Fintech Insider Podcast by 11:FS
00:17:43 · Aurélien Bonnel
[00:17:43] S4: Yeah, so the system is inherently fragile and was designed as such, right? In the 1930s when you had a bank run, it would occur over the course of weeks, right? Because people had to get off work, they had to line up outside of the branch, and orderly get their money slowly until the bank run had occurred. You know, the past few years, the bank runs that have happened have been over the course of days, because you've had basically API interactions that triggered the money movements. With agents, it's even worse actually because they all get the same signal at the same time. In engineering, there's this concept called the thundering herd. It's basically everyone gets a signal at once and they all rush to do the same thing. That's exactly what's gonna happen with agents. You're gonna have bank runs potentially triggered in seconds, not hours, not days, not weeks, seconds. So it exposes a system that is inherently quite fragile, The fractional reserve system, the fractional banking system that we have right now. Basically banks keep, you know, 10, 20% defined by capital ratios. They keep that as cash and the rest is loaned out. If you cannot sustain the agents all redeeming their money at once because they got an event from Bloomberg that triggered them to do this, then your system is at risk. Right? And so I think we're gonna see an evolution of this system. We're gonna see more narrow banks that do not lend like N3xt, that are coming up. As far as I know, are none in Europe right now. I hope to see some in the next few years. But globally, that this it's the same problem everywhere. It's the same system we're all running on and it needs to change to adapt to be at, you know, agent speed rather than human speed.
00:19:37 · Ross Gallagher
[00:19:37] S1: Merusha, I'm really keen to get your thoughts and I suppose your build on that as well, and I suppose leaning on your your sort of institutional expertise.
00:19:47 · Merusha Naidu
[00:19:47] S3: Yeah. Absolutely. So I think what Aurelien was saying is so true. So the infrastructure is key to keeping all of this together. Right? If you don't have stable infrastructure, which a, the AI agents can tap into quite easily. So access. Right? But also the platforms to actually, you know, service this many transactions within a very short space of time. I think that is absolutely key. So when we look at, you know, the more traditional players, they do, like, was saying, they do run on three or four days. This is like real time payments. You've gotta have the infrastructure, the capabilities, the guardrails in place to make sure that, a, the system can handle it, b, there are rules in place to also mitigate any kind of fraud or agents that are doing, you know, transactions which they shouldn't necessarily be doing. But it's also making sure that from a... Like, for example, from a processing perspective, do you have the right partner who can ensure that you can man... Manage all of these global payments at the same time with all of the guardrails while still making it fundamentally easy for the AI agents to trigger and successfully make these payments. So I think all of those components together is what is gonna make this a success. And I think that's one of the challenges so far that we've seen, you know, with a lot of players who have tried or have started to enable AI agents. They've done it in a way that is very intentional, but intentionally small so they can actually test their own infrastructure and see what works and what doesn't work. The first movers are the ones who are saying, actually, we need to either change a partner out because it's not working, or we need to rethink our overall architecture. And so that's one of the reasons why we see it being slightly slower than I think the market expects it to be. But I think once you get that infrastructure layer right, this is gonna be an area that completely changes how we look at banking and and servicing customers on a daily basis.
00:21:53 · Ross Gallagher, S5
[00:21:53] S1: Yeah. And I think that point that point about testing and learning, and doing that in a controlled environment before you start scaling this stuff out, I think is, is gonna be so key. Alright. Look. I think we have done a pretty neat job of, setting the scene. We're gonna take a very quick break, but we will be back with you very shortly. In part two, we're gonna explore how banks might actually need to reinvent themselves, where the opportunities to make money could actually emerge, and then what a profitable banking model might actually look like in an increasingly automated world. [00:22:29] S5: Launching innovative payment products is only the beginning. Real progress comes from what happens next. Paymentology fuels continuous compounding momentum as a cloud native issuer processor, helping banks and fintechs grow beyond go live with global reach, local know how, and the power of its Loom platform. Powered by more than 500 paymentologists across 68 countries, they make card programs work naturally across markets. Want to keep the momentum going? Head to paymentology.com.
00:22:29 · S5, S0, Ross Gallagher
[00:22:29] S5: Launching innovative payment products is only the beginning. Real progress comes from what happens next. Paymentology fuels continuous compounding momentum as a cloud native issuer processor, helping banks and fintechs grow beyond go live with global reach, local know how, and the power of its Loom platform. Powered by more than 500 paymentologists across 68 countries, they make card programs work naturally across markets. Want to keep the momentum going? Head to paymentology.com. [00:23:03] S0: This episode is brought to you by Google Chrome. You think you know a browser, but Gemini Chrome, that's new. They can help you with practically anything on the web, like restoring a vintage motorcycle from a 50 page restoration block or finally break down that long article you've had open for weeks, Gemini and Chrome is here for it. Ready to make anything online make sense? There's no place like Chrome. Check responses, setup required, compatibility, availability varies 18 plus. [00:23:31] S1: Welcome back. Merusha, we ended the last segment with you. I'm going to start this segment with you as well. If AI agents increasingly decide where customers save, borrow and invest, how do you think that competition then between those financial institutions begins to change?
00:23:49 · Merusha Naidu
[00:23:49] S3: Yeah. I think when we look at, you know, how financial institutions are going to start competing effectively, is they really need to focus on how do they make themselves visible. Right? Because that visibility is key to making sure that AI agents, a, can actually access the information or the services, products, features, everything from commercials to what makes the service better than any other. Right? And I think that's one of the areas. I mean, even if you go on ChatGPT now and try to, you know, look for a holiday, if you ask the same question a couple of times, you're gonna get different answers. Right? Now you can't afford for that to happen when it comes to your banking infrastructure or your banking services. You need to have... You know, you need to trust that the AI agent is bringing back the best results. Now when it comes to the banks that we're working with... So for example, we have Mal Bank in The Middle East. They are an AI native bank. So they have built in AI agent capabilities from the start. Right? And so what they've made sure is their bank is visible at all times. They specifically built it in a way that AI agents can receive the data that they need in the best way possible to give them the best scoring. Right? And so they've typically thought through an agent being able to access information by an agent. Right? So optimizing the way the data is actually transferred, accessible, and visible. Right? What they've also done is they've looked at what are the key decision makers or drivers that AI agents usually use when making a decision around ranking rating. And so when we look at their proposition, we can already see outstanding results from the way their agents are operating, from the search results that we're seeing. So I think it really... There... There's a big difference between AI native banks and banks that are adding AI agents onto their propositions. Because the banks that are adding AI agents onto their propositions, they are a lot more nervous, and their guardrails need to be better than an AI native bank because there are so many moving parts within a traditional bank. Right? There are so many different rules, configurations. All of those things need to be, you know, reviewed. They have to have guardrails around the actual product service features, commercials, then put an AI agent on top of it.