Brandon Arvanaghi (Meow) on Banking AI Agents (EP.715)

Today on the podcast, I sat down with Brandon Arvanaghi, the co-founder and CEO of Meow, a Castle Island portfolio company. Meow is a financial technology company that provides a suite of products for businesses, and that includes business checking accounts, stablecoin accounts, and global treasury solutions. This week, the company announced the launch of a new product, bank accounts for AI agents. In this conversation, we discussed the convergence of AI and financial market infrastructure and what Meow is seeing in this fast-growing part of the market. I think you'll enjoy this one. Without further ado, here's my conversation with Brandon from Meow.

Matt Walsh and Nick Carter are partners at Castle Island Ventures. All views expressed by them or the guests on this podcast are solely their opinions and do not reflect the opinions of Castle Island Ventures. Guests and hosts may maintain positions in the assets discussed in this podcast. You should not treat any opinion expressed by anyone on this podcast as a specific inducement to make a particular investment or follow a particular strategy, but only as an expression of their personal opinion. This podcast is for informational purposes only.

Brought down by bad mortgage investments, Lehman, which has 25,000 employees, will be liquidated. The federal government loans American International Group, AIG, $85 billion. This is a different kind of market, and the Fed is asleep! The federal government is stepping in to stabilize Fannie Mae and Freddie Mac, the two mortgage giants that have been threatened by the housing crisis.

The Bank of England has pumped £75 billion more into Britain's ailing economy with a new round of quantitative easing.

You print a couple trillion dollars, and all of a sudden people start to worry. So out of this worry, we have something called the Bitcoin. Bitcoin.

Brandon, welcome back to the podcast. A lot has changed since the last time you were on, huh?

A lot has changed.

So I feel like you guys have 5 or 6 new products probably since the last time you were on. Could you just set the lay of the land in terms of what does Meow look like today?

Yeah, it's interesting because they all seem like new products from the outside, but when you peek under the hood, this is all the same thing. So what we just launched, we're the first fintech to offer bank accounts for AI agents. The first fintech in history. This is a moment just like back in the day where people thought it was crazy to open a bank account from their mobile phone. That was unheard of. It was like, I could just walk to the branch next door. Who needs this? Oh, it might be unsafe to do it from my phone. We became the first to allow an AI agent. So your favorite LLM, be it Claude, Gemini, ChatGPT, to open an actual bank account for you. Now that is opening up a ton of doors. I think the Vibe Coders, the Indie Hackers, etc. They want no friction. They want to be able to spin up a business, open a bank account, accept payments all in one prompt. We're at the forefront of that. So the entire app, we're rethinking from an agent-first perspective. So it's not that anything's new on the app, but it's just this is the logical conclusion of what finance looks like and we're just trying to get there first.

It's pretty incredible. When you started Meow, AI was a word, but we didn't have LLMs. The whole industry has evolved. What's it been like to just operate a startup in this environment?

You can go so much faster than you could before. It's crazy too, because when— I'm not going to name names, but there were some AI companies that came out with AI coders, etc., first, and they all were terrible. We were early adopters to them. It was just spaghetti code, terrible, etc. Then there was like an inflection point. I think it was with one of the cloud models where it started getting good. It went from bad to good overnight, and that's when things really opened up. So we've invested a lot in our infrastructure to make it AI-friendly, basically allowing code to be written, code to be reviewed. If there's an alert, it's automatically triaged and PR is submitted for it, end-to-end tested. It's absolutely Bananaville. Our engineers are not really writing code by hand as much anymore, if at all. They're reviewing code by hand. They have agents to help them, but it's just crazy what's happened in the past year and a half.

It's really interesting. And I'll probably stumble over this in terms of the use cases. A startup or a company comes on board your platform. They have a general corporate account, right? But you wouldn't want the corporate account permissioned onto these agents for obvious reasons. You wouldn't want your agents to have access to the full bank account. Am I thinking about that the right way, that the reason you would want your agent to have an account is so that you can dynamically permission it, give it access to certain amounts of capital, put a human in a loop potentially for large balance transactions?

