Know Your Agent Is the Most Important Problem in Payments

Sy Taylor  0:10  
Welcome to Tokenized, the show focused on stablecoins and the institutional adoption of tokenized real-world assets. But this is our agentic commerce series, and I am Simon Taylor, head of marketing to Devon Tempo, and author of Brain Food, and I'm joined as always by my agentic commerce co-host Bam Azizi, CEO and founder at Mesh. How the heck you doing, Bam?

Bam Azizi  0:30  
Thank you, Simon. It was mouthful introduction. Excited for this.

Sy Taylor  0:35  
Hey, what can I say? I'm living on very little sleep and a lot of hope right now, and a lot of hope that we're going to have a great conversation, and I don't think we'll need it because joining us today is the fantastic Alfonso Gomez Jordana, who is the co-founder at Crossman. How are you doing, sir?

Alfonso Gómez  0:51  
I'm doing great. Thank you for having me again.

Sy Taylor  0:54  
Thank you, man. Like I think this is going to get really, really deep into the weeds. But before we do that, I've just got to remind viewers and listeners that views and opinions of our contributors today are their own and might not reflect those of companies they represent. Please don't take anything we say as tax, legal, or financial advice. And of course, a reminder that this episode is made possible by Visa and our friends at Mesh. This episode is brought to you by Mesh, the global crypto payments network. Mesh makes it possible for anyone or any agent to get paid or pay instantly from any wallet on any chain anywhere in the world. You can learn more at meshpay.com. All right, Alfonso, set the scene for me here. You guys have been building in this agentic commerce space for a little while. What is it you're actually building, and what opportunity are you trying to get hold of?

Alfonso Gómez  1:49  
So to catch up the audience, if they didn't watch the other one, in general, Crawsman, we believe that most economic activity is going to be intermediated by agency in the future. The reason I think to us is is obvious is that making financial decisions is hard, and when you have an intelligent being in your pocket that can take better decisions for you, it's a no-brainer that it can do better use of your money. No, and so at Crossman, we are a financial infrastructure company, and our focus has been on helping the agents or the agent platform side. In particular, we we see ourselves as you can call it something like a PSP for agents with the tools that we're developing there. Basically, be the helping hand for companies that are hosting their own agents, developing their own agents on anything that's related to financial aspects. So, particular, we've been iterating our tools here for over two years as the space changes by the week, but right now our offering is focused on enabling agent customers to delegate payment methods, credit card and stable coins today, adding a few more in the future to their agent, and then letting the agent use those to purchase things, to buy things online with credit card securely and compliantly, to be able to do machine payments on Tempo with MPP, X402, etc., all in a single box that aggregates all this functionality fully, whatever. 100%.

Sy Taylor  3:14  
Bam, I know you've been obsessed with agency commerce forever and and and the stablecoin side of it, but I'm interested so much in what Alfonso is saying about stablecoins and virtual cards together. Kind of curious for both of you, gents, really, your thoughts on that coexistence.

Alfonso Gómez  3:32  
On our end, funly, I think I may have mentioned this last time, but when we started powering or trying to power agentic commerce, we were 100 on stablecoin land, giving agents a wallet, giving them payment APIs with stablecoins and whatnot. And what we found is that while agents may be able to use stablecoins and you know understand what to do with them, their agent owners, the humans behind them, and the platforms that are developing those agents, no one wanted to touch stablecoins. No, what we realized back then is when you're thinking about the GenTech payments, there's two legs in a payment where before there was only one. There's the what does the human delegate to the agent, and then how does the agent execute the payment? And we don't necessarily think that both may be the same payment method, no. And here's where I get to where where each one will be used. Our expectation is that for the first leg, which is a human delegate a payment method to their agent, I think the vast majority of that is going to be the payment methods that we are familiar with that have points, that have customer protections. I think most of that is going to be cards, maybe bank accounts. There's other alternative payment methods as well. Now for the second leg, the one that the agent executes, I think there the agents are going to be pretty comfortable using sometimes cards, and if they can extract a delegated card from our virtual card from the original payment method, and other times they'll use stable coins, and they will just use the path of least resistance. No, and so we can talk later about where. Each of them will be least resistance, but high level, my answer will be human to agent probably cards for the majority of the use cases from agent onwards for agent to pay against the merchant. That's what we're going to be seeing, I think, a mix.

