Self-Driving Finance: The Fintech Promise AI Finally Makes Possible

.Co. Amayas, great to see you. Welcome to Stablehouse Live.

Thanks for having me, guys. It's fun.

Yeah. We awesome. We love the QED team, and, it's been great to get to know know the team a little bit more in the recent months.

We were in Mexico City recently and had some had some run ins with your team. So it's great to meet you. Yeah. Well, we we love Mexico City. We've got great companies down there. I'm sure you guys Bitso has been a huge sponsor and and host of all all things crypto and stable coins down there, and we just, been long Absolutely. Long and happy investors there.

So bring us into into your world, because I I feel like the QED is like been like an onion. Right? I keep on peeling back. I'm like, I I I I think I have it figured out, and then I really I come to find out, you know, you're focused on AgenTeq US, then there's somebody else focused primarily on The U on, you know, Africa. So where where do you fit it into this puzzle right now with QED? Like, where do you spend your time thinking, talking, engaging?

Your your LinkedIn is very clear, Agentsi Commerce, but it's probably a lot bigger. Well, it is. I mean, at some level, our mission is just really simple. We wanna be the best fintech investors in the world.

So that means we've gotta be everywhere where there's fintech. We've gotta be there. We've gotta be on the edge. And we just think that there's so much potential here, and it comes from two big commitments. Right? Two big beliefs. The first belief is, like, AI is happening. Tokenization is happening, but AI is not gonna conquer finance on its own. Right? There have to be finance specific technologies.

Tokenization is obviously gonna be one of those and it's really taking off, especially it's sort of broken containment from crypto to the real world. That's just like a super exciting moment for us. And then the second thing, is that because of these two technologies happening together, I think we're in a position to finally admit that fintech has been a bit disappointing for the last fifteen years. Because the promises we all made at sort of the inception of fintech, like, we're gonna eliminate friction.

We're gonna democratize access. We're gonna you know, you're gonna have a smart wallet. You're gonna have machine learning everywhere. Like, we have to admit, mostly, what we did is we took web forms, we put them on your phone. Like, we got rid of paper. And by the way, NewBank, an amazing company. Credit card, an amazing company. You know? Stripe, an amazing company. It's not that we didn't create great companies. It's not that we didn't move the ball forward.

But if you think about the promises that people in fintech were making in 2010 to 2012, it's only now in this era where we could make those promises real, where money could move without you having to fill out your name and credit card 10 times. Right? So that's what makes me so excited is we're about to enter the era where the promises we've been making for more than a decade are actually maybe gonna come true.

What what was the going back to the original promises that were being made, like, at that time, was it a, you know, what we are now starting to see, which we can dive into, we've you know, people thought that that was gonna happen sooner. Was there a misunderstanding of the problem? Like, what what caused this delay in your view? And and, like, why is this why has this taken so long to finally arrive?

Look, from wherever you are, like, we are people who live at the frontier of technology, all three of us. Right? And so wherever you are, we become intoxicated by just how powerful the latest technologies are. Right? Think about how excited we were about the possibilities of machine learning in 2014, 1516. How excited we've been in about the possibilities of blockchain. You know, look, I explained bitcoin to the treasury secretary in 2014.

Like the first person who ever explained bitcoin to a treasury secretary was me. So like we get excited about these technologies, but because of that, we often don't see the limitations in the technologies. And I think if you go back and one one exercise we've been doing is we've going back to some of our companies that are ten years old. And we're saying, let's look at your deck from 2016. Let's look at your deck from 2014. And what's awesome and amazing is those decks tended to be promising AI based automation.

They tended to be promising, you know, artificial intelligence. I mean, of my great companies, a company called Ocrolis, they do data extraction, analytics, AI for lenders. I've got a a picture of the CEO, Sam Bobley, standing in front of of a sign from nine years ago saying AI based underwriting for your mortgage. And actually, Oculus was doing that at the frontier of what was possible then, but it's it's night and day, you know, to in terms of what they can actually do now.

Another example of this, you know, one of these holy grail ideas we always talk about, Nigel's talked about this, Nigel's my boss, cofounder of Capital One, is the idea that so much money is made by banks because people don't have the time and attention to move it into the right pocket. Right? They don't have time and attention to actually get a dollar here into a yield bearing instrument over there because they know there's gonna be friction here and there and just easier to leave it there. Right?

So the holy grail of self driving finance, I mean people in fintech have been talking about this since I don't know when. Right? I'll I'll tell you a story.

