SERV Reasoning with IXS

our track partner, I excess finance and dig into one of the areas that we're really most excited about at OvenServe which is a genetic finance. AI agents are really moving from just assisting us with financial decisions to managing and deploying capital at scale. So there's really a lot for us to unpack here around what that looks like across various asset classes like RWA's tokenized assets, financial markets more broadly. So as mentioned we have joining us today, Tim Hathner, founder and CEO of OvenServe, who I'm sure most of you guys already know and Julian Kwan, CEO of IXS Finance. So yeah, we're just going to keep things pretty open, talk about where we things are headed, what both teams are building and how devs and builders can really start experimenting with our tech. So yeah, let's kick things off with a quick control, Tim, maybe you can just kick us off and do the honors for anyone discovering OvenServe first time maybe just joined our community. Can you just give us an overview of what we're building with serve and the bigger vision? Absolutely, thanks, Adam. Cool. So yeah, so at OvenServe we're building what we call the reasoning layer for agents and we tackle kind of three dimensions within that that's around kind of the main issues that plague agents today, AI agents and kind of hamper their adoption across different spaces and verticals and those are sort of reliability and performance, cost and auditability or observability kind of interchangeable there. And we essentially embed or add on top of OVA agents and their models, like basically the AI models they use, we implement like custom reasoning graphs and architectures and agent harnesses and context engineering features to basically uplift the intelligence of the AI models that the agents are using, which as a consequence makes them more reliable and more performant. And as a consequence of that allows institutions and builders and companies to use smaller models for their agents that cut down on costs, cut down on latency and improve the performance and reliability of their systems. And as a final piece we introduced this auditability aspect, which is particularly crucial for sort of like critical use cases and larger scale use cases where agentech finance being one of them in particular where the decisions that agents are making need to be tracked and traced and need to be able to point to something. So different stakeholders can understand why results were produced, why certain outputs were put forth compared to other decisions that the agent could have made, which allows for greater visibility, transparency and accountability ultimately as well. So that if something goes wrong, you know where to look who to ask and what to do. And yeah, the goal of what we're doing is to essentially yeah become the kind of like default intelligence and performance layer for AI agents across the space and unlock the adoption curve of AI agents in real settings and enterprise context that currently they're not able to do because they don't basically possess the properties that we're building for. So yeah, that's us in Anacha, we'll pass it on to Julian. Hey guys, can you, I guess you can hear me. Yeah, hey, Ron. Thanks for joining in, taking time out of your day. I'm Julian Kwan, I'm co-founder of IXS Finance. IXS started as the first real-world asset Dex about four or five years ago, that was the original business model. The back in that day, we hadn't actually, as an industry, seen any of the financial instruments and the fixed income and institutional assets at that point in time. So the I think we lost Julian. Yeah, I think we lost you. Oh well, yeah, let us ping Julian. Oh, you there? Oh, you're back Julian. So you lose me just that a bit of a glitch. Okay, sorry, yeah, there's something wrong with XC, but yeah, so what happened was in the last year or two we started to see kind of the institutionalization of the RWA space. What that really meant was much higher quality assets, publicly traded assets, you know, money market funds and treasuries, a lot of fixed income yield. They're a really good product market for stablecoins that earn no yield. And that's where our platform kind of evolved to. And then what that has meant is that, and then the beginning of the year, we were looking, we implemented a lot of AI into the business starting January last year, 18 months ago, 20 months ago. But we really sat down and thought about what's this whole agent growth. It's really getting exciting. I'm excited to sort of really coming in alive in Q1 this year, I would say, or a lot of it did. We started seeing a lot of stuff going on blockchain related. We met the serve guys thought they were building something really valuable. We saw a lot of people launching agents kind of investment agent type kind of projects. A lot of them were, you know, sort of like an ICO utility type token that was tied to an actual agent that was doing trading and then parts of the, I guess the profits would be shared with the token holders. There was a bunch of people experimenting in different models. So we originally then thought, you know, hang on, we need to get, we need to look at what we could do, what we wanted to do for the industry. So we have, you know, the regulated infrastructure, and we started building real estate vaults and then initially those vaults were for human investors. Then we thought, well, why don't we provide real estate vaults for the agent economy. So agents could interact with these vaults and there'd be sort of different parameters and metrics to that. And that's what I access is focused on in relation to the agent world and the agent economy. It's building like additional products on top of what we need to have any way to do this. And so that was sort of how we got into kind of the space we didn't ourselves want to build another agent that there's lots of different products and projects out there doing that. We want to provide, we see, we