect to this. What are you getting a sense from in the marketplace? You know, I think that would have been my intuition as well as like maybe the larger category leaders are leaning forward and the smaller guys sit back a little bit. That has surprisingly, again, not been my direct experience.
It seems like there is a lot of curiosity and a lot of willingness to invest kind of across the board. You know, some of our clients are very small institutions that you wouldn’t think of as, you know, folks that would be on the cutting edge of AI and leaning into AI. And yet, you know, they’re some of the most enthusiastic about figuring out what the capabilities of the new technology are, how to incorporate into their systems.
Now, obviously it’s not right for every single organization. A lot of, you know, I think the latest stat I’ve seen, I can’t source this right now, is something like 90% of the AI PLCs that’s happening right now won’t go into production in the future just because there’s a lot of exploration. But I think at the same time, you know, there are a lot of tools that these guys can start adopting today that both kind of checks the box for getting, you know, exposure to AI, to the organization, to start building that organizational muscle as well as provides, you know, immediate or very near term ROI on their investment.
And, you know, I always kind of recommend if you’re not, or even if you are, you know, if you’re not a heavy data science shop and tech shop, you know, partner with someone, work with a service provider or work with a product company like Spring Labs, the shameless plug there. And I think that’s a great way to kind of make sure that whatever you’re, you know, implementing within the system has been vetted and tested at other organizations and can provide kind of immediate ROI without you having to kind of take that risk of exploration completely blind. Yeah, maybe I would just add one thing, Jason.
I think this is important because, you know, despite the fact that, you know, we’re not a small startup at this point in our journey, like every other organization, we don’t have unlimited, you know, resources, whether that’s, you know, capital or people, which is why, you know, our partnership with Spring has been so fruitful on a two-sided way, right? They’ve built, you know, tools, you know, that we could, you know, seamlessly integrate into our environment. We’ve literally had one, one and a half people, maybe one and a quarter people working with Spring over the last, you know, couple of quarters to kind of go live. And so even if you’re, like I said, I was going even very early in the mid stage, understand that resources are even, you know, sort of more constrained.
I think ultimately at the end of the day, you sort of got to get into this, you got to get into this space, right? It’s inevitably going to continue to go in this direction. And I think there will be winners and losers, or it might just be too late to catch up. And so you’re always, you know, better, I think, in our experiences.
It’s just diving right in. Yeah. I don’t know if you listened to the episode where we talked about the AI digital divide, those who adopt early keep pulling further ahead, those who are the laggards, who think they’re going to catch up when it becomes more certain, will never catch up.
Good luck, yeah, for sure. Yeah, I mean, you saw that. I mean, again, like I keep pulling back to like the previous, you know, tech adoption cycles.
I mean, obviously I think history is a great teacher in terms of what happens in the future. Like you saw this with machine learning, right? That was like the last AI wave was more traditional machine learning for like underwriting, fraud prevention, collection segmentation, all of that. I mean, there’s kind of three distinct groups of finance institutions of FIs out there.
There were like the super early adopters like a bunch. And I remember we were like building infrastructure to deploy AI models for real-time consumer underwriting before there was even like Python libraries or before there was even like code for it. And you had to build really from the ground up.
And there were very small handful of those. And that was probably in this like 2012 to 2014 timeframe. And then you kind of had your mid to late adopters that were like rolling these things out after there was infrastructure probably in this like 2018 to like, you know, 2020 to 2022 timeframe.
And then you kind of have the folks were still on logistic regression models, you know, and maybe they, you know, get onto this at some point in the future. Maybe kind of the process is so embedded it’s difficult for them to ever do that. And I think you’re going to see similar things with AI.
And a lot of the early adopters reaped a lot of the benefit. And a lot of the mid adopters got some of the benefit. And then again, you have some organizations still on legacy technology.
Well, I think we’re at an interesting inflection point with the community banking in particular that this may, if you are one of those laggards, it probably could be part of the death now. And if you’re in that mid group, it may be allows you to sustain an unsustainable model a little bit longer. And that early adopter said it might allow you to actually cross the chasm and reinvent, you know, the business of banking in particular.