Bam Azizi  5:15  
My thoughts are like we have all the tools to serve humans as of today. I think the the future of agentic commerce would be very very different. I think we we won't have like browsers anymore. Everything will happen on the API level. What Alfonso and team are building is very interesting because they're building toolkit for the future of economy, which either agents interacting with other humans or agents interacting with other agents. One thing would basically won't change, which is like stablecoin is going to play an important role as a settlement layer for a lot of those transactions. And the more we advance, the more I'm confident that that would be part of the the recipe of agentic commerce.

Sy Taylor  5:58  
What I like is that separation, Alfonso, of like human to agent, but there's also human to agent. Human gives agent their card, and then agent buys like the instinct or the muse example. Agent buys thing is very different to buys resource, buys tool, buys something from another agent. And I think it's those use cases that people kind of get mixed up, so I'd love you to to just kind of unpick that a little bit for me and talk about like who you're seeing, who's already working with you, that's doing those different flows and and what those flows look like.

Alfonso Gómez  6:33  
At least 90% of the volume we see today on agentic side is all cards. No, it's similar to the Instinct Flow, we do something similar to if you're familiar with Stripe Link. The only difference is like it's more white label, so you don't see the cross main brand everywhere. But there's many iMessage agents or things like that that they're asking, hey, can you save a card on file somewhere and then later go and buy a shoe for you with a virtual card for it. I would say that's the vast majority of the traffic today, especially after Instinct launched. Prior to this, the majority was actually on the stablecoin side. On stablecoin land, we see a little bit of X402 and things like that, but I would say that today that's still rounding error, and I don't think I think that will change in the future. But I don't think we've had the instinct moment for machine payments, or at least it's not something that we've seen.

Sy Taylor  7:29  
What a fascinating way to think about it! It's like you still have conviction in the like machine payments, but you've not had the instinct moment. But like, talk about that like toolkit you've built, because there's there's genuinely quite a lot of people trying to do stuff in agentic commerce these days. So, who's trying to do it? What's the developer trying to do, and what's in the toolkit?

Alfonso Gómez  7:47  
Our toolkit is basically like the compilation of all the products that we've been building. But high level, I start with the card side because it's what has most demand today. We have in in a single SDK. You're an agent developer. We only work with agents, not with merchants or the other side, you're an agent developer and you integrate this toolkit. You get one the ability to securely collect cards in a PCI vault, so that you're compliant and and whatnot. Second, you get the ability to obtain virtual cards or it's more like card tokens from Visa, Intelligent Commerce, Mastercard, Agent Pay, and a few other networks are launching soon. And these are like imagine the same way when you use Apple Pay, you can get a virtual card number. It's the same thing. You get a virtual card number that your agent can use, with the additional feature that it has guardrails built in. That card, that virtual card, may only have let's say $200, or may only work to buy electronic goods or things like this. So that's kind of getting the payment method to the agent. And then the other part we have is we also have an execution API. It's a we call it agent checkout API, and you can think of it as a. I think the right mental model would be a sub agent that your agent can talk to via MCP or REST API, where you say, "Hey, buy me this shoe on Amazon," okay? And what this subagen basically does is it wraps basically everything under the market that you can think of to buy that shoe as fast as possible with as minimal token spend as possible. So it wraps UCP if a if there's a headless checkout that supports it. It also uses browser automation with a custom harness that's optimized for checkouts, trained over 10s of 1000s, hundreds of 1000s of checkouts. It also uses stealth, residential proxies, ISP proxies in orchestration, and then all the KYA protocols as well. So we basically aggregated all the tool toolbox of all the things you you need to do to maximize the conversion, to maximize an agent being able to buy something, and then can do this in under one in around one cent of inference costs. So basically, if you want your agent to be able to buy things, you can get the payment methods with us, and then you can also execute the order, get the agent to our sub agent to execute it for you. The last thing that the SDK contains is all. So in the same integration, you get a wallet for the agent, and you can on ramp and off ramp from those save card and files. So you save a card to buy a shoe, and then the next day the agent is like, "Hey, I need five bucks to get a video credits at Krea AI for X402 or something like that." The agent can know how to can know to request, hey, I need $5 in USDC from this safe card that I already had, and then can go and execute it. No, and all of this is kind of building this, this like this toolkit. And there's a couple things coming. Probably by when this podcast is out, there's probably one or two more things coming out as well.

