584 Hot Takes on Banking Regulation, Standards, AI, and Fraud

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.

Today, another episode powered by U.S. Bank recorded at the Stena Center for Financial Technologies Fintech Exchange in Utah. We kick off with my good friend, Bill Goldfeder, CEO of the American Fintech Council, and Seema Gandhi, senior advisor at FS Vector and co-founder of the Coalition of Financial Ecosystem Standards, to talk about everyone’s favorite topic, regulation, and would standards actually make a difference? In the second half, John Son, CEO at SpringLabs, and Simon Taylor, head of content and strategy for Sardine and I, talk about AI, fraud, and the future of financial services in a world run by robots. So, what don’t you like about the hot sauce? If you want the truth, it’s not kosher, right? Number one.

Fair. Right? Which is fine. No, I’m not going to file any formal complaints.

But if we were offering alcoholic beverages like Fireball shots before we went on the air, you’d get a better interview. That is true. You get the same brand.

I mean, you get the same hot take, you know, same. You would get the hot take. Yeah, you get the same thing.

Every time you say hot takes, you need to take a shot of Fireball. Now you’re speaking my language. So, the very first Money 2020, it was Shamir and I. So, Simple had just announced, but not yet launched.

Perk Street was live. I forget who it was from hire one. And the moderator, like, walks out with a giant bottle of tequila and says, anytime anyone mentions their own company, you have to take a shot.

I like that. So, now imagine trying to have a 45 minute panel conversation without referring to your own business, including it wasn’t just like the proper name. It’s anytime you refer to your business.

So, but we’re already getting into the hot takes, just not the hot takes we were talking about. Not the ones you wanted. He just said hot takes.

Hot takes. Yeah, he’s a hot take. I think there’s a bunch of hot takes coming up in the industry.

What’s really telling is we had the LA Labs Banking as a Service Center of Excellence was last week. And we had the folks from the Kennedy School that Ian Maloney from your group had introduced us to. And we’re talking about this direct versus indirect.

What was really interesting at a high level, that group said, that actually doesn’t model doesn’t exist anymore as we think about it. Ian’s still responding to this. It’s like, as we think about it, there’s kind of this definitive line.

There is banking as a service evolution up to 2024. And then it’s kind of like a BCAD sort of thing. It’s like, but 2025 going forward, the world is going to look a heck of a lot different.

Now, as two people have been involved in this for a really long time. Sue, I’d love your take first. Do you see like kind of a drawn line of like 2024 and before 2025 and after? From a tech infrastructure perspective? Across the board, it just feels like 2024 is to fintech as COVID was to, you know, digital transformation and kind of remote work.

I think my take is maybe folks are looking at 2024 as a moment because you had this climax of regulatory activity. And then you had the promise of a new administration coming in to ease some of that pressure. And now it feels like, OK, there’s a moment and maybe 2021 is that moment.

I don’t know. I think maybe it’s a natural evolution of you had just a massive pace of new technology meeting older traditional operating models that came to a head. There are cracks that became very apparent that probably everyone in the industry, I’m sure.

Maybe even a crevasse, not just a crack, but a crevasse. Or maybe multiple that manifested and you couldn’t put your head in the sand anymore. I mean, I think most insiders would say the Synapse thing was like pretty obvious many years ago.

Everyone knew. Everyone knew, right? And it happened. And so finally, like the things that people have been worrying about or talking about, you couldn’t ignore anymore.

So in that sense, maybe 2024 was the defining moment of we actually had to do something here. And so I think that’s fair to say. Like, I think there’s been a shift in terms of how folks are talking about things.

But what I want to say is I do think it’s clear that innovation is here to stay. Like, new models are here to stay. And I don’t know that it’s like something’s changed in 2024 in terms of the model, but in terms of how we all work together, yeah, maybe that was it.

Phil, before you answer, I do have to call out because Sima said Synapse, and for the audience that can’t see this out of range, Jason McCoola is sitting over there. I’m waiting for him to dive in, grab a microphone and say, you said Synapse three times in the mirror, and Jason McCoola appears. I mean, it’s weird saying the word Synapse.

I kind of feel like it’s like the one who shall not be named. The one who shall not be named. So Phil, I would say as an outside observer to AFC, something has definitely changed because the arc of AFC has changed, right? Like, I can look at this.