You know, in 2009 so I was at the treasury department in 2009, and my boss, who's now Michael Barr, the who you know, Michael Barr, who's now fed governor, He and I were brainstorming the idea for a government website that would analyze your real credit card spend and then tell you what the right credit card for you to use was not based on your aspirations, but on your actual behavior. Right?

Not based on the points you wanted to earn, but on your payment history and the fact that you actually sometimes need to revolve. Now that was kind of a crazy idea in 2009. It was not like we didn't create a government website to do that. We didn't wanna do, you know, creditcard.gov. But that's when Credit Karma was founded. And Credit Karma still to this day cannot do the feature I just described. But I'm an investor in a company called Kudos that can't and that's brand new.

So Kudos can analyze your actual spend and tell you, you shouldn't have an Amex Platinum because rationally, your Chase Sapphire is where you should put your high, you know, your high cost spend, and then you should move to a cash back card over here. Nobody in the market is doing that today, And more importantly, nobody could do it until we had AI. And that's where things get really exciting. And then you move to agentic commerce, which isn't happening yet.

But I think, you know, I talked about this, the idea of a flood. There's gonna be a flood of interest, but the storm hasn't started. So when's it coming? We don't know.

Would love to dive into kudos. It sounds like you're you've made Oh, yeah. I was gonna say uniquely into this and and would love to kind of compare and contrast what's happening here versus, the statement with Credit Karma. Like, are they,

what is now possible? What are they using? Go ahead. I'd also say just like this idea, like this this gap between what was possible, what's now possible with the technology aspect. What are the other barriers though? I think especially from like a psychological behavioral aspect with some of this, you know, lot people are a little tense with AI. I don't know if I want to do this. So, yeah, how how do we break down those barriers? Okay. Maybe this is now possible. Now that it's possible, what what does it take?

What does it get take to people, you know, using this technology, whether it's kudos, whether it's more agentic focus, you know, tooling and platforms and apps? And then also, how does this influence maybe even the next wave of the next generation of founders, the way they think about what companies they should go build? You know, is it is it taking Credit Karma and helping it evolve? Like, which is what kudos feels like it is. Yeah. I'd love to just understand more about how you're thinking through those.

Maybe that's something Yeah. So look, think Question. This comes down to an idea I call the physics of finance.

Finance is special because finance is about dealing with other people's money. And so finance always resolves to trust. And that was, you know, that's why, you know, it's so central to the founding myth of of of any crypto project of all the blockchain technology is that how do we find new and better ways to create trust. But it's the same dynamic on product. Right?

Like, I mean, one thing that I'm really interested in right now is, you know, this this new wave of personal Concierge's Instinct is is sort of the one I'm using right now. And something has happened that has made in Silicon Valley, people are trusting Instinct in a way they were nervous about OpenClaw. And I don't think it's the tech. I think it's something about the way the product gets strung together, the way the models have moved forward just in four or six months.

And so what used to be hard is now easy, and that creates a massive unlock and a change in consumer behavior. And so that's what we see with kudos. You know, if you go into if you just think about the processing power, the analytical capability to truly do mass customization, and mass customization are one of those old promises. This is tailored to you. Think about when you were like do you remember Helix Sleep, one of those, like, direct to consumer mattresses?

And they were like, oh, we we we, you know, we do a sleep quiz and we tune the mattress exactly to you. Well, honestly, it was four questions and they had three mattresses.

Right?

The illusion of choice and, like Yeah. The illusion of choice. But but if you think about what Kudos can do, they literally take your transaction history. They run it through an engine that has the actual policies, costs, points, rewards of every credit card in your wallet, and they tell you the real economics of if you shifted spend from here to there. And Credit which is a great company. Right? Our our firm was made on Credit Karma. But Credit Karma still is locked into a form.

They ask you what kind of a card, what's your FICO, and then they recommend some cards that might work for you.

Well, then I also think about it from the perspective of, you know, SEO. Like, where are they in what cards are they incentivized to recommend? Is it the best card, or is it gonna be the most popular card that they're getting the highest kickback from? Because, you know, we're obviously involved in the neobate conversation, and then there's a ton of cashback wars going out. Are you using the best card for you, or are you using the the best card for the time because it gives you the best cash back?

And how do you start to think about, like I said, where your finances go, especially with how sticky banking is. Like, I don't I don't really want to switch a card every three months because a promo period a promo period ended.

