All right, so I'm gonna go ahead and quiet that beautiful hold music, and we're going to get everyone's face on screen here. All right. Oop, and we got a squeaky chair. That's me. Hey, John. Well, that's a perfect segue. So I'm going to go ahead and hand this over. We have a star lineup here to walk through the Zeus release. So, John, Carl, Jonathan, I'm gonna go ahead and hand it over to you guys. Quickly tell us what you do here at Bridgeline, what we're gonna be talking about today, and then we will jump right in. Excellent. John, why don't you start off there? Okay, terrific. Hi, everyone, Carl and Jonathan, good to talk to you again. I've always enjoyed these conversations, so, looking forward to this one, too. Thank you everyone for your time today. I'm John Murcott. I run products and strategy, and looking forward to telling you about Zeus today. I'll hand it off to Carl. I see you in that order, Carl, as I- There you go. ... go down the left-hand side here. Thanks, John. Yeah, so I'm Carl Prizzi. I head up the growth team here at HawkSearch, and we're super excited to be announcing our latest Zeus release here with its advanced AI-powered capabilities. And I tell you, in my position, I get to attend a lot of trade shows for all different industries, and I think regardless of the industry, the hot topic lately has always been AI. How can you leverage AI to help your growth and your business, and streamline operations? So many of you attending today may have great interest in leveraging AI, but you just don't know where to start, so we'll simplify that for you. We'll show you how easy it is to get started with enhancing the product and content discovery on your website using AI. So, I'm really looking forward to the time here today and always great joining you guys, so appreciate it. Yeah. I'll, I guess I'll bring up the rear here last, but I certainly hope not least. My name is Jonathan Meyer. I'm one of the senior solutions engineers for the Bridgeline team. Been working with the HawkSearch tool for close to 10 years now this year, and so I think it's kind of a fitting cap for that 10-year anniversary that we'd be rolling out this great new features with Zeus, really diving deeper into all of that AI functionality, and very exciting future to look forward to here. So, Carl, I'll pass it back to you. Why don't you give us a little background here on HawkSearch in general? Sure, yeah. Before we really jump into the details of this latest release, I figured I'd spend some time just giving everyone a quick background on HawkSearch. For those that don't know, we're all about helping our customers grow and be more successful online. We have tools to help dve more traffic to your site, increase order value, drive conversion. We have thousands of implementations for powering the product and content discovery, as well as merchandising capabilities on your website. So, we have a growing ecosystem of technology platforms and partners, agencies, SIs, that really can implement HawkSearch on any web application. If we jump to the next slide, just a deeper dive into some of our detailed features here. It's a bit of an eye chart, so apologies for that, but we offer really a comprehensive solution for powering B2B and B2C experiences with advanced machine learning capabilities that understands user behavior, what's trending from a search standpoint, surfaces the most popular categories and items, really even before the user starts clicking on the search field. Or, you know, we can have recommendation widgets, or even when they click on the search field, presents, you know, a lot of those top trending items. We can dynamically adjust the ranking of results based on trending data and the actions that are taken throughout that customer journey. We handle complex product catalogs with tens of thousands to millions of products for our customers. We can handle use cases like partial part number searching, customer groups or entitlements. You know, these are very common and really essential use cases in any sort of B2B scenario, to make sure we're showing the correct products and pricing to the specific audiences. We can handle a hierarchy of sites and distribute the search indexing and merchandising rules down from a parent site down to child sites, so really powering a complete enterprise. Really helpful for franchise organizations or distributor networks that may have location-based catalogs. We can also handle federated search models, where you have content, products, data in different silos. You want to have a unified search experience for that end user. You know, we can really make that easy, really streamline that user experience. When it comes to data services, we have a pretty unique offering there for normalizing data that may, again, come from different third-party sources, different silos, different suppliers, really helping to normalize that. Pretty unique feature for handling units of measure. You can think of how somebody can search on a size or dimension, all the different ways they can enter that in. So we help simplify that experience, and we have built-in libraries to really help normalize all that. Ultimately, we also have personalized product recommendations based on search history, interests, you know, really that personalized experience there, past