I was going to say, that’s a testament to his personality, have you not known that? It’s funny. I’m a fan of yours. I read your LinkedIn posts and I’m always fascinated.

There’s a political trick that you have embraced, right? Where Joe the carpenter, like, oh, I was talking to a three billion asset size bank, this is only two billion, right? Meaning you try to make what you write human and real in terms of providing some intel and insight for your readers. And by the way, I very much appreciate that. I live that in terms of the way I operate at AFC all the time, right? I spend my days, maybe probably too much, talking to my members all day, every day, not just the CEOs and not just the chief compliance officers, but across the board to fully understand what’s going on, what are your challenges, what do you need? How can AFC provide you value? Interestingly enough, I spoke to someone last week, it was one of our banks, who said the same regulator that came into my bank four months ago to perform our exam is going to be the exact same regulator that comes in two months.

Am I to think that whatever’s happening in Washington at the very highest levels is going to get all the way down to that examiner by three months from now when he’s ready to come back? He’s probably not. It’s going to take time for the evolution and the shift that we all see and hear and talk about to happen. It doesn’t happen overnight.

At our summit, you heard Yelena McWilliams say that it took almost nine, 10 months to get the most basic reports run, to gather certain information. Now, I think the beauty of having acting chairman Travis Hill, he was there once before, he was vice chairman, he understands the bureaucracy. I think it’ll move quicker, at least at the FDIC level, not to say to the Fed and the OCC.

But my point to you is that I think everybody so is trying to create that change, right? Like, you know, we’re going to will it so, it’s going to happen. And from 2024 to 2025, and I don’t think it’ll happen as quickly. I think rightfully so, it should be a bit more thoughtful.

It should be with more collaboration and doing it the right way will actually build a foundation that has some lasting power. So, let me play devil’s advocate. I don’t actually, I’m going to throw a hot take that I don’t necessarily actually believe, but this is the, when we had started this, right? The ideas, let’s workshop an idea.

So, play this out with me, because I’m not sure where it ends. Under Yelena, the effort to modernize the FDIC, at the top, there was vision that even, you know, before the coup, you know, three years in had not reached a level. And that was at a point when there was enthusiasm and optimism around it.

Then if you look at, I’m going to just get myself in all kinds of hot water where this one goes, but then like the coup happens, then you end up under Grunberg where like, it becomes the inquisition that anyone that was pro-innovation was, you know, pushed out. And so, even someone as vice chair Hill, acting chair Hill, who would, you know, liked his opening speech, it feels like people are actually going to be a bit gun shy to go embrace that. And I don’t know if I fully believe what I just said, but I’m playing out like a counter argument to you where we’re now in a world in a post-SVB, post-first republic, where regulators lost their jobs in a way that you hadn’t seen before because of improper oversight.

Are you going to see people at the regional and below embracing of this change? Because to your point, it’s like certainly not happening over two to three months. I mean, my take is Yelena was way ahead of her time. And what she did was so hard because she was talking the language that nobody wanted to hear.

That was like, and Marty Grunberg came in and that was like the day of traditional regulation. And I think that what we saw in the last four years is that things are so broken. We have examiners that come in and focus on things that they shouldn’t be looking at.

Oh, is your contract evergreen or not? Like, let’s have a conversation on what’s, it’s just like. But that’s the CYA mentality of people who are scared. It got so bad is the point where you cannot run a business that way.

Like it’s breaking. We’ve had so many community banks shut down. You know, the core of innovation and financial services has been between bank and FedTech partnerships.

That’s not going away. And so you’re looking at a reality like in the face where I think the people, maybe the innovators that might’ve been pushed out four years ago, this is their time now. Like Yelena’s predictions, like this is her moment.

If she came in and laid out what she did four years ago, I think that the time is just better. The work that we’ve done on standards last year, 2024. I couldn’t have had that conversation in 2023 for something.

Something has changed where it’s gotten so bad that if we want to have a thriving and competitive financial services ecosystem, we have to change how things work. So let me just say, firstly, I love what you said. Cause I completely agree with you.

And I think it’s not just what Yelena was trying to do when she tried to do it. Cause I agree. She was way ahead of her time and there just wasn’t enough around her to support it.