Bam Azizi  10:36  
One question for you. So in that world of like coexisting card and basically stablecoin layer payment. Do you do you see the card type transactions can get any benefit of blockchain technology through a smart contract? Whether you want to add card rails or some thresholds or some rules and regulations, can blockchain play any role on on those type of transactions,

Alfonso Gómez  11:01  
I think they do, and in fact, both Pisa and Mastercard have spoken about, I think, more experimental proposals or protocols around this. I think in general, blockchains can provide like a global coordination layer, no, like a a ledger where you can maybe have things like store things like reputation or or things like this. At the same time, I would say the challenge here is going to be something becoming a standard and people like jumping the extra hoops to get there. And I would say beyond theoretical, I haven't seen anything practical to date. Now everything moves really fast. We are on an AI takeoff, and it's possible that AI chooses to use this in next two months and it blows away. It blows up, but at this time, I do see the theoretical advantage. But in practice, everyone agreeing on it, I think it's a bit far out. But I wonder what you think, Bam. As you you had someone seems like you were interested in this.

Bam Azizi  11:52  
No, no. Like it's obvious that we will have like cards and a stable coin coexist. But was just beyond money. I thought like maybe blockchain can help help on that end. Like Simon knows my take. Like we, I believe that before in next 12 months we will have like agents transacting with each other more than agent transacting with merchants or or humans, and as a result of that we will see a big spike in terms of like microtransactions that you cannot use traditional rails. So you have to use stablecoin if you want to send like 0.0005 cents. You cannot use Visa card, Mastercard, or traditional wire transfers or bank transfers. All boils down to basically when we will have more agents transacting with each other versus agents transacting with humans or merchants, like what we are seeing with Instinct and Muse.

Sy Taylor  12:51  
I think that way that Alfonso framed it of like we're waiting for the Instinct Muse moment for the machine-to-machine commerce. But to Alfonso's point, superintelligence could be in like explosive mode in two months and takeoff scenario. Just when you think we're pacing the frontier, outcomes Zuck to be like, nah, look at this. It's called wonder. It's state of the art. So yeah, I'm kind of curious. Bringing it back to payments, Alfonso, the like liability here because card networks they can kind of make liability make sense. There's a lot of things that go wrong if an agent buys the wrong thing, or an agent is compromised, or you know you've given an agent a budget and it runs off and does a bunch of things. How do you think about that emerging space? Because it feels like it's still messy and there's like 15 different protocols. I came across x4 101 the other day for identity, and now there's PAP personal agent protocol, and there's obviously MCP. There's all these things like a verifiable intent from Mastercard. What did the user intend versus what did the agent intend? Curious is like where all of that sits in your head.

Alfonso Gómez  14:08  
I think this is very similar to, or it can be analogous to, you know, when they see these swarms of OpenAI hacking other companies, and there's this, you know, OpenAI and Anthropic saying let's get regulated. We need to figure out who's liable, but then I think the government or someone was like, "Well, you are liable. You're the one that hacked other companies." No, and I think ultimately the legal framework for these things needs to be figured out, and legal as well as like card network framework for these things is still iterating, as far as I can tell. But high level, my opinion is that we're going to have to look into the details of a transaction and attribute who screwed up. No, that we're going to need actually better attribution and proofs of when someone said "buy me a potato," and the agent bought them a tomato. Was it that the person actually explained it incorrectly? Was that the. Interpreted the intent incorrectly, or ultimately was the the merchant delivered the wrong thing? No, I think we're going to have to get to that level of complexity in order to be able to make this this work down the line, which brings complications. You may need new protocols like PAP or verifiable intent, but also you need something that works universally, ideally, and all these platforms can can implement, and then some kind of review that they're implementing it properly. So it's going to be, I think, messy until it actually works out. To me, the interesting question is more like, when is it's not how is this going to be resolved, but when is this going to become a big issue, and who's going to be the one with the stick instead of the carrot?