order history. We can ingest all that data and really provide for that personalized experience, helping to drive increased AOVs, cart sizes, increased order value, many merchandising features built in here. So again, it's a big laundry list of features, but anyone interested to learn more outside of what we're showing you here today, feel free to reach out. We can set up a quick demo and show you all the great capabilities of the platform. I think you would say, "But wait, there's more. Yeah, is that it for me now? I think we're done. We're done here. Yeah. So, yeah, again, as you can see, there's a ton of features here, and, and, we've been leveraging AI for years. It's funny, it's just the latest trend when I go to trade shows. Like, "AI," we've been doing it for years, but we have... We're continuing to innovate, even more AI capabilities. So, so today we're excited to release even more innovative AI-driven features to help users find that right product or the right content of interest. Introducing Image Search, Concept Search. I don't want to steal their thunder, so, I have John and Jonathan here. You know, it's really gonna be the John and Jonathan show. I'll chime in every so often but, I'll pass the torch over to the Johns here to, to continue forward. Yeah, yeah. So we're, we're gonna be diving deeper into kind of what we mean by, you know, AI-powered Concept Search and Image Search. There's a lot of, you know, and to, to Carl's point there, it's- it is kind of the big buzz thing that's going around right now, and so we want to make sure that we're kind of delivering these advanced features for our clients and, and, and really presenting new forward-facing features. But also, we want to make sure that, you know, the proof is in the pudding. We're, we're not talking about, you know, kind of smoke and mirrors here. So we're gonna be showing you practical demonstrations of this working in a, in an environment for you, so you can see kind of... You know, when we're talking about Concept Search, not only are we gonna talk about that, we're gonna show you what that actually means. Same thing for the Image Search. We're very excited about this functionality here. So let's get into this a little bit. Let's give a little bit of context first here. You know, you saw the webinar. It's called, you know, Meet Zeus. Well, what are we talking about with that? This is kind of our code name for this upcoming release. The next release is going to be called Athena. I think the one after that, we're looking at calling Hermes. We're kind of going for this classical antiquity theming because this really is a fundamental kind of, you know, shift in how we're looking at product discovery working within HawkSearch. And so we're kind of going for that big, epic, grandiose kind of naming conventions here. And, John, I think you found out in kind of researching names to be calling these, that both Zeus and Athena have birds associated with them. I know. So What, what could be more perfect? Right. I know! If it was only a hawk and not an owl, but hey- Mm-hmm ... that's, it's close enough, as far as- They have wings. We'll take it. Yeah, thankfully. We'll take it. Yeah, there you go. The Concept Search- There you go. ... would understand that they are related. Yeah. Excellent. Well, let's go ahead and talk about what we mean by kind of this approach. You know, if we're talking about changing how HawkSearch really works from a product discovery perspective, what does that mean? We're really looking at this kind of as a concept of smart search, and we're referring to this as the smart search trifecta. The Keyword Search that you kind of are used to within a classic kind of product discovery site search engine, where someone comes to the site, they go to that search box, they put in what they're looking for. You know, it's really looking at kind of, well, matches against the product descriptions, and the names of the products, and, you know, attributes around that. That's kind of your classic Keyword Search, and HawkSearch is gonna continue to support that going forward. We actually have a lot of existing AI functionality baked into that, like Carl was talking about, for tracking, you know, what are your visitors actually purchasing, and what are they interacting with, and using those to help with the kind of relevancy improvements, by using that AI. So we're not gonna be abandoning that Keyword Search functionality. There's a lot of great use cases there. And again, you know, that's, I think, kind of the expected functionality, but we want to be looking towards the future. And so we're branching out and leveraging more of these kind of AI for generative AI, for some of the internal tooling. Looking at AI to drive kind of visual search, understanding what images can be uploaded, and of course, leveraging those kind of large language models that you've been hearing in the news, when talking about, like, ChatGPT and all of that. Leveraging those large language models to understand the concepts and deliver more natural language processing. So really kind of looking at this, the product discovery aspect of being a search trifecta driven by your classic Keyword Search, improved through AI, Concept Search, understanding the intent of what your visitors are actually looking for, and