Yeah, you can configure it however you want. It's exactly right. I would say what's the logical conclusion though? The logical conclusion is people are not going to be logging into meow.com. They're not going to be logging into brex.com. They're going to have their AI personal assistant, the same one that does the hair appointment for them, the nail appointment, get their partner a gift. It'll also say, hey, you received a bill, by the way, for this much in your email. They applied the 20% discount they promised. You haven't paid them yet and it's due in 10 days. So I've already scheduled it. It'll be out in 10 days. So honestly, agents are going to be extensions of ourselves that can work 24/7 without any coddling. So that is the future is end-to-end control, just the same that you would if you're logging into an app like Meow in 4 years, 5 years, because we done something wrong is the way I view that.

That's really interesting. So you think the future looks like the whole layer is almost obfuscated and human just needs to be in the loop to onboard the agents essentially upfront and onboard the agents over time.

I think that will be like at the beginning of the loop and then you won't have to like re-KYC an agent across like 1,000 apps. But right now the human is a harness, so to speak. They're like a safety measure is we're just the early days of this tech. But in the long run, it's pretty obvious to me, it's a bug, not a feature for a human to be involved in any of these monotonous, terrible things like banking. Who cares, right? Back office, who cares? It's something people don't want to deal with, so why would they?

When you look at maybe some of your most advanced customers, the customers that are really leaning into AI, what are these agents actually doing? I mean, maybe bring that to life a little bit in terms of what a best-in-class company would be using AI agents for.

Great question. So before we became the first fintech to support bank accounts or AI agents, so agentic bank account openings, we had customers who were spinning up their local cloud environments to create dashboards They wanted to view everything. They wanted to use our API, feed in all the data from their bank account for their company, and create internal dashboards, create alerts internally, basically fly coding their own command and control center for the office of the CFO. So that was really interesting. They weren't really interested in our dashboard. They wanted to fine tune it locally. And we had non-technical CFOs spinning that up. This is mostly view only. Some people started delving into the world of let me try to make it more agentic using the API to transfer money, et cetera. But now we've come to the table in a big way, being the first to launch bank account openings and agentic bank account openings. So now people are actually signing up agentially. We have a channel where we just get notification of sign up and account opening, which is crazy. This is a new channel, didn't exist but for Thursday of last week.

What was the hardest part about just building out the product for the agents? Were there obstacles? How are agents different than humans in that respect?

It's very different. There's a lot of trial and error that goes into getting prompts and interacting with an agent. it's something of a black box, these models under the hood. So our engineers would have to do a lot of guessing and checking, and there's a certain feel that comes for it over time. The more you work with AI, the more you understand it. It's almost like spiritual, like how to work with these things. So getting comfortable and understanding what makes it tick was the hardest part, basically. We're in a good spot there.

When you look at just the landscape of the big model companies, it seems every few months there's just a seismic shift, and one— Anthropic pulls in the lead, OpenAI catches up. How are you seeing that evolving? Are you building this stack to be agnostic? Do you have a preferred partner there? How do you think about it?

We went heavy with the cloud branding because they seem to be the far and away winner right now, especially for businesses. But we built it to be agnostic behind the scenes. You're exactly right. So even though our branding is cloud-oriented, it's agnostic. And yeah, I'd be terrified to just be a pure software company right now. This is a crazy world we're entering.

Do you see it impacting broader financial services in a major way?

I do, because fintech is only 3 to 5% of the market for financial services. And we've been talking for a long time about, oh, banks are going to have to wise up because the meows, the the Brexits, like they're all coming to eat the lunch of the incumbents. That was inevitable. Anyone with any sense of this space knew that was inevitable. But the rate of that now with agents opening bank accounts now in India, it becomes primitive technology compared to the US Air Force basically is what's happening here. So the rate of the flipping, so to speak, of traditional incumbents to the new age tech company and fintech companies I think that's the most obvious bet basically of the next 2, 3 years. It's going to be unlike anything that we've seen.

You guys have obviously been at the forefront of digital assets as well in your career, really were I think the first company to enable stablecoin payments. I got a kick out of, there was a Y Combinator company that just got funded a couple days ago and they made a big press release about, "We got our funding in stablecoins." I was thinking, I think Meow's been enabling this for 2 years, but maybe talk a little bit about this convergence of stablecoins on your platform and how you see that intercepting with AI agents.