Sy Taylor  15:39  
So I can argue it's already a bit of an issue with Amazon blocking Muse and J.P. Morgan in their earnings call saying they're now going to block personal agents as well. Like if companies are blocking it, it's already an issue. If shareholders are asking about it, like the volume might be low with what eight, 10 million users on Muse or whatever, and similar sort of number on instinct around the world, but I also know that I think Target has said that the liability sits with the consumer, no matter what the agent does. Like, just it's with you, which is not necessarily what Reggie says. And so, there is going to be like a lot of stuff for the courts to maybe figure out in the interim of like getting these these standards in place, so I agree with the overall messy, but but I can sort of see. I think that problem on the consumer side is nearer than we think versus maybe a little bit further away on on some other stuff. So yeah, really really interested there on that protocol soup. And how are you choosing between you know, sort of the browser automation, something like a UCP, something like Shop Pay and the wallets, like because that seems like another layer of complexity. Of there's Web MCP now. Unpack what some of these things are, and then talk to me about like what their relative pros and cons

Alfonso Gómez  17:01  
are. So I think this is part of the soup of standards, specifically around how do you execute a checkout. I would say UCP is probably the leading protocol now of being able to do a checkout through an API call, being able to specify to an agent, hey, this is it. It covers very broad scope, and different clients may implement it differently, but high level it can cover from discovering these other things I offer down to execution. Like go and make this API call, send me a payment credential inside, and I'll buy it for you. WebMCP has focused a bit more on the discovery and not the payment execution, and then Shopify now has an MCP that Shopify releases a thing every week, but I believe that the latest is Shopify stores expose UCP. I believe MCP as well, and then they also have an MCP that you can install on the buyer side to be able to buy on their stores. Although that's not required, an agent could also buy through the other standards. This is very complicated, and it changes every week. I would say today the only winning strategy, if you want high conversion, is like to aggregate everything. Unfortunately, probably in a year, maybe UCP and browser automation are sufficient. I hope, but unfortunately, right now there's more protocols I think than than purchases. Or at least that was true two months ago. Now there's a bit a bit more with instinct, but the other thing I would say though is the reality is browser automation is still like 99% if not more than all these commerce. We barely see in the real world any store that supports UCP down to checkout. The only ones that do are Shopify with pretty much with Shop Pay as a required payment method in the majority of cases, so reality, even though this looks very messed today, it's browser automation, and if you get blocked, it doesn't work. And on our end, we're doing everything we can to iron out, and we can get probably 10, 20% more conversion than a standard browser automation. But it's still going to be this cat and mouse game until the entire industry updates. No,

Sy Taylor  19:03  
conversion is just the key metric, isn't it? Like if if something is converting worse than than you actually going to the website yourself, you're not going to use it. And if agents are converting worse than having something on their website, they're going to hate it because this is like losing sales. I don't want this thing.

Alfonso Gómez  19:19  
And one interesting thing we're seeing going back to what you said about Chase and Amazon, is that there's actually been a very strong reaction amongst merchants against Muse and Instinct first, and the reason is right now the agentic commerce is starting, and the reality is we're at the phase where people at Twitter talk about it, but the incremental volume that these folks are getting is still close to zero, whereas they see a lot of you know the compliance teams, the risk teams, and all of that like see all the alarms. So everyone, all the big ones are taking the conservative approach. It's only Shopify and maybe like smaller merchants that are taking the more progressive approach here. And what what I'm curious about is like how far. This is gonna go, and if the others are gonna regret it, or are they are they gonna switch quickly, or or be kind of the like Amazon maybe be like a year, two years until they accept it? I don't know what's gonna happen there.