Unless it's so easy that you just put it on autopilot. And that is an example of the kind of the promise, right, of what's possible now versus what's what was possible. Right? I mean, there was a a pretty good, but not, you know, didn't ever break out fintech company called click switch. Right? Make it easier to switch your bank account. Very cool tool, but it was kinda before its time. Whereas imagine a thing there's another, you know, kind of gen two point o fintech company called MaxMyInterest. Great company.

Exactly. It's the idea of moving in. But that was a little harder, a little hard time, a little harder to break through, a little harder to execute. And in the age of AI, those kinds of promises are just possible. And then I think you're exactly to your point. We've got a layer on trust. Right? There's so much broken trust in our society. There's so much broken trust in our culture, and it's not like AI is making trust easier. Yeah.

So I'm pretty bullish on what's possible, but boy, there's gonna be some really interesting barriers to overcome. And I think that's one of the reasons why we haven't seen more agentic commerce right now, is people are very worried about the trust dynamic.

Yeah. I wanna get into that a little bit more. I do wanna ask, Amayas, have you seen this company called Rivo, r I v o, finance? Really early stage company. I've I've been trying to get them. Drew, I don't know if you wanna bring this up. It's rivofi.com.

Oh, yeah.

This it it very much reminds me of, Acorn's app. But, again, how do you put idle cash to work without switching your bank account and making sure that that they're essentially letting you keep your bank account, earning on funds, and then moving them back when you need to plan around bills. I I've never used this. No affiliation with the team. I've came across some prospecting trying to find new guests.

But this is the first company that kinda came to mind when you said, you know, again, how are we optimizing our finance? How are we using AI to, you know, maybe move our funds around while using the same bank account and really just optimizing, you know, our returns and kinda setting it and forgetting it to a degree?

Yeah. And look. There's huge economic returns to set like, setting it for I mean, you know, in investing, there's huge economic returns to setting and forgetting. Right? But in bank deposits, those economic returns go to the bank. I mean, I will I I'm embarrassed to admit this, but my primary banking relationship is from the same bank that I signed up with when I was nine years old. It's Bank of America.

And I recently, was we're having this conversation, and I said, let me go ask Bank of America whether they're offering like, whether I'm actually at least in the highest rate the Bank of America will offer for a deposit. The highest rate, I'm in it, it's under 10 bips.

Wow.

I mean, that's it. That's embarrassing for me. Right? Yeah. I I have since corrected that. I haven't moved my primary banking relationship, but I moved most of my liquidity out. Know, because there's, you know, another good example of this where, I'm on the board of a company called Atomic Invest. Atomic is an API first broker dealer.

They power the high yield savings for, crypto companies like Uphold, for direct to consumer companies like NerdWallet, and and for other companies in our portfolio like a treasure Meow, you know, which offers high yield business checking. And it's that kind of technology which needs to just be everywhere to your point, whether it's Rivo or someone else. This is the kind of idea who whose time is now. And the idea isn't new, but the ability to execute on it is. And that's what makes us excited as fintech investors.

Yeah. Can you speak a little bit more to the unique advantages that you're kind of hitting on a little bit with founders right now who are building businesses, entering the market. The, you know, the Meow is a perfect example, like these companies who are taking a a different approach or a first principles approach that is more, you know, perhaps stablecoin native or AI forward. Like, what are you seeing founders do really well right now, and what are you hoping to see more of?

And, like, would love to just hear your yeah. Maybe your perspective or advice to founders who are actively raising or or, you know, things that as a as an investor or a partner on the other side of the table, you guys are really looking for right now in this crazy, like, intersection of AI and money and and what the real opportunity is that you want more founders to go chase?

We love the idea. I love the way you put it, AI and money. Right? There are emerging rules, emerging properties of AI. It's incredibly powerful. It's incredibly creative. But, you know, we're what are we in now? Three and a half years into really using this technology every day? We can't stop it from hallucinating. Heck, it just committed felonies. Like, we don't know how to control this technology sort of as a general matter if you just let it loose. So we've got to lock it down.

And then the opposite, there is a physics of finance. If you are dealing with other people's money, yes, there are regulatory rules, but all of those regulatory rules emerge from very old traditions. And that's what gets us excited. Can you bridge the power of AI with the solemn guarantees of finance, with the mechanistic systems of record? Can you find a way to bridge those two things? Because that's where we think the present interacts with the future, and that's where the money gets made in our view.

