Transcript
Hi Mike Matchett with Small World Big Data and I'm excited to be here today talking about AI. Of course, AI is a thing, but also event driven architecture because we talk a lot about how we can do development in lots of different ways, and we talk a lot about AI and AI agent networks and stuff. But how would you really support that with your architecture underneath the hood? We've got some hints and tricks today from AxonIQ.IO, and we've got Jessica here today. Just to hold on a minute and we'll get started. So. Awesome. Thank you so much, Mike, for having me on, I appreciate it. Jessica Reaves, CEO of AxonIQ and based in Austin and excited to be here. All right, Jessica, uh, let's just unpack a little bit when we say event driven architecture and explain like level set with that. Yeah. Uh, there are different architectures out there that people have, and we're talking about, like, how you stitch together various parts of a larger application, right? Do they make remote procedure calls? Do they make rest calls? Do they do they use microservices, Soap, whatever else? But there's this thing called event driven architecture. Maybe just level set a little bit what we're talking about with just that. Yeah. No. Great. Um, so event driven architecture, I think the easiest way to explain it is what it's not. So I think in in traditional Crud based systems, it's very much focused on the current state of what is so right now. And it forgets everything else in event driven architecture or event sourced architecture. It basically has that persistent memory of we're definitely interested in the state of right now, but also IT architects in a way that it tracks and almost audits and ledgers, all of the events that your system and the application has been doing when they've done it, how they've done it. And from there you can derive a lot of insights from that, especially when you think about agents now making decisions without, you know, without oversight. Um, you definitely need that context. And that is what event driven architecture really provides. So, so, Jessica, one of the things I have is, is with AxonIQ, it's not necessarily something brand new, right? You guys have been around doing this for a half dozen years or more. Uh, so would you look would you look at what you say is kind of hardened? Is this is this really proven, or are you still feel like you're on the bleeding edge of open source here? I'd say both. I think that we're always bleeding edge, just the innovation of open source. But our time has come. We've been working on this for almost a decade, and it's truly battle tested. We have, um, we're represented in used in over 80% of fortune 100 companies, over 60% of fortune 500 companies. And it's mission critical type applications. Shuns, right? It is the heart of a company application, so definitely not new. We're not the new kid on the block. I think that we need to get the the gospel out there of events we're seeing of event driven, so people really understand the benefit and isn't as scared about making that change. Awesome, awesome. All right, so before we get to really the goodness of putting this under agents, I just want to just clarify, I spent a lifetime working in system management solutions for it. And by by doing this event driven architecture and basically instrumenting what's going on, we can do a better job of security. We can do capacity planning, we can look for performance bottlenecks, we can do compliance. We can do we can do lots of it enables a lot of the ilities that we talk about, whereas other architectures don't that. So it seems like a natural fit already for somebody looking for an architectural foundation. Um, and we just don't have time to get into all that today. But I'd love to have you come back and we talk about just even any one of those topics? Yeah, no. For sure. So let's move. Let's go on to the exciting part for today, which is okay, AI, AI and AI agent and AI agent networks. I mean, when when when an AI Jen AI does something today, it's kind of like a one shot thing. It goes off and it builds this big package and returns it to you. So what what do we get when we put an event driven architecture underneath our AI agent applications? Yeah. So I, I would say three things. So one is that business context and lineage, almost as I said before, really understanding the micro events that occur to current state and that then that leads to auditability, replayability, etc.. Um, then I would say consistency and control as well as just performance. So it we have standardized ways when you think about events, queries and events throughout the architecture, which then prevails through the application itself, that then gives you performance gains, right? So the read write confusion of application is kind of moved to the side. And you can you can absolutely speed up. And we have a very large Swedish furniture company that I'm not allowed to say they use AxonIQ software and not AxonIQ software on the same use case, and they were 2400 times faster in terms of performance, which is pretty mind blowing. And then the last thing to your point about the world of agents now is really just that orchestration, right? Of really understanding, you know, um, the sequencing of them. What data and events, um, do they need access, do they need. And that standardization of all of that? I always say we are the plumbing, the unsexy kind of gnarly underbelly of of development. But it's so important because AI does not happen without a really solid foundation. And that starts with your architecture. And that starts with event source architecture. Yeah. And if and if I can, you know, just sort of think ahead a little bit when we're working with an agent, an AI and, you know, writing our prompts and stuff, at one point we're asking it for code or we're asking it to do a set of steps. But at this point, we can start to ask it to, to execute a series of essentially events or to execute a series of things which become events. And if I start thinking that way, even when I'm interacting with my AI prompts, when building code, I'm going to build better, better architecture anyway. Whether I even have an event driven system under there. If I'm just event thinking I'm going to, I'm going to end up with in a better place, because now I've got different componentry and people tend to think of like, you know, just analogously, getting to microservices and designing everything as a microservice, and then maybe they get some of that goodness. Right. But if you just stick with microservices and that's kind of your end all be architecture. You still don't get the visibility into what's going on or the replayability necessarily, or anything. You just you get a huge 10,000 set of microservices, right? Which gives you some deployable control as an IT person, but doesn't really give you some of those other things that we can talk about. Um, just, just, just briefly and again, for the sake of people who aren't as familiar with events, tell us, uh, what in, you know, maybe just give us an example of an event store. And