Sy Taylor  20:10  
Yeah, I mean, I can't say I blame some of these merchants given the amount of things that could go wrong. You know, Amazon was saying something like, "We don't know who this customer is. It hasn't identified itself. Is it acting on behalf of somebody that is our customer? Is it fraudulent? We've no idea. So you know, like, let's pump the brakes on this. And of course, they've got their distribution network. They can get something to you same day. And they're 20% of commerce in the United States. So they're in a position where they can say no and hold out, and everybody's going to come and hopefully bend the knee and and want to work with them, versus some of the smaller merchants that are mid-sized that might not be sophisticated enough to block, but might not be also sophisticated enough to figure out what Shop Pay is doing and lean into some of those protocols. So there's like this messy middle that's going to be really interesting to watch. I'm going to bring us to part two shortly because I think there's a whole lot more on this this risk conversation that we could jump into. But before we do, I'm just going to take a quick pause while we hear from our sponsors that make this series possible. Stablecoin operations usually mean a wallet from one vendor and on ramp from another, and then controls stitched together across all of them. Visa's stablecoin platform fixes this fundamentally. You can mint, move, and manage stablecoins across OpenUSD, and you remain your own custodian, all in one single environment. Then, stablecoin linked cards let you spend balances anywhere Visa is accepted. That's Visa, the global leader in payments, and of course, sponsor of this show. You can find out more@visa.com forward slash crypto.

Bam Azizi  21:59  
Alfonso, that was a great conversation, very eye-opening. I think KYA or Know Your Agent the same way that we have KYC or Know Your Customer is going to help a lot in terms of like detecting fraud or preventing fraud or future fraudulent activities. What's your take on that? And also my follow-up question in terms of like adoption. Do you agree with Simon that like companies like Amazon that they are also a merchant as well, right? Amazon is doing a lot

Alfonso Gómez  22:30  
of selling a lot of its own item out of their own inventory, and like maybe they can take a bullet or two to make sure that the agentic commerce become mature enough to be at the same level of like human level activities from fraud detection, from conversion rates, and all sort of like key metrics, right? So, do you think big companies can play an important role to nurture agentic commerce? So, I think to me, KYE agent reputation is the most important problem to solve to enable agentic commerce long term. No, and so right now we're figuring out okay how to collect payment securely, how to actually execute the purchase. No, but all these things, you know, I have no doubt that they will be sorted out in you know maybe a year time or or something like this. The interesting part later is, yeah, determining who's this person buying on behalf of. Like a lot of the margin that merchants get is actually derived from better understanding the customer and discriminating in one way or another. From very directly, like you see maybe like the shopping mall to you know, Amazon. I think a significant chunk of their profit is actually ads and how they reorient the page much more than you know just having the item for you.

Sy Taylor  23:49  
$68 billion of their revenue annually is advertising. So if there's eyeballs not on the website, that's $68 billion they potentially lose. So yeah, that's that's quite an important thing that you're pointing out there.

Alfonso Gómez  24:04  
There's a chance that those business models just go away. Maybe there's no value you can extract from that. The other one is I do think part of it will remain, but it will require really understanding who's the agent and who's it buying on behalf of. So I believe that economic forces will require us to do this, and I think that maybe the unblock, to some extent, for Amazon and others to get comfortable is being able to access more data than they have to today. No, and that's, for example, what Chase said publicly. But to me, the challenge is who's going to initiate it. No, going to your second question, who's going to be the one that triggers it? My guess, not that it's Amazon going to be. It's more like it's going to be OpenAI or one of those that maybe, maybe maybe it's Muse if it gets enough traction. I'm not sure, but I think it's going to be one of those platforms that starts driving enough meaningful volume. Will be the one that you know will collaborate with one of these. Figure out okay, I'll give you two of the five data points that you need, and will probably trigger it more than the merchant side. Will be my guess. No, because merchants at the end of the day don't have an incentive. They they have an incentive for things to remain the way they are. No, to like continue increasing 10% year over year that adds revenue and those things. It's the agents, the ones that are poised to benefit from disrupting the current model. No,