a visual search here. Very exciting stuff. Let's give you a little bit more background in terms of what this is going to entail, though, what our Zeus release is really looking at here, 'cause it is such a fundamental shift. We're expanding out kind of our classic HawkSearch Workbench, including areas that now have a dedicated Hawk AI feature. I think we've internally joked about this 'cause we kind of just started calling it Hawkeye, Hawk AI, and then we realized that there's a generational divide there when you say Hawkeye. John, I think me and you, being older gentlemen, unfortunately, refer to this as Hawkeye from M*A*S*H. And then everyone, like, our younger development team was like, "Oh, no, you mean, like, the Marvel character, right?" So, just kind of a little bit of fun there. But really, we're looking at kind of improving that functionality within the HawkSearch Workbench, creating these dedicated areas so that our clients that are using HawkSearch can have quick access to these exciting features here. So updating the dashboard, exposing the configuration for this, while keeping that same kind of ease of use that HawkSearch is known for, and really we want to make sure that we're explaining what these features are here, as you can see in the screenshot that we've applied. So what do we mean by Concept Search? There's been this kind of preamble that we've been going through here, kind of setting the table. Let's get into the meat and potatoes here. What is Concept Search? And John, maybe I'll pass to you. I've been talking for a few minutes here. Why don't you run us through... kind of what you believe and what Concept Search is within HawkSearch? Yeah, and there's a lot of terms that are also used, natural language understanding. I think it was NLP before, a new kind of like NLU, natural language understanding. Mm. And it really is looking beyond the keywords, as you were just saying before, and converting that request, and you can kind of see this in action right here, into a vector. So this is not gonna go into all the technical details here, but finding a match within that large language library that you were talking about before, and doing that kind of upfront search, and it's just a different mindset. Like, right before the call, we were talking about, you know, can this AI figure out, like, missing metadata that I have? And you have examples, Jonathan, where- Yeah ... rather than kind of work backwards from your age and the size of the jacket you may want, or whatever- Yeah ... those details are, hoping, of course, that your data model supports that. If you could approach it more like, "I have a question, I have a problem, or I have a plan in mind, and do you have products that match that?" So I don't want to give away everything you're gonna-... do on the demo, but the idea of saying, like, you know, "I have a six-year-old, and we're going on a trip in the mountains," like, the system would understand the way that that question is converted. That, okay, I understand the size of the jacket. I understand it's gonna be cold. You know, that type of thing. So it's really approaching a query, not working backwards from the way you think that search works. Like, that's how we all think now. Yeah. We go to Google, it's like, well, I better ask the question the way Google knows, you know, how to respond, although it's getting a little smarter these days. But rather it's, "No, why don't I just say what my issue is, and then see what the responses are?" And what's so great about it is, number one, it does return great results. Mm. But number two, it returns results you didn't even think of. Yeah. I know you'll get through this when you do the demo, but we're both very familiar with the library of products- Yeah ... that we've been demoing forever, and it wasn't until we started to expose this Concept Search that I was like, "Wow!"... I didn't realize this data model had food in it. Yeah. Because it never understood cooking, and now I'm giving away- Yeah ... everything you're gonna cover. No. But the point is, it's like you discover things you didn't know, and we kind of- Yeah ... talked about product discovery, and now it's, like, really taking that next level to help discover. Sorry to interrupt, but- No, please ... Carl, I know one of our bigger customers mentioned, yeah, it's sort of like a conversational search, right? Like, I'm asking- Sure ... just phrasing long-tail phrases that I'm putting out there, and it's responding to that, and it's intelligent to know, like, what are you asking, and interpreting that, and then returning the right results, right? So it's just, yeah, the use cases in my mind are just endless as far as the capabilities here. Yeah. I think what's kind of best about this is that this isn't putting any kind of additional pressure on our clients either. If you're with HawkSearch, this Concept Search is using the existing data that we're already using to drive the Keyword Search functionality. So this doesn't put a bunch of pressure on your team, on your development team, to come back in and say, "Oh, well, now we need to spin up all this additional data that then HawkSearch is gonna have to ingest," and that's not the case. What we're doing is we're taking that existing product catalog that we're already using to drive that