What I would say to you is, is that she did it in a very thoughtful way, different than maybe some of the other agency leads at the time, may have thought about the way in which they have to engage with their, with their organization in that it wasn’t a free for all. It wasn’t like, okay, we’re open for business. It was okay.

We’re going to be able to, we’re going to give it a much more thoughtful approach to how we discuss and have these conversations. Look, I think vice-chairman Hill has had the ability to learn from her and to have worked with her and to sort of seen where she struggled or where she hit those roadblocks from the bureaucracy, not by a fault of her own and now maybe can overcome them. But I think you, we’ve got to be mindful of the thoughtful approach because look, the biggest fear for us.

And I say this on behalf of all of my members is the perception of deregulation, right? Perception of deregulation is a very scary thing. Not just because what you saw, director Chopra had sent out to all the States, basically giving them permission to regulate again. Um, which, you know, is just fascinating to me.

I mean, I’m going to be nice and not, sort of get myself into trouble, but like, like all of a sudden the States have a right to step back into their regulatory authority. I mean, I think a lot of States now, though, unfortunately, the perception of deregulation will have them like sort of pushing back too much, right. And sort of stifling innovation and reasonable economic activity and sort of, you know, creating access in a way that I think wouldn’t be right for a state to step into.

I think Phil’s point about, you know, you can’t take a, Oh, deregulation, you know, like time to party mentality. That’s not, this is the time to educate and build for the future. So that’s absolutely right.

And yeah, I’ve heard the same thing around some of these States, but the interesting about the States is that they, most of them have a dual mandate around safety, soundness, consumer protection and competition. And I think we don’t have that at the federal level. And so what we’ve seen is just an inordinate amount of focus on consumer protection and safety and soundness, and you can, you know, regulate to a straitjacket.

And so that’s taken a big hit on competition, which in some ways also encourages consumer protections because people can vote with product choice. And we’ve just been missing that part at the federal level. And we sometimes miss at the prudential level, the mandate is not to protect the individual bank and ensure competitiveness.

It is to protect the system. Hold on. I mean, but the system, let’s be honest for a moment.

If we’re really going to be honest was, you know, did, was Synapse create some sort of world economic incident that I’ve, that I didn’t see. Right. Like I, I agree.

It was terrible. Sorry. You stay right there.

I agree. It was awful on a lot of levels. Right.

But it, I mean, let’s put it all in perspective. I think, I think that’s exactly right. Like what, I think a hundred million in funds lost, but a billion dollars, a capital one and interests that CFPB identify that wasn’t paid out.

Yeah. Oh, CFPB different than the others. My point is the magnitude of what we’re talking about and the disproportionate burden.

It’s so easy to hit on the new kids. Look, I, I, I oftentimes, I mean, Jonah crane, the Clara’s group has done a lot in terms of talking about the, the, the disproportionate attack on innovate banks were in the engaging in innovation. There was not a week that goes by or, you know, whenever the consent orders come out with the FDIC is the last Friday of the month, the FCC is a little bit different schedule.

There’s always a big brand name on list almost every single time. And no one ever talks about it. And the magnitude of damage that has caused is five synapses.

Right. And again, I don’t take for granted one nickel of, of consumers money that got lost. Right.

But I, when we, we sort of judge an entire ecosystem based on the bad actions of a single player or, or a few players, if it’s the wrong approach to take. Well, having spent a lot of time in this part of the industry, the, I think you can look at and say, synapse is the exception, not the rule to the industry, but it is also indicative of the fact that regulators should have played a role in this, you know, because to SEMA’s point, we all saw this coming. This was the Austin powers, you know, steamroller coming in slow motion down the industry of what’s about to happen.

I mean, this is a tough one to talk about on a video recording, but regulators had a seat at the table in that situation. And they did nothing. Well, I would look, we created the American FinTech Council with the idea of creating a standards.

I say this a lot where standards based organizations or the industry can define itself long before a regulator or someone on the outside can define it. We spend a lot of time talking about that. A regulator or an examiner can possibly figure out something they don’t understand or comprehend, right.

Or didn’t have the right education. And we wrote our letter yesterday to acting chairman Hill talking specifically about examiner education. The examiner can examine if he doesn’t understand the nature of the products, right.

And so big deal that there was examiners within a few community banks that maybe could have seen something, but they didn’t even know what they were looking for. Right. We talked about this, like know what you know and know what you don’t know.