Sy Taylor  25:17  
well, the agents and the the consumers, because there's that whole sort of, I didn't read the policy and know I could get a refund, but only if I get it in before 3am. the next day type of stuff that an agent will quite happily do. I mean, Muse is even selling itself on this ability to stay on hold for two hours because you couldn't be bothered, or like you have life, frankly, to deal with, and you can shop around and get a better price. The best price isn't always on Amazon, but the fact is, it's probably there, whatever it is you're looking for. So you don't have to do a really big search. I think dealing with that, like having agents that are just infinitely patient and can find all of that value for you, is adversarial to the business model of people that monetize that inertia, and that's sort of going away a little bit. But but I wonder about that that merchant side. You know, there's obviously what Shop Pay is doing. I also think about the developer side as well, or the B 2b side. And when you when you're buying there outside of the classic commerce, so if I'm sitting in Codex or Claude Code in a couple of years' time. What am I giving my agents? What's what am I using, and and how am I preparing as a merchant or as somebody who's created an agent for it to take sales? What what does that look like? If we've got this like instinct pattern for consumers now, is it is it looking different for the developer merchant sort of ecosystem?

Alfonso Gómez  26:40  
One thing I did want to mention with regards to the prior point is I think even though part of this conversation may sound pessimistic, okay, Amazon is blocking it, what have you. To me, the interesting thing is to start looking at. I think there's going to be merchants like when e-commerce came and when mobile commerce came. My guess is like the merchants who embrace this and figure out what works and doesn't work for agentic are just going to be like massive companies, and I do see some very big like retail brands and things like that that have like full teams are fully committed to it. I think it's going to be interesting to see which merchants actually end up becoming 10 times bigger thanks to this, and which ones get disrupted. No, but I don't think this is going to be negative for merchants or something like that, which may have sound before. I think it's just going to create a lot of changes. Now, to your question, how does it look up from B 2b side? And I think your question is more around if an agent wanted to make you money, like accept payment, or are you saying?

Sy Taylor  27:31  
So I was imagining I'm a developer. I want to make agents. It might need to buy things, so I want to give it a card. Is there a wallet sitting in there, and it might need to sell things. So it's like that two-way door of I've created an agent. I've probably done it inside one of these coding harnesses. Is my wallet baked into that, and is my account baked into that? And then it receives money, or one of my agents receives money. What's that account structure look like? What's that sort of looking like? Do you see what I mean?

Alfonso Gómez  28:03  
Yeah, yeah. I mean, today the reality is there's nothing like standardized for all of this. To make payments, you can do the same thing that consumers do. So, I mean, you can put your corporate card in there, or you could give a stable coins if you wanted to, and you can buy with stable coins. You can buy APIs and things like this. The interesting thing is acquiring, like making money. Today, to the best of my knowledge, agents can only acquire legally, or you could say legally, or illegally with stablecoins. So an agent can trivially the same wallet that they're using to pay, use it to accept through MPP and Tempo or X402 and Solana or Base and whatnot, and this is just a like something that the agent can do trivially. So if you want your agent to accept payments in stablecoins, they can do directly. Now the challenge is who's going to pay for that, no? So I think other agents that have a wallet will be the main. If they're selling something to humans, then the best thing the agent can do is create you an account in a traditional acquire in a like PSP like Stripe or Checker.com, but unfortunately that's going to require you as a human to do KYC, KYB, and all of that. There's nothing right now seamless that I've seen for payment acquiring for agents. Something like maybe a Square terminal or your phone. Those things don't exist today that I've seen for agents, but I'm sure it will be coming in the next few months as well.

Bam Azizi  29:24  
One question for you: Do you see? Like, I really I was fascinated by the point that you said, like the merchants who disrupt, basically leverage the agentic commerce or agentic toolkit will disrupt the the incumbents. When mobile commerce happened, we had new type of companies like Uber and DoorDash, they come to life, right? And they become like worth like 10s of billions of dollars. Do you see new type of merchants are coming to the market that they didn't exist today or till now, and then now they're going to open like or create a new industry or new. Segment in the market that they are agentic native.