And then the second thing, which is just, you know, sad but true, there was a moment two years ago where you could just say, like, because AI. And, like, your customers would be like, I'm buying. You've got AI. I'm buying. Right? And so the the mantra to every founder was just get AI into your product. Add a chatbot. Add something. Add something that makes it look like AI. AI is the thing that's winning. I'm sad to say, AI is just a normal technology. Everybody's using it.

So now customers are back to their state of nature. They're trying to figure out who is the best going to solve my problem. And that means, we like to say old lessons are good lessons. The oldest lesson of y c is build something your customers want. And man, that's a great lesson for founders. AI is just a way to deliver for your customers better. And that's what we think. So it's it's sad, but true. Like, you can't just throw AI at it anymore.

You have to figure out the old hard problem, which is what can I do for my customers that other people can't? And that's the way to make money.

Do do you think in this conversation and, is there something deceptive to a degree if you're building a product that's very AI forward and you're not disclosing that to the end consumer? I think about especially from, like, a data perspective when you like, I agree with you wholeheartedly that, like, hey. Build the best product.

Probably don't put AI on it, but there is the part of me in the back of my head that goes, but if you're inadvertently using ChatGPT, you're using these that are using your data to train their models, do you do we need disclosures around that? Do we need do do consumers need to be made aware of that? And there there's there's a little bit of attention there for me because I do want the best product, But like you said, AI is a little bit we we can't tame it. I don't want more and more data out there.

How how do you think through that problem?

I think this is a really hard problem. I think it is a little bit case by case. And so let's take at the base level. I don't think people need to disclose, hey, FYI, our code was written by AI. Right? If your code isn't written by AI, what are you doing?

Right? Yeah.

At the other end of the spectrum, you know, we see very consistent feedback that regular consumers in most conversations actually prefer an AI agent to a human agent. That's counterintuitive. I talked to an HR leader at one of my scaled companies yesterday. They have an opt in to AI interviews for the first round. 95% of candidates opt in. And the reason she said is that they you know, the best candidates aren't are are already employed.

And so if they opt in, they get an immediate interview, so they're not waiting to schedule. They can do the interview on their time, whether that's 5AM or 5PM or midnight on Saturday. It doesn't matter. They get the interview, they do it and it happens immediately. And so that's kind of an example where, yes, fully disclosed, fully opt in, but people are actually preferring it. Now when you get to the next thing and you say like, you know, who am I interacting with? Is it a bot? Is it a human? Can I know?

Those are places where I think it does get complicated. And I think, at least for now, companies are gonna make different choices. And I'm not sure what the best practice is. Right? Like, if you Yeah. If you're answering one question on a website, do you meet do you really want it to spend twenty seconds, like, explaining to you that it's an AI before it just answers your simple product question? I don't know.

So I I think this is but this is a place where I think a lot of companies are gonna make mistakes because they're going to get that trust equation wrong. They're gonna go too far. They're and and I think that's it's a it's a tightrope for sure.

Yeah. I think about perfect example right here. I just used this service called Pack Up and Go. They do, like, secret planning trips. They're very upfront of, yeah. We don't use AI. A person is planning your trip. Now I think about if somebody wants to launch a competitor to them, and maybe they're augmenting all of their agents with AI, and they're also positioning us, but we don't but AI is not planning your trips because the person's still reviewing it. And it's like, well, which one is true?

And there's like it all revolves around trust. And there there's something that makes you inherently wanna trust it more when you know that AI wasn't all the way involved. And and I I don't have a good answer too, but it's just branding, perception, trust. How do you position your company while, you know, being I mean, do the pickles taste better

because they were handmade in Brooklyn? Like, I don't know.

Yeah. Sometimes.

What does that mean? I love a I love a pickle from the farmer's market. But I also love, you know, you know Yeah. I do think you will you will find, each company is gonna have to fill its brand promise in its own way. And I'll say, like, you know, look, we're in fintech. We are focused very much on, can you deliver with the accuracy that finance requires? Many of our companies are actually selling their customers in a b to b context peace of mind. If you work with us, you don't have to worry about AI.

You know, an example on this screen is a company called Collective. Collective is a back office for, you know, accounting, tax, some back office stuff for solopreneurs and and very small biz like, very, very small businesses. So maybe they would have two or three or five employees, but really it's for solopreneurs. And if you're a solopreneur, you're probably pretty excited about AI. But also, do you really wanna have to figure out, Claude, for small business?