Keyword Search functionality and running it through these different AI technologies to understand this additional to drive this additional functionality. It's super powerful and super easy, which I think is kind of it almost sounds too good to be true, where it's, oh, we don't have to spin up a whole new data set, a whole new thing. No, we're just using the existing data, running it through this technology to drive the functionality. It's great. And one of the kind of extensions of that is that this - you'll note the last bullet point here, is that this supports 50 languages. So if we move on to the next slide here, it's really incredible to me that one of the things we've consistently run into in the Keyword Search space is this idea of multiple languages and how do visitors search. You know, maybe, maybe English isn't my first language. Maybe I speak Spanish primarily, but the catalog of content that I'm searching against isn't in Spanish, and that's always been a gap. You always either had to rely on, like, Google to do, like, a page translation after the fact, or you had to go out and hire a translation service to go ahead and, you know, create all that new data and translate all of that, your product catalog, so that someone could search in Spanish. And then you need to have a dropdown to toggle between those different languages, and that's not the case with the Concept Search here. And, John, I'll leave it to you to speak a little bit more of the technical details there. But what's really happening is that this allows a visitor to come to the site, search in what their primary language is. So if I search in Spanish, and we're capturing that at the search request and transforming that to say, "Oh, well, they... This is in Spanish, we're gonna transform that basically back into English," and that then allows us to search against the same product catalog and deliver consistently similar results, regardless of how that visitor submits that search term, with up to 50 different languages. It's super powerful, allows for that, again, that expansion of product discovery, where you're not tied to, "Oh, my catalog is in English, so I have to just kind of force that to be the way someone searches on this." We can now support up to 50 different languages out of the box. Someone puts that search term in their language, and even if your product catalog is still within it, it still contains primarily English content, we can pull back in those same results. Yeah, maybe just to add, Jonathan, the technical aspect to that, 'cause we are kind of thinking of these products here as nodes inside the system, and it's a conceptual way to find what's nearest. These are terms that are common- Mm ... in the world we're living in now. And there's a sentence transformer, that's the technology, that sits in front of the LLM itself. Yeah. That's how, whether it's Hindi, whether it's Chinese, whether it's Spanish, you were saying before. We have a lot of fun because, of course, we have team members around the world. Yeah. We have a salesperson in Israel, so it's like: "Hey, just try it in Hebrew, see how it works." It's pretty scary how well it comes back with the results. So you get the best of both worlds. First of all, the Concept Search just overall, and then that you can do it in multiple languages is really powerful. And I'll just highlight one more time, Jonathan, 'cause I remember- Yeah ... when we were first talking about this, of course, the first question back was like: "Do I need another engine? Like, do I have to set up... You know, what do I gotta do here? Yeah. This sounds like I gotta set something up, either on my side or more, buy more technology. It is essentially adding fields inside your existing instance of HawkSearch that store all of that vector information. So that's why it's really powerful for Carl and Carl's team. You know, this is something that can be done economically, and is already inside the system. So no changes on your indexing side, your whole process from that perspective. And then once again, not to get too much into the weeds here, but on a lookup side, like the API, believe it or not, you're hitting the same search endpoint. Yeah. You're essentially passing in a different parameter. Whereas before, of course, it was always assuming you're making a Keyword Search. Now, you say: "What kind of search type is this? Is it an Image Search? Is it a Concept Search?" And it's smart enough to return the same results, the same JSON that comes back from before, it's the same JSON you see now. Yep. And maybe, I think that might be a good transition, 'cause it's not just this Concept Search that we're talking about there. You just mentioned the Image Search, and I think that's really one of the most impressive features of this functionality. The large language model understanding the terminology that someone is using is super important and really kind of helps expand that product discovery. But, you know, you can't get better than, "Hey, this is what I'm looking for. Let me take a picture of it, and you find what matches against this." And this is also functionality as a part of this Zeus release. And again, just to kind of build off of that point that you're talking about, this isn't a big lift because we're using the existing product catalog images