Alfonso Gómez  30:03  
I think so. I think one of the ways I would think about it is maybe connecting with Simon's PowerPoint. But I'm thinking here about the lovable vibe coder. Maybe people use lovable aim or codex, cloud code, or what have you. You can be at the toilet and say like, create me like a you know GTA six before GTA six launches like that's the capabilities they have now. So of course you many people will be asking create a you know store to sell maybe custom T-shirts maybe their food etc. And I think those people went back to the earlier point I think they're not gonna have the patience to go to a traditional KYB or KYC underwriting process and everything from a PSP, so I think there is a gap there in the market of if I can create a full production, a full website, and everything you know from the bus and from the toilet in like five minutes, am I going to now spend hours figuring out all this paperwork? I think there's an opportunity gap there that once it's covered will probably enable new kind of merchants. Let's call them buy merchants. Like you can create a store sell things services, which much lower friction than before. I think there may be like micro businesses that people start creating. I would expect there's going to be a lot of that. The other one is on machine payments services native for agents. Admittedly, I think it's going a bit slower than I thought. But being able to, you know, build things that are just for your agent. If you think your agent gets stuck on this or that, being able to tell them, all right, if you pay me one cent or $1, I can serve you directly, and then the agent is the one that discovers them and whatnot. I think there, there's also going to be a lot of interesting opportunities. I'm again surprised that there hasn't been that much. Maybe we're missing something. I don't know if it's the App Store construct or something like that. So people start using those tools a bit more. But when that comes, I think that will be definitely be a new kind of merchant. Even though maybe it's not like an Amazon competitor. Maybe they're selling something else now.

Sy Taylor  31:58  
I'm going to change tack here ever so slightly and talk a little bit about your founding contributor position you have in the legal context protocol with the AAA Google and Circle. Why is that? What is this and why?

Alfonso Gómez  32:16  
So here we're talking about payments. Payment is really transmitting value, like terminating money from point A to point B, but when you think about commerce, agentic commerce per se, or commerce in reality, you're really actually executing a contract. You're saying, okay, I'm paying you $100 for a shoe, and then you're liable for sending me this shoe in an agreed time period with certain return policy, etc., etc. There's actually many more things that you're agreeing to, no. And this is still very abstract and loose. But my thought here, I mean, I'm one one of many, and this organization is in the same way that we are starting to put together protocols how you move the money or how you represent a you know tokenize a credit card to pay and how the liability is going to shift, which is one of those conditions. I think encoding in a standardized way. What are the terms and conditions? What are the terms that the agent saw? The attestation that those have been seen. You know, maybe like an interpretation that is unambiguous to machines. Those kind of things I think become important down the line. You know, right now when an agent helps you find a shoe, finds the cheapest one or the one that's closer to your favorite color brand or what have you, but once we're past that point, I think they're gonna need to be comparing legal requirements and what legal conditions and whatnot. And I think this just makes it more accessible, more adaptable, and whatnot. No, so I think it's important. Maybe step we have payment methods, then executing checkouts, then KYA. This is probably the fourth one, but probably will be comfortable in the future as well.

Sy Taylor  33:43  
So interested to see. There's like so many emerging spaces. I don't know how you stay on top of it all, Alfonso. Day to day, there's like whack-a-mole or something.

Alfonso Gómez  33:54  
I clearly have an agent, an agent swarm, just to keep me up to date on the things that are coming up. I assume everyone in this space is doing the same, and likewise, Simon. Like I heard you guys, and you also keep releasing stuff all the time. I also find it hard to keep up as well.

Sy Taylor  34:07  
Yeah, no, Tempo just dropped zones today, and there's a million things coming from BAM and Mesh, and that's just a couple of organizations in this space. Google dropped their Gemini four a day ago, as we record this, Muse Wonder was two, three days ago. It's it's extremely hard, and I think that's why a lot of enterprises are at that sort of block and engaging kind of phase. Is they they just don't know where it wants to go. But Bam, I actually want to ask you: How does Mesh think about some of that recourse when agents pay in stablecoins, because we talked a little bit there about what we're doing with the legal context protocol, do you think that smart contracts and stablecoins have fertile ground here, and or do we need to do just good old fashioned standards work and get adoption for something?