Do you really wanna figure out when it's gonna make your accounting right, when it's gonna make your accounting wrong? Do you wanna error check that yourself? Probably not. And so what collective can do is say, hey. All that back office stuff that you know should be better with AI, well, don't worry. We've we've put that together, and we stand behind it. And you look at this claim. Right? Members save an average of $10,000 on their business taxes. That's a really powerful claim.

Now, of course, they're doing it with AI, but they're also standing behind the promise and they're taking a thing that the average solopreneur doesn't want to deal with, which is optimizing their back office. Because what should the solopreneur doing be doing with AI? They should be worried about how do I use AI to serve my customer, to delight my customer, Not how do I use AI to optimize my back office. So let collective optimize your back office and they offer the peace of mind.

You will always be on the front end, on the on the front edge, on the frontier of optimization with AI because you're working with us. And then you can go focus on doing the experiment, doing the thing that makes you different in your community. I

know we're kind of, we're rounding rounding the end of this a little bit. I just have to ask about stablecoins, just high level. Like, would love your comments on how you've seen stablecoins as a movement or an asset class, if you will, really affect the portfolio. You like, QED seems to be in a really unique position where you're you're you're helping back new companies who are stablecoin native, like the Meows and the Augustuses and these companies who are starting from that perspective on the world.

But then it seems like you also have this really interesting position where you can go to fintechs who maybe started five, ten years ago, who can radically accelerate, stay defensible, and like adopt this new technology. Like, how have you seen both sides of this and how are you seeing stablecoins like within the portfolio right now?

One thing I love about stablecoins is stablecoins solve a very real problem in the world in addition to the promises they're making about what's possible in the future. And the very real problem they solve in the world is prior to stablecoins, there was no US dollar denominated twenty four seven global payment rail. Very simple promise. That is something the world really needs. And a lot of the stablecoin usage is coming from that. Now there's other stablecoin usage. A lot of it is crypto based. Right?

We need a settlement layer for for crypto trading. Okay. That's interesting. To me, that's most interesting as what it portends for other real world assets. And then there's other things that can come from this Programmability. Right? What about instant payment? You know, most business payments don't need to be instant. But boy, if they were instant, if I could trust that they would be instant, well, could I run my business differently? Could I run my liquidity differently? Could I negotiate differently?

Imagine the following conversation. Right? We're to b to b conversation. You say, if you sign today, I'll give you a discount. Great. If the cash hits my bank account today, I'll give you a bigger discount. Well, prior to sale points, you couldn't make that promise because even if I sign and wire right this day, it's gonna be five days before it settles with any guarantees. Right? So just imagine, like, what is possible in the world of business negotiation when you add programmability control on top of instant.

Yeah. So I think today, it's mostly about global payments, and that is where we see absolute ripping demand. But tomorrow, it's about the extra control, the extra certainty that comes back to that issue of trust. How do I know exactly where the money is, exactly where I need it to be, and can I move those things instantly? That's a very powerful tool. It makes us super bullish on this space.

It's awesome.

We're making it rain as you need know. You had a few areas that I wanted to go. Maybe next time we'll have you on. I would love to have, like, a US centric conversation. Oh, absolutely. I think a lot of times we go emerging markets, and I think especially with payments domestically, it's something that I'm like I said, merchants want their money faster. They want lower processing fees and all the world of AI. So maybe that'll be part two of my thanks for Part two. Well, anytime, guys.

Great to great to be on with you. Yeah. Great to meet you. Thanks for all the work you're doing to support the space. There's so many opportunities there. Companies we need to cover that are not as on chain native, but more fintech and bleeding into this world? Happy to Yeah. Look, I'll I'll do one I'll do one big, one big pitch here.

Everybody in the crypto ecosystem has gotta be worried about fraud, gotta be worried about AML. Like, the more that crypto becomes, you know, US connected. And so we've got this amazing company called Footprint, that is the identity layer and fraud layer that connects the hard stuff of that old KYC, know your customer security, and the modern stuff of have an agent who can do investigations, who can figure out whether there's a US nexus to a person.

So, like, it's such a good example of someone who is crossing that whole hard and old mechanistic stuff with the modern and the truly agentic workflows and already working with some of the biggest names in crypto. So great one for your for your audience.

Yeah. Shout out. That sounds like great. The invite has been made. The invite has been You'll see a clip.

I mean, Eli is like the most fun. He's super fun.

Cool, man. Maez, thanks for the time. Thanks, guys. Cheers. Thanks for coming on.