to analyze those and deliver those results. The way that this kind of works, again, to just take the high level here, is that we're actually leveraging the AI to analyze both the image that is being uploaded from the visitor and the image that is a part of your product catalog. That kind of bridges the gap there between maybe it doesn't look exactly the same, maybe the perspective is different. It really can let the, you know, the AI say, "Oh, well, we think that there is, you know, some gloves in this image," and even then look at the product catalog and say, "Oh, we think these are also gloves," even if the... You know, I'm looking at it upside down, I've got it turned around. It's not exactly the same perspective of what we have in our product catalog. Because the AI's analyzed the image that's been uploaded and the product catalog images, it can bridge that gap. We'll show you this, 'cause I think it's super impressive, as part of that functionality there.... and this does work in mobile. Please, go ahead, Carl. Yeah, so I was gonna mention the mobile experience. So, you know, we're seeing a lot of traction in the industrial supplier vertical, specifically electrical suppliers, plumbing, HVAC. You know, you have technicians in the field that are working on something, a complex mechanism, and they have to replace a part. They could just whip out their mobile phone, take a- Mm-hmm ... quick picture, you know, and upload it all through the UI that we make available, right? And just returns the exact part that they're looking for. So no longer trying to figure out, "What part number is this?" You know, like- Yeah ... take a picture of it, and boom, it, it shows up, right? So it's- Yeah ... super powerful stuff. Again, you know, could be not the best lighting conditions, could be at a different angle. Yeah, and having that ability to take the photo from your mobile device or upload from an existing photo that you might have, I think really kind of expands that use case, like you said, in that B2B environment where we see people out in the field. They've got the part in their hand. They might not know exactly what it's called. Let me just take a photo of it and have the Image Search find that. Really powerful stuff. John, anything you'd want to add to the initial Image Search functionality here? I'll just mention it briefly. I know, Jonathan, you go into more detail in the demo itself, but it's not trying to match the picture. Mm-hmm. It really—you said it already, but just to kind of emphasize, it really is trying to understand what's in the picture because, you know, comments you've made before is like, "Well, maybe the picture is at this angle, but the picture I... Yeah. You know, so it's not just saying, like, "Oh, is there a, an exact match to this picture? Yeah. It's looking for what's close, and it's looking at all of the objects in the picture, just to emphasize that. Very good point there, and I think that's one of the big differentiators of why this technology is really kind of coming to fruition right now, is that in the past, it was always, "I took a photo," and it just tried to find whatever looked the closest to the photo that you uploaded. It really wasn't understanding what was in that image. We've reached the point now where the AI can actually say, "Oh, we think these are... There's a lamp in here. We think there's gloves in here. We think..." You know, whatever you've taken a photo of, it's analyzing the image itself, not just saying, "Oh, well, this looks close to that." It just kind of changes the game in terms of this Image Search functionality. But the cool thing there is that because this is analyzing the image in both the image that's being uploaded from the visitor and the image that lives in your product catalog, it allows us to do this other cool feature that we call Image Finder, which allows you to describe the image that you're looking for. You know, maybe you don't always have the thing in front of you that you're looking for, and you don't really know what it is to do a Keyword Search on it, but you can look at it and say, "Oh, I'm looking for this," or, you know, the example that we have here listed was, you know, you might have been over at a friend's house, and your son comes back, and he's like, "Oh, I was reading this book, and it had a butterfly on the cover." They don't know what the name is. They don't know what the author is. You can't go to the search box and say, "I need a book with a butterfly on the cover." It's not gonna understand that. With an Image Search, though, an Image Finder like this, you can put that kind of a term in. You know, "Looking for a book with a butterfly on the cover," and it will understand the concepts of what you're talking about there and use the images that are in your product catalog and that it's analyzed to say, "Oh, we know that this is a book and that this has a butterfly on the cover. This might be what you're looking for," and returning that. We're gonna show you some really cool examples as a part of that demonstration with this. But, John, I know you love to explain the multimodal model. So I'm gonna let you do that. Yeah, and I also love that example 'cause it's funny. Of course, we have a lot of these conversations, and we talk to customers, et cetera. And on the one hand, you