one of the things you can do if you can replay events or go back and and analyze events after the fact. Yeah, yeah. So, um, a good example. And a lot of our use cases are either in financial services, which are highly, you know, transactions very heavy. Um, throughput. Yeah. Real time really matters, right, in terms of how you're pricing things. Um, also highly governed type use cases and commerce, when you think about multiple stakeholders of what is the price, is it shipped out, is it late, etc. the billing aspect of it, all of it. Um, so a good kind of event use case there. I would say that if you take a, let's say a grocer, a commerce type use case where what is coming in from a supplier perspective. Right. Um, what is the not only how much supply do they have, but when are the actual no kidding shipments coming in then goes to distribution of like where are they being distributed to. Then it goes to like implementation and just orchestration of those warehouses. Then it goes to thinking about the demand and what you're seeing historically. And those can be events too, right, of how much is being bought of said things. Are strawberries being bought a lot right now in this in this quarter. Right. And that then affects the actual Pricing of things, so the supply and demand aspect of it. That then also will help when you think about the tagging from a consumer level, if you go into a grocery store and sometimes you see, you know, last sale or brand new, that then interacts with those types of tagging and labeling systems as well. So you can imagine that an event store in this use case is super powerful because you can see from season to season, because a lot of times groceries are seasonal. You think about the holidays or just seasonality of produce. Um, the replayability of what happened last year or the year before that versus the current state. If you were doing it just from more of a credit based system, current state is strawberries are doing this and performance versus truly understanding the whole story of the peaks and valleys of what's going on. And Um, and not just from the sales perspective through all of the different stakeholders that really matter when you think about, you know, providing strawberries to an end client, to an end person, go into the grocery store. So there's so many vectors there that you really need to. I think the beauty of an event store there is you can cut it into different kind of angles to then analyze what's going on in your business and make make it better. Right. And of course, and of course, the question in my mind is like, gosh, that sounds like that's a big job and a challenge and how would you do that? And of course, the answer is AI, right? This is just is just where we're going when these things come together is if you can empower what we're doing with AI, whether it's, you know, generative AI or generative AI, plus some other things, and the automation that's going around a lot of generative AI with, uh, what you see, what I, what I would call a time series, uh, data history of what's happening. You can do a lot better predictions. You can do a lot better reasoning with what's going on. People talk about Rag today by giving it the right series of documents or or calling out to the APIs. But here we're saying, look, we can actually give you the events that have happened in the past in your application, and you can apply artificial intelligence to that in the future to actually make really superior decisions and operations. Right. Hyper efficiency and all the rest of it. I always use the analogy of like, do you want to make life decisions based off of your state of your life today? Or would you want the whole history of growing up and like all of the events of your childhood that has made you the person you are today? That principle is true inside application development as well, especially when you think about agents making decisions for you. You want them to understand the whole narrative, not just today's narrative. Uh, actually, I'm not sure I want to remember my childhood in that detail. Fair enough. But but you know, I know that that's the way brains are made, right? Somewhere down the core of my being are the impressions in early memories, and we build upon that. And that's how human intelligence is formed. So yes, I can fully on board with that, that we, we don't want to forget, um, the big things and the important things, and we want to build on those. So very cool. Uh, so, Jessica, let me ask you this, uh, and there's so much to dive into here, but, uh, it's really exciting that that you're, you're starting to see the real value of event driven architecture, not just for what it is as a, as a, as a development style, but for empowering kind of a new future of intelligence, which, which just means that our AI can be super intelligent. Uh, and when when it's calculating what we should do next or what should happen to optimize things. So I'm really interested that if someone wants to dive a little bit more into this, they want to learn a little bit more about either, you know, the core, uh, exotic, uh, event driven architecture or how this is applying to AI and how you see this is going, uh, even get, uh, just as a hint. There's some cool things coming, guys, uh, get on the list to see what's coming next. Where would you point him out? Yeah, definitely. Go to the website, which is AxonIQ and it's AxonIQ. Many people think it's AxonIQ, but it's AxonIQ.io and you'll see a lot of cool things coming out. Um, one of which I'll sneak peek. I think some of the event sourcing, event driven architecture people are scared about, it's a pretty advanced topic. It has a steep learning curve. We're coming out with some capabilities that we handle. A lot of that lift for you, while also teaching you and showing you what's going on behind the scenes. So that way it helps speed people up. They don't have to, um, learn all about event driven architecture to get the benefit of it, but at the same time learn more about it so they can be smarter and more educated as as they do their next thing. Right. I mean, this is something that you might want to be doing regardless of AI, but with AI in the motion and definitely not going away anytime soon, right? This is getting more and more. This may be something that you absolutely want to start to prioritize and get to the top of. Uh, the developer food chain sooner and start to start to explore it, right? I'd say every day pass. You're behind a bit when you think about how fast the AI revolution is going. And I really do think that back end development is changed. It is changing as we know it, just like how we moved away from monoliths, right? I think event driven event sourcing is the new future, especially when you think about the dawn of AI. And like you said, that's not going away. Thank you so much for being here today, Jessica. I look forward to seeing what you guys are coming up with, because I know there's some really great things in the pipeline. Uh, and hope you come back soon. Awesome. Thank you so much. All right. Take care.