People have different philosophies on this. We've always tried to be the most practical. We started deep crypto, then we went deep tradfi, and we found ourselves at the most practical intersection of business checking accounts, actual FDIC-insured business checking accounts from our partner banks, Cross River Bank, Grasshopper Bank, plus the native on and off ramp, the free USDC that you could send and receive from it, the ability to hold USDC as well. There are some people who go deep crypto, only crypto. So we think the future is only stablecoins. No one's going to need a bank account. I was going to get a So I see merits to every side. My thesis is it's important for Meow to be the first to allow agents to open business checking accounts, actual banks, not e-money accounts, not wallets, not virtual accounts. Because I think there's a lot that's predicated on the banking system in the US, for better or worse. Some people think that it's all going to be wallets. And one of the big reasons everyone's pointing to only wallets, only stablecoins, is because of the friction involved with KYC, KYC. So an agent doesn't need to have an identity right now. We don't need to wait for Congress to figure that out. They can fund a wallet elsewhere. It's like they could have some pre-funded funds somewhere, put it in the wallet or receive payments in a wallet and send money in the wallet without requiring a human or KYC or KYB. We, I think, have turned that on its head a little bit because with this Agentic bank account opening, man, I don't think most people thought that was possible, but we've done. I don't think banks are going away, number one. And number two, I think No one is really thinking about this actually being the long term, which is what Miao brought to the table.

That makes sense. Do you imagine a world where AI agents will have a bank account, they'll also have maybe a stablecoin wallet, they might have a card that facilitates maybe a deferred settlement? What does the landscape look like there?

It's a good question. I think they'll do the path of least resistance for themselves. It's the classic question of if you have a bank account, then are you also holding stablecoins separately? And if so, why? So maybe there's a protocol, like a DeFi protocol you want to earn yield on with your stablecoin wallet. If it's just to get exposure to dollars, maybe you're happy with the bank account. I think AI agents are going to be ruthless negotiators. I think they're going to pit every fintech and banking against each other. They're going to open accounts at multiple. They're going to say, hey, this fintech offered me this percent. I'm going to churn instantly if you don't give me the same amount contractually. So we'll have to deal with that. Yeah. Why not? That's the future we've been building towards as well. We've tried to build this like Costco, so to speak.

Well, you guys have also made it pretty sticky for your customers, right? Offering other services like record keeping and 409A evaluations. Maybe talk a little bit about that strategy.

Exactly. So we also layered in bookkeeping, tax, 409A. In the AI world, I don't think you need to be like a single thesis or single threaded as a company. I think it's counterintuitive, it's contrarian, but I think being multi-threaded is very important. It's not enough to do one thing very well. You need to be multi-thesis, multi-threaded. And what we found is founders, companies don't want to deal with things. And we try to take all those things off their plate and now make it agentic. So they don't want to deal with banking, make it agentic. They don't want to deal with bookkeeping, tax, let us handle it for you. That's becoming more agentic by the day on our platform. At the end of the day, the CP is switching from the founder to the agent, and that's what we're focusing on.

It's a really interesting time from an entrepreneur's perspective because you can make an argument that you're almost better off just starting from scratch and building a company today as opposed to having started one 3 years ago and have all this legacy tech debt. Maybe just talk about what you're seeing with the founders that are starting businesses and the various cohorts of Meow customers and what some of the challenges are there for maybe some that didn't start early.

That's a very astute point. There's this sweet spot and there's this like danger zone of if you started 4 years ago and you're not in fintech, for example, you're just a pure software company 4 years ago, you don't have the partnership with a bank, you don't have your own licensing. You don't have something that requires humans basically to sit with the company where that 4 years was necessary, you're in deep shit from my perspective because all you have is a legacy infrastructure, legacy processes, probably hundreds of people, and now you're competing with people who can set up an AI-first infrastructure which allows you to code quickly, alert quickly, develop quickly. The slope difference is insane. There's this sweet spot that we're frankly in, thankfully, of having been around to vertically integrate the Meow Card that requires a partnership with a a bank processor, have multiple partner banks, get very close with their teams over a long period of time, have compliance frameworks they've agreed to. There's a lot of compliance-oriented— I don't wanna say moats, 'cause nothing's a moat in my opinion in the long run. There's a lot of compliance-oriented things that we're able to do now that you couldn't just do out of the gate as a new company here. You need to have some trust basically, or some relationship aspect there. But if you don't have that, or a license, or if you're not a hardware company or something, You're in deep trouble. The companies starting today will outslope you in record time.