know, especially in the B2B world, it's like, "We don't have a lot of books with butterflies on covers." But on packaging, and this is a way that I often think of things, B2B or otherwise, it's like: What's that packaging, you know, that has the thing on it? Yeah. Most of the time, that's part of the picture that you have in your catalog, and it happened to be one of our large customers. They had butterflies on their packaging, and it's a customer that has just a very large number of different types of products and, you know, trying to remember, like, "What is the SKU, or which one is this?" But it's like, "Oh, it's the one that has the butterfly." It seems like a maybe a use case you wouldn't normally see in B2B, but in that case, they actually happened to have a butterfly on the packaging. So sometimes when I'm doing these searches, I'm actually searching for the packaging 'cause I know what the packaging looks like. Yeah ... when I'm ordering this. On the technical side, I know multimodal model, and we hear about these large language models, et cetera, but most of the time, each model, each kind of, segment of information is oriented towards one thing, like the large language models we were looking at before. For Concept Search, that's a very particular type of model. We happen to use a BERT model. B-E-R-T is the model. But some models have multiple modals. Yeah. It just means, in this case, when you're searching for the butterfly or whatever it may be, it's not looking at metadata. It's not looking at the file name. It's not looking at the description, although it could, but it's not. It's... When it is creating the vector of that picture, it's also including in there the description of it, so it understands, what am I actually looking at here? And we've had fun, Jonathan, looking for- Right ... you know, pictures that have numbers in it, or, you know, there's a way you can start over time to, like, mix and match your metaphors. It's like, well, maybe the picture has the SKU number on it. Yeah. You know, that kind of thing. It really starts to get interesting here. Yeah. That's... We're very much looking forward to showing you guys that, in the demonstration here in just a few minutes. But little bit more that we want to cover here. From the configuration settings, again, we'll touch on this very briefly, but we wanna make sure that you guys understand when thinking about... You know, we've talked about this a couple of different times now, where we're gonna be leveraging the existing data that we're already using to drive Keyword Search. And so this functionality, this extension, adding in concept and Image Search, it is a simple extension of the tool in the HawkSearch world. We're basically then just saying, from a simple on/off toggle, "Tell us what's the information that you want to be used for the Concept Search." Obviously, the Image Search is a little bit more straightforward. There's usually just only one or maybe two fields that we might be looking at for that, but it's a simple config toggle to say, "Hey, this is the information that we're looking to use," and then we take care of the heavy lifting on our end to take that data and run it through those large language models or through those multimodal models and understanding what that actually then comes back with. So ease of use, still paramount when it comes to how HawkSearch is delivering this functionality. ... maybe I'll just add one- Yeah ... super quick, detail to that, is I know I said it already, but for people who are very much in the day-to-day, so like, "Well, we run an indexing process, and we already have the index set up," like, "Do I have to do anything, our nightly index?" or whatever your timeline might be for doing that. Essentially, what we did is we added a pipeline to that process, and when we look at every document that is gonna be indexed, it says, "Oh, is that supposed to also have Concept Search? Oh, it is. Oh, okay, let me go through the fields that were selected in the screen you were just looking at, Jonathan- Mm-hmm ... and let me create a vector for that and put it in another field, so that if someone does a Concept Search, it will have that information available. Yep, and what if, what if the image changes, right? If they re-upload the image and, and we're, we're pushing that to the pipeline as well, right? For any deltas or anything like that. Yes. That's an advantage in general with HawkSearch. You can just do a partial index, maybe even just one product, and yes, when you do that partial index, of course, you may have other changes as well. Yeah. But regardless, if the image changed and maybe the other description or what have you didn't change, it would still regenerate a vector of that image. Nice. Yeah, really, really cool stuff there, and again, straightforward. You know, it doesn't require, you know, heavy lifting on anyone's side here to get this information up and running here. But it's not just kind of from the search end of things. We also wanted to take the advantage of kind of this new technology to bring some additional functionality into the back end, into the workbench sections of HawkSearch. So we've come up with our content generation, which is using generative AI to allow visitors to, or allow our users, I should say, to be able to