These examiners don’t walk in there with any sort of vulnerability to say, okay, guys, what are you doing? Like, how are we like, explain to me what you’re doing so I could properly understand it, understand the risk and then give you the appropriate sort of mechanism to move forward. I mean, that’s definitely a problem. And I’ll quote a regulator that, you know, said a lot of these examiners are file clerks that were promoted.

Right. So, yes. But I think my challenge is when we talk about, well, synapse happened, we need more regulatory authority and we need more superficial authority.

We need more rules. And it’s like, well, that’s a very different solve than education. And maybe there’s something happening here and we need public private partnership and the opportunity to have experts come in and supplement the process.

That’s a very different conversation than more straight jacket. And I can I and I I went in to visit about six months ago with the Marty Grunberg FDIC. And I’ll say it like that just to make sure that I’m making a clear sort of differentiation.

And we started it was one of the first times and it took us a long time from in my tenure to see where we finally got in the door, had a conversation. We sat at a conference table with about 15 people from FDIC staff, various levels. And the tone and the theme of the meeting was we don’t have to change anything.

The rules exist regardless of whether you’re innovating or not. Thinking, you know, sort of structure exists and everybody needs to work within it. Right.

Yeah. Great. I don’t agree with that necessarily.

Right. But fine. I understand that as your position.

And then you put out the synapse rule. Right. Oh, which, by the way, a lot of people argued and we argued in our letter.

Number one, I think process wise, you collected RFIs and then you still pushed out a rule mid RFI process. But but like if the rules exist, then why are we doing this? Right. And if banks should be responsible reconciliation, why do we need a new rule to tell us that banks need to be responsible reconciliation? And so don’t tell me on one hand that like, oh, we don’t need to do anything to understand or embrace innovation and yet push out what everybody called immediately the synapse rule.

Well said. So something that three of us care quite about a bit is the maturation of the industry. And one of those things that we hotly debate, normally not on camera with microphones, is this idea of standards.

And before we run out of time, let’s talk about like, what is the role? Sima, let’s start with you. Where do you see standards fitting in as a tool to mature the industry? I mean, I think today bank fintech partnerships, big non-bank partnerships, embedded finance set in this weird area where what existing regulatory guidance we have is within TPRM, which is, you know, written decades ago when a bank controlled its full stack. And that’s not the world we live in today.

And so the idea of a partnership is very different than a vendor supply management one. And, you know, I’m still at the end of the day, a market first, let the market dictate. And so standards help the market dictate because now we’re all talking about the core compliance categories, the things within each category that needs to be satisfied in a very clear and transparent way.

So that when your examiner comes in, they won’t have a conversation. Everything is laid out clearly. They don’t have to necessarily become the expert.

There’s a third party that’s come in and shown that expertise so they can have a conversation about risk management. So I think standards are the way forward where, you know, with partnership by regulators, by industry, that you solve that gap and you give clarity to the market so that we’re not playing whack-a-mole and doing redundant compliance and compliance theater. And the most hotly, like, job posting right now, compliance, compliance.

Like, you want to find a job, just do compliance. Like, you’ll have competing offers, right? That’s ridiculous. That’s not solving a problem.

So I think standards that evolve and can be more nimble and signal to the industry what best practices are and also give reassurance to regulators that there is rigorous thought and process and, you know, compliance practice is going to be important for true risk management. So, Phil, I’ve heard that AFC is the premier standards-setting organization. Not standards-setting, standards-based.

Standards-based organization. How do you see the role of standards playing out? Let me just say, and I’ve completely, and I hope my wife is not listening, fallen in love with Seema and the work that she does and the energy that she brings to standards. I don’t think you did a good enough job of kind of plugging yourself in the work that you’ve done over the last number of months with CFES in terms of actually sort of building it out.

I think a lot of people talk about standards and few people actually act in terms of how we’re going to do it, how we’re going to build the collaboration, how we’re going to build out those standards. Because I have good mentorship. Not please.

Not even close. What we have done, and again, we were founded as a standards-based organization, was less in this area. While we’ve evolved into the banking and finance space, it started with fintech lending, right? And how to think about fintech lenders who are partnering with banks.