Bam Azizi  34:57  
Yeah, I think they have. They truly. I think smart contracts and public ledgers can help a lot when it comes to KYA and also agent identity and also the the fraud layers. I think it has a tremendous amount of value that can can solve. It's not as sexy as payment, so that's why people don't talk about it that often, but definitely that's something that can be trusted and is very versatile in terms of the technology that we can use to build things around KYA and fraud prevention. The other thing that it's a tailwind for us at Mesh is like the whole added value that Mesh has is like the more fragmented the market is, the more value we can add because we're abstracting all that complexity, and agentic like blockchain itself is very very fragmented. That's why mesh exists. So agentic is going to add more fragmentations with all the protocols and all the layers that you're building on top. Like as you mentioned, like a small small merchant doesn't have the capability that Shopify or Stripe has to build, or Amazon has to build all the layers around it, but they want to get the benefit of it, right? So something like Mesh or Crossman can abstract all of the complexity, so users and customers and businesses can get the benefit and all the upside, but none of the downside, which is the technology and compliance risk, so I think mesh can add tremendous amount of value on top of the blockchain and offer a simple solution to merchants. But we definitely can't do it alone, so that's why like we are partnering with companies like Crossman to make it happen.

Sy Taylor  36:37  
Yeah, I think there's so much to build here and so much complexity. People are going to need a little bit of help. Dare I ask this, Alfonso? Maybe I'll frame it slightly differently. Actually, what needs to happen the next 12 months to normalize agents purchasing things for businesses and consumers? Like, what do you think is going to be the next thing we need to happen? So we've had like the instinct muse moment, but what happens now?

Alfonso Gómez  37:02  
I think to some extent, probably commerce of Muse. My guess is it is mostly a prosumer thing. Instinct, for sure, is no. It's like VC's founders and you know maybe like their friends and family. My guess is people buying on Muse, maybe outside of those buying in Facebook Marketplace, are probably the same. The reason why why people try new payment methods and things like that, which is like promotions, like free shit. I think to me the biggest unblock at this time is not technical, is not KYA, is not this or that. It's like people starting to throw promotions at this, for which obviously the economics need to be you know aligned. Like for someone to throw promotion, they need to be making money on it. The moment you start doing it, like I've seen, you know, my mother try the most obscure payment methods you can think of, much more complicated than stablecoins, like maybe obscure bank account somewhere, something that you have to go to a store and buy something to be able to pay just because she got like $2 off on something. I think to me that's the biggest unblock on bringing this mainstream is people saying, "Hey, you know, I it's not like my agent bought this thing. It's actually I got this something for free, and then they get hooked, and then it suddenly becomes the more normal thing." No, so that would be my maybe like hot take. Is that we're talking about technology here, but what people want is like something much more visceral than their agent buying something. It's like saving money or something like that.

Sy Taylor  38:19  
Yeah, those use cases are sort of still around the edges a little bit, but it's where the value is. You know, they always say what the nerds are doing at the weekend, everybody will be doing by default in 10 years. So it's all about time horizons here, and who knows in the age of AI what happens next. Alfonso, I'm sure we could keep peppering you with questions and keep this conversation going forever, but I really do think there's a lot of people trying to dip their toe into this, a little bit lost. If they do want to kind of get their hands on what you're building with Crossmint and your APIs, how do people get a hold of you? How do they get started?

Alfonso Gómez  38:54  
Well, so everything I talked about is public, and I think pretty much self serve. So they can just Google it or ask ChatGPT or Claude or the agent of choice to find it. But otherwise, go to crosswind.com, check in documentations, the agents section, or follow me on Twitter. I'm Alfon G J, Alfon as in Alfonso, G as in Gomez, J as Jordana. Shoot me a DM and happy to connect you as well. All

Sy Taylor  39:19  
right, bam. How about you and all things mesh?

Bam Azizi  39:22  
Yeah, you can follow us on LinkedIn and X. My handle is bam as easy mesh, so easy to find across

Sy Taylor  39:28  
platforms. You can find me at S Y Taylor on all of the socials and at brainfood.xyz, our sparkly new domain. And if you haven't already, you'll find a lot more of this show if you subscribe on Apple, Spotify, YouTube, or wherever you get your podcasts. And finally, if you enjoyed this, please leave us a like or a review. It means a lot, and it helps others find the show too. Thank you, and we'll catch you next time.