see agents who are putting idle capital into defy vaults and defy yields and earn programs. But when it comes to RWA, you know, it requires licenses and security. So we sort of a big opportunity. to step into that first and that's what we did. So that's how that's how we got here and, you know, obviously, came to do a lot more with with servant, you know, the, the, what we realized very early on like anything new, there's a lot of experimentation, there's a lot of people doing all kinds of different things. I just finished a long tour of America, meeting a whole bunch of different groups, people building bank, banking services for agent, but building, building trading services for agents, like all the things that, you know, human investors or human would need, but with the agent spin on it. And at the end of the day, every single agent at the moment is predominantly kind of owned by a human. There might be a human that uses agents to spin up a whole bunch of sub agents, but at the end of the day, they're not just topping out, they're not coming from nowhere. And especially on the investment space, they're looking to make money for somebody. So it is this interesting layer. I don't think it's as crazy and galactic as a lot of people think at least in the investment side of the world. And that's how focused. Thank you, Julian. And thank you, Tim, for shedding some light on what we're building with open-serve and I access respectively. So to take a little step back and let's say maybe you're new to our communities, you haven't really researched a lot about agentech finance and what the opportunity is at hand. What does that actually mean beyond, you know, let's say crypto use cases or payments, trading about simple automation, gardener predicts, divide 23, at least 15% of day-to-day transactions will be made autonomously now. Bit of a bold claim there. We know these predictions go, but that's definitely the trajectory of the industry is going on. So I guess what changes when agents go from like assistant humans to actually making these financial decisions and managing capital. I know Julian, you've mentioned compliance there and making sure it's regulated. So I'd love to take both your takes, Tim and Julian on this. Yeah, sure. But yeah, I can add to you. Sounds good, thanks. Cool. So yeah, I definitely agree with what Julian just said at the end there that, you know, I think we like to sort of like mystify what agents really are today and kind of imagine them as into a galactic autonomous being that traverses our world and makes decisions by themselves which at the moment is not the case, right? And we're kind of on like on a curve, right? On like a sort of like quote unquote, like autonomous curve. And right now agents are not particularly autonomous. They're still relying on a ton of guidance from humans and there's a lot of guardrails baked in because it's such a nascent technology. And we know at the moment with how sort of like reliable and trustworthy these systems are that those guardrails and human sort of like human contribution is still super, super important. But you know, when we look into the future that is likely where we're going to get more and more as we can trust the underlying infrastructure and technology more and more that these agents are going to become more autonomous and require less human involvement and make their own decisions in an increasing manner. And you know, the reality is we don't know where where that ends up, right? Like to what degree will agents actually be performing tasks on their own under their own sort of directive under their own creativity. And we're going to have to see where that kind of lands in the future because I'm sure one can imagine if we get to that state and then agents start acting maliciously and you know, holding sums of funds for ransom or what can cut those what that we're definitely going to be throttling back the if possible, right? Throttling trying to throttle back the autonomy that these agents have. So what that means in terms of sort of like actual applications in this space is yeah, hard to say what it really looks like looking further out. But right now we're super early on in this curve. And I think right now, yeah, we're looking at the early days of trading bots and simple payments and sort of like commercial interactions and transactions. And it's just going to expand into a larger, into a larger slice of what finance is, right? Like finance is a massive massive industry spanning different geographies and different types of finance. You know, we're talking about Web 3 finance, which is an entirely different world at the moment until they too until, you know, both of them merge. But is is largely a different world than what exists in, you know, regulated banking in somewhere, right? So we're talking yeah, payments, commerce, you know, even back office stuff for an exchange transactions. Like imagine some large entity or sovereign world fund is has some agent that's you know, autonomously gathering price data and aggregating news feeds and, you know, executing 4x transactions to swap it in out of different currencies and maintain certain like risk profiles against there, whatever their directive is. And it's going to expand into a whole number of things all the way down to like customer support at a bank, right? Like you're not going to talk to a human anymore. You're going to be talking to an agent to process your banking loan or check your credit score or whatever that may be. And as the sort of capabilities of these agents continue to improve, that will only expand into more domains based on that kind of like trust coefficient, I suppose you could frame it as. And yeah, we'll yeah, go ahead Julian. I think yeah, I think we love. Yeah, I think it needs to be a speaker. Yeah, just add on to invite to speak. I guess Tim in the meanwhile, as we get Julian back on, you know, you briefly touched on serve reasoning, the availability, the availability.