It started with buy now, pay later companies and what kind of standards we’re building around those. Earned wage access. AFC and our membership developed, I think it’s 17 points of standards in terms of how do you define the right way for a standard to set.

I don’t want to get ahead of my skis. I love some of the work that CFES is doing, and I’m hopeful that we’re going to find ways to work together. We have just over 30 innovative banks in our membership today.

I think that number is going to grow pretty quickly over the next couple of months. And so I think it’s critically important that people like Seeman, the work that she’s doing is sort of taken seriously. A lot of people, like I said, have talked about it.

And so to see it actually playing a real role in the conversation as quickly as it has, credit to her. And we’re excited to be a part of it. Fantastic.

Thank you both for joining with your hot takes. Thanks. All right.

So we’re three days into a sea change of administration in the U.S. And I dare say the U.S. likes to be U.S.-centric, but I think this is kind of the administration change heard around the world already, for better or for worse, going to stay out of the politics of it. But one of the interesting things that I thought happened so quickly was President Trump leaning in on AI and one of his biggest fanboys, Elon, immediately taking aim at a half-billion dollar initiative. Half a trillion.

Half a trillion. 500 billion. 500 billion.

Yeah. But like, Simon, break down Stargate. All the bad puns can be yours in terms of, like, you think deeply about this.

Like, this setup, was it ill-conceived? Like, what’s the genesis of why, like, these two that, you know, Elon helped them rise to power again or rise again, and boom, they’re disagreeing on AI. So obviously, open AI has done a pretty good job at lobbying, but as have all of the other hyperscalers to get to this point, 500 billion dollars is a massive amount of money. But anybody in AI will tell you that the core issue is energy.

We need data centers and those data centers need energy, and that’s going to require a massive amount of capital. Each one these data centers that Stargate is building, and they’re building six of them, I believe, or seven, requires at least six nuclear reactors worth of power. Six gigawatts each.

It’s an incredible amount of energy that’s required. So you would need a massive amount of cash just for the infrastructure build-out. So Elon fires back because he sees Larry Eliasson on stage from Oracle, he sees, you know, Microsoft are involved, SoftBank are involved, and he says, well, SoftBank are not good for the money.

And then the next day at Davos, Satya Nadella says, well, I’m good for my 80 billion. And it’s absolutely incredible. So there’s a few things going on here, but I really think it’s about that infrastructure build-out to get you to the stage.

And I think Elon’s just fired back at like, he doesn’t like Sam Altman. Sam Altman was standing next to the president. It comes down to who gets to stand closest.

Yeah. He was on the stage with the guy, John, what’s your take on this Target debacle? I mean, I think it’s pretty clear that the excitement behind AI is kind of there. I mean, the investment certainly kind of reflects that.

I mean, I think it’s going to take that quantum of investment to really get us to kind of the next level, what AI is capable of. And, you know, one of the big debates, you know, right now is kind of, you know, is the transformer architecture, is that kind of what gets us to AGI and kind of beyond, or is there going to be some intermediary, or is there going to be some, you know, new infrastructure or new platform that kind of arises as a result of it? And I think there’s kind of recent news around, for example, DeepSeeks are one that really kind of appends, you know, what we thought about the transformer architecture. There’s apparently a lot more extensibility than we thought.

And in that scenario, pouring $500 billion of investment into it could really get you to kind of that next level. Well, it could, but also scaling appears to be broken on the training side, but it has shifted to inference. And DeepSeek R1 and O1 Pro and O3 have really shown that instead of trying to get more data, because we’ve run out of data on the internet, certainly written data, then you just spend more time thinking about it and really thinking about it.

If you ever saw Hitchhiker’s Guide to the Galaxy, deep thought, like you could solve the universe and everything if you can just think about it enough, that we’re kind of in that space. But what’s really holding us back on the model size is getting bigger and the inference and everything else is just data center capacity, chip capacity. So we need that infrastructure.

So it’s going to play out in different ways. But you were talking, Jason, earlier before about like how many YC companies you’re seeing. What’s happening in the banking industry and the finance industry? This is what’s crazy right now is you’re watching things unfold on multiple levels, right? There’s the Battle of the Titans going on between all of the big players getting in and whether it be from Copilot and how that interacts with ChatGPT and OpenAI and where Claude fits in and what AWS is.