One last thought from each of you, what do you think the industry writ large, whether that is the AI industry of sorts, right? The tech techno folks who think about it or financial institutions, what don’t they understand about artificial intelligence? And Matt, why don’t we start with you? Well, I think we can take that question in a number of different ways. And I sort of hate to kind of go back to what I’ve been kind of harping on throughout our discussion. But again, in my experience here, what you don’t know is what you don’t know, right? At the end of the day and being very, very intentional around where you deploy that technology internally with the right level of guardrails, with the right partner or not, if you decide to build it internally and really stair-step your way into that deployment.
It does take some time and attention to get the underlying technology sort of dialed in, right? And I think it’s kind of never evolving or learning process, but this sort of misconception that you could kind of flip a switch and you’ll benefit overnight across the business is simply just wrong in practice. Eric, how about you? Well, Matt just took mine. I’ll just say it a little differently.
From the get-go, it was amazing kind of what the tools were spitting out, but we’d spent months and months kind of trading the models, refining what kind of outputs were useful and which ones weren’t. And so I think that there’s some magic to this, that these new tools are just kind of amazingly powerful and surprise us kind of every day that you play around with them. But also to really get the most out of them, there’s a lot of engagement, like a lot of very deliberate interaction between kind of the tools and the people back and forth to get them dialed in.
John, bring us home. Yeah, I mean, I think going off of what Matt and Derek said, but also just introducing kind of a new thought here, financial services, as you guys all know, is kind of a different space. For some reason, everyone kind of from outside the industry, from the tech industry always looks at financial services and they’re like, oh, there’s so many things we could be doing better if we were doing financial services, right? And I think the AI kind of trend is no different.
And we were all there at one point, right? I mean, I think before getting into financial services, we’d all kind of looked at the industry and said, there’s so many things we can use technology to improve. And then once you get into the industry, you realize there’s kind of a lot of challenges because of the structure and the regulatory complexity and the various moving pieces that’s not just like deploying an e-commerce site and selling widgets to consumers. So, I mean, a lot of startups are going to run into that and immediately fail.
And some from, you know, like all of us in the room here are going to embrace the badness and kind of like be a part of this kind of ecosystem. And I think it’s going to be no different from AI. I mean, we’re seeing a lot of these, what, you know, we call, what the industry rather I should say, calls like chat GPT wrappers, like GPT wrappers as applications.
And some of those things work pretty well outside of financial services. But in financial services, a solution that’s 60, 70% viable and accurate might as well be 0% accurate because it’s not going to get any adoption. So I think the hurdle to building a successful AI product for financial services is that much higher.
So I think that’s kind of one thing to kind of keep in mind as we’re, you know, growing in this space. I think I’m going to say one warning that’s also accidentally a plug for John. What John said was absolutely proved right for us.
I don’t think that it would be a very good idea to try to really develop one of these tools with people, with partners, with like tech partners that aren’t really, really well-versed in financial services. I think that there were so many things as we were deploying it, need to understand like which things matters, what things you’re allowed to do, what processes have to happen inside the banks that are just really, really unique to the financial services and to the regulated financial services. So I think, I hope that people will be careful not to just start throwing these tools on there that are built by people that aren’t really, really well acquainted with the nuances of working in the regulated financial system.
Fantastic. Well, thank you all for joining today, Derek. Thanks for taking the time on short notice, Matt, you as well.
And, you know, John, always a pleasure to have you on the show. That’s it for another week of the world’s number one fintech podcast and radio show, Breaking Banks. This episode was produced by our US-based production team, including producer Lisbeth Severance, audio engineer Kevin Hirsham, with social media support from Carlo Navarro and Sylvie Johnson.
If you liked this episode, don’t forget to tweet it out or post it on your favorite social media. Or leave us a five-star review on iTunes, Google Podcasts, Facebook, or wherever it is that you listen to our show. Those actions help other people find our podcast.
And in return, that helps us build an audience that can be supported by sponsorship so we can continue to provide you with our award-winning content every week. Thanks again for joining us. We’ll see you on Breaking Banks next week.