Yeah, I'm sure you're getting questions from entrepreneurs in your network around just how do I get my organization up to speed here? Like, how do I get people to start building more in this agent category? Do you have any typical advice that you're giving to founders that maybe feel like they're a step behind in terms of this AI adoption?

It's downstream of good engineers at your company. The good engineers at your company will already have ideas on how to make your infrastructure like more AI-friendly, basically make it such that you can share context across different parts. The agents need sufficient context, basically. Good engineers know that. So it won't be like a founder, frankly, understanding that. It'll be like your best engineer saying, "Oh, this is so obvious. We need to do this yesterday." Hire good engineers, good things happen. Every company in this industry, basically.

I think about this a lot from the perspective of the banks themselves, and some of your partner banks must be seeing this. There's a big narrative with this market structure bill around stablecoin yield pulling away deposits. And maybe that happens, maybe it doesn't. Didn't necessarily happen that much with the money market funds, 20 years ago. But I think really the deposit flight could potentially be just the banks that are very slow to adopt to the startups here where there's going to be this migration of just pulling assets out of banks that are not flexible enough to allow some of these use cases.

That's true. I'm generally very bearish on banks' abilities to attract deposits absent a really good fintech partner or absent lending. It seems to me that lending is like the last kind of salvo for a lot of these banks are going after Series B companies or large companies offering a line of credit. Basically, that's tried and true. People will crawl over glass for when you extend risk to them. But the ability to just attract deposits without that, that comes from a tech experience, that comes from ground game, that comes from distribution, that comes from things that banks are in deeper shit than ever from my perspective. So some of the fintech banks will do very well because they partner with good fintechs and they understand that the marriage of tech company and bank, that's a marriage, very good marriage. Like we both do things the others can't. But the regular banks that are not lending, I, I don't think people are opening accounts based on the branch nearby anymore. It's based on opening your phone and getting a Robinhood account these days.

To that point, it was interesting, this Brex transaction and that trading to Capital One. Obviously they must have been seeing some of that threat to make an acquisition like that. I imagine you'll see more of that.

For sure. There's a big delta between Capital One in terms of tech forwardness and Most banks though, I think a lot of them have their heads in the sand, but I think there's a category of banks similar to Capital One. We're probably going to see some follow-on behavior in the next year, I'd say.

Yeah, that makes sense. So what's next for Meow? Where do you go from here? Obviously this is a huge swing, bank accounts and virtual cards for AI agents, but where do you see this going over time?

I see this getting massive. AI agents opening bank accounts, AI agents issuing virtual cards, sending and receiving money. You can do that all on meow.com today. So we're just obsessing over that experience and we want to be the only ones doing this, the best ones doing this. We've set ourselves up well to do this with multiple partner banks basically in the long run. So just get more and more agents opening bank accounts, becoming the default there.

It seems like it's in a time and place from a regulatory perspective, you're sailing with the wind here, right? You're not sailing against an administration that's pushing against banking as a service partners. I think it probably would've been really hard to build this business 3 or 4 years ago.

You never want to be too long certain, like administration, especially in the US. There tends to be like a ping pong dynamic that goes back and forth. So you never want to be totally contingent on one party being in power or something like that. So it's important to do things right regardless, such that you have a compounding business. And in either case, we tend to stay away from this only works in this environment, this only works in that environment situation.

Brandon Arvanaghi, this has been great. You guys are crushing it. Where can we send people to learn more about this offering? And I guess your whole suite of offerings.

Yeah, go to meow.com/MCP. The first agents that onboard, we're going to be very close with the customers there. We'll do case studies with you. We'll promote what you did in the process. You get to be one of the earliest customers in history for this. It's like opening the first online bank accounts, basically, from my perspective. So meow.com/MCP, come in and then ping us in Intercom in the widget in the bottom right.

We're Meow customers, but we need to get onboarded to the AI agent thing. I feel like we're a step behind if we're not onboarded in the first cohort here.

Let's do it. Appreciate it, man.

Awesome. Thanks, Brandon.

Bye.

Thanks for listening to another episode of On the Brink with Castle Island. To learn more about Castle Island, visit castleisland.vc. And to listen to all of our podcast episodes, please visit castleisland.vc/podcast or just click on the tab on our website. Thanks for listening.