All of these big players are fighting this macro winner-take-all. And then on the opposite extreme, I don’t know how many YC companies that we’ve seen where we are AI for X point solution. We’re definitely running out of fingers and toes for the number we’ve seen for compliance alone.

We built a map for this. It becomes it’s like the, wait a second, we can’t have AI for every point solution, right? Because one of the things I know about banks is they’re not going to buy point solutions all the time, particularly when it’s sold as an enterprise level solution, but it doesn’t solve an enterprise level problem. And I think that’s where we’re having the quandary, which is none of them have matured enough to solve an enterprise level problem, but that’s what they’re charging for it.

But isn’t that because it’s really good at the other side of the budget, rather than buying software, you spend most of your capital in compliance on people. And if the AI can do some of the work of a person, then that displaces some of that budget. But they’re selling it in an enterprise sales cycle with an enterprise cost, but it’s a point solution.

And therein lies the rub, right? And I have to give credit, Barb had said this on stage with me, Infinivate is like, if you’re selling a point solution, it has to be bought like a point solution and cost a point solution. And that’s hard when compliance has risk and enterprise risk. So it has to be closer or do more.

Yeah, love that in a GLC. I totally, totally agree with that, by the way. If you have a point solution, it shouldn’t be sold like an all encompassing enterprise site license for like millions of dollars per year.

That’s kind of not the right economic structure for these types of products. But I think ultimately, I hear you on this dichotomy between there’s folks investing in underlying infrastructure, all of the battle of the titans investing in infrastructure. There’s all of these tiny startups doing vertical AI.

And I think ultimately what it comes down to is who is responsible for delivering last mile value. Is the bank responsible for taking open AI’s model and delivering last mile value, or is the startup responsible for delivering last mile value by being that interface layer between the technology and the processes and the human capital, as well as technologies within the bank? So I know you’re not going to turn this into a commercial, but tell me, how is Spring Labs though, like when you’re approaching that, because you’re kind of in the middle here, right? You’re bigger than the bread box, but you’re not the Volvo, right? In terms of the Titan. How are you finding the last mile? Are you being expected to solve it? Does it vary by customer? How do you handle that? I mean, I think a lot of these similar enterprise kind of SaaS enterprise B2B kind of SaaS solutions, you’re not selling on product capability.

You’re selling ROI. You’re selling on what’s the end result to the bank. And I think to your point, the bank can potentially take like an open AI model and kind of go implement it themselves, but then they’re really responsible for figuring out how that delivers value against their existing processes.

And I think where we’re at is we’re really trying to aggregate kind of enough insights about how things are done in the compliance world to be opinionated as to how that value is delivered. And I think that’s really going to be the difference between the vertical AI guys who make it and the vertical AI guys who kind of get eaten by open AI, is are you a chat GPT wrapper that’s making it easier to access the technology, or are you opinionated as to your solution and really delivering kind of value through that solution? We’re seeing similar things at Sardine, fraud and compliance is often based on how well do you understand the craft? How well do you understand the issue? How well do you really understand fraud? How well do you really understand BSA, AML? What are the workflows that you have to deal with? What’s the other part of the organization going to think when you do this? What’s your existing workflow? What’s the data you need? And if you’ve got that foundation, then building generative AI or agents around that can be additive and can deliver ROI. But just like an agent by themselves, I mean, maybe one day we end up in a world where somebody hires an agent, then it’s just like hiring an agent, like BPO, like outsourcing.

I don’t think we’re there for a little while, like certainly in the banks, but I tell you in the FinTech space, very, very different. And the reason you see so many small YC companies is because they’re all getting traction and they’re all getting to a million, 10 million ARR really quickly. The interesting thing is going to be what happens next.

How many companies can get to 10 million ARR and then get stuck there that we’ve seen in FinTech over the past couple of decades? Like it’s just a story as old as time. A story as old as time, but more prevalent. Frank Rotman and I had this conversation just before the holidays over breakfast is what used to be the zero to one met something.

Zero to one doesn’t really mean anything anymore. And in disclosure, you work for Sardine, I’m an investor in Sardine, but I think there is an important, I have to say that, right? Yeah, yeah, no. But part of our thesis was, as you said, the craft piece of it, wasn’t the technology, love the technology.

You know, Supes wowed me with that years ago, but it was the understanding of the craft piece of this. I find that kind of interesting right now is we’re selling some of most complex technology in banking that it’s ever had, right? We’re not hitting applied materials, other complexities, right? That are far more technical, but this is as technical as it gets when it comes to banking, right? And yet it’s, there’s a human centric element to it and a cultural element that tends to be the rate limiting factor. And I want to like, let’s round this out and let’s talk about how do organizations need to change? And do we think that every organization can make that leap to be, you know, an AI driven organization, which probably will determine some of the winners and losers? It’s almost like, hey, yeah, you can be in the manufacturing business, but if you’re not adopting electricity, you’re probably not going to get around that much longer.

Yeah, absolutely. I think there’s things that they can adopt today to kind of take little baby steps. I’m a big fan of the baby steps model.

It doesn’t have to be change transformation to become an AI company or anything like that. There’s stuff that substitutes almost one for one with the things that you were doing today, except gives you better efficiency, gives you better accuracy, gives you better scalability. Agent training is kind of always been a huge pain point in FinTech.

It’s not just about, you know, can you establish the right set of processes or anything like that? There’s a translation layer between your policies and your processes and your humans. And, you know, repeatedly training folks to kind of be able to execute on that workflow with consistency and with fidelity over time has proven to be very, very difficult. You run into the situation where every time you scale or your volumes change or your needs change, you have to go retrain a bunch of people and that’s something that… All the broad patterns change.

Yeah. It changes all the time. Exactly.

Exactly. And I think that’s something that AI can kind of help you with, you know, day one is to kind of remove some of that. I was speaking to a head of fraud of a sardine client just yesterday who said one of the things they’ve really always struggled with was exactly that issue.

They were relatively small. They relied a lot on outsourcing their agents and offshoring them. They spent all of their time training them.

Whenever a new model came up, the senior leaders were developing rules and patterns to try and kind of deal with it. And they knew exactly what they were looking for. And what they found with the agents was, yeah, they can train the agents so much faster and the quality bar tends to be higher.

So you’ve still got these quality internal people that can see that the world has shifted, what they need to do, but their ability to do it instead of relying on an outsourcer and trying to flex their staff, they’re flexing with technology. And I think that’s a powerful thing to think about. Well, it’s an interesting dichotomy because I hadn’t thought about they’re improving the execution of the models that were human driven and the human insight.

So we’re going to almost see kind of both ends where AI can find patterns faster and better, human in the middle and the execution also becomes an agent. Is that a good description? That’s spot on. And it’s exactly what’s in production for us today.

And I’m sure for you as well. It’s that little bit of intelligence in the middle is the difference maker. Fantastic.

Absolutely. I mean, I think that’s exactly right. I mean, one of the most common evaluation metrics for whether AI models are performing like you expected is human eval, right? It’s like a human eval metric.

So that assumes that the best answer is still one provided by a knowledgeable person looking at this, because at the end of the day, you’re kind of replicating, you know, what an expert in the space would kind of be doing. Now, that might not always be the case. Maybe in the future, there’s some other evaluation metric that kind of makes more sense.

But today, this kind of, you know, human in the kind of middle of the stack process very much kind of echoes what we’re seeing out there. Well, I can’t wait to continue this conversation here next year at the AI conference that Spring Labs helped put on with University of Utah, which was fantastic. And I think we’re probably not even sure what we’re going to be talking about next year, because so much is going to have changed by next fall.

Anything you want to say in close that you’d say you learned coming out of last year’s AI conference? Yeah, I mean, I think last year’s AI conference had kind of a very specific theme to it, which is what are the practical applications that people are looking at exploring. And I think the goal for this year is everyone should definitely try to make an effort to come out for this great show. But what we’re trying to do this year is kind of almost 12 months later.

Now, what’s actually happened? So everyone talked about what they’re going to do last year, what was real, what wasn’t real, what delivered ROI, what turned out to be a flop. And I think this year, we’re going to focus much more on practicums, real case studies, what has kind of moved the needle, what hasn’t for folks. Heck yeah.

Look forward to participating next year as well. And thanks for joining us. Thanks for having us.

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 U.S.-based production team, including producer Elizabeth Severance, audio engineer Kevin Hirsham, with social media support from Sylvie Johnson. If you like 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.