generate content using a kind of a prompt experience. So coming in and saying, "Hey, I need a paragraph about those gloves that we were talking about," or, "I need a paragraph about that book with the butterfly. I'm out... I've got a brand page and, you know, I just need to create a bunch of information really quickly. Give me, you know, two paragraphs on why this brand is super important. You know, leveraging the generative AI then at that point to generate that content for you. It's built right into the system, so you can drop that in for, like, a merchandising banner, drop that in on a landing page. It's an extension of the existing functionality within HawkSearch, and you can use those for, you know, like SEO prompts and product details. Really powerful functionality and really kind of just kind of extending the existing functionality with HawkSearch by leveraging this generative AI tool in here. We'll be showing you examples of this, 'cause again, very important to kind of make sure that you guys understand the ease of use here, and just kind of how this, you know, create that content. Lots of other things that we can talk about. Changing the tone of the content, which I always think is kind of fun. If maybe you've already got existing content, and you just want to make sure, does this read okay? You can proofread it, you can enhance it, you can summarize it. Maybe you've got a three-paragraph description from your vendor about how great they are, and you need to break that down into just a single paragraph. We can have the AI analyze that and condense that down into a summary for you. So really cool stuff there. The other thing that we'll talk about here, because I know we want to get into the actual demonstration here, is the AI synonym generator. So, in the past, the way that this has always worked is that, you know, it'd be like, it's, you know, "Oh, they did this search, and then they did this other search," and it was really kind of a bulky way of doing it. It wasn't really very effective. With this introduction of large language models, we're able to now understand better what are the concepts that are related to the terms that we're seeing, and so we're really adopting this idea of the AI-generated synonyms, where basically you can go into the existing synonym tool within HawkSearch, give it the term that you're looking for, and it will then suggest alternatives by using that large language model to understand what is related to those terms. It's a very simple process then to just select those terms from that suggestion and add them to it. Importantly, though, I always think this is an important distinction, is that you're not locked into only using the AI-generated terms. You can still create your own. Maybe you've got other terms that you know you want to add. What the AI synonym generator is doing here in this regard is it's also help kind of catching those ones you might not have thought of or just streamlining it, 'cause it's really simple to select those instead of typing out all those individual terms. And again, we'll be showing you this as part of the demonstration as well. And this is just kind of the beginning of the conversation here when it comes to kind of where we're going with HawkSearch. As you've heard us say at the start of this call here, this is really our kind of initial release within Zeus here. So we're gonna be looking forward to future improvements on this, adding additional functionality with Athena and with Hermes. But let's get into the practical demonstration. What do you say, John? Should we start showing some things off here? Well, first of all, great background with the lightning. Yeah! That's a new one. I, I like that. We're really getting into it here. Maybe I'll just... Yeah, a little lightning update too, and this gets technical before Jonathan gets into the demo, but those two Gen AI features that Jonathan was just talking about, they actually do go through a middleware that we expose as well. So this isn't designed to be so technical-... but if people are interested, that endpoint is available as well. So you could imagine there are some prompts going on underneath the covers. Jonathan mentioned it briefly, but there is an endpoint that you could use as well that takes advantage of some of our expertise in making those Gen AI requests. But anyway, Jonathan, jump. Please jump right in. Yeah, yeah. So let's, let's kind of start from the top and work our way down, just like how we did through the, the actual, slide deck that you guys just walked through here. So we'll start with the Keyword Search. Well, again, we'll make sure that you guys understand here, this is something that we're not abandoning. This is something that's still very important to us. But again, the limitation with the Keyword Search is that, you can provide your kind of basic search terms, and it will find the matches and drive the faceting and filtering and pull back in the relevant results, but this is again looking for an exact match on that term. So I provided the term hiking shoes. We're now looking at the names of those products, where to find a match of it. We're looking in the categories of those products to find a match. That term has to exist somewhere for us to say, "Yes, this is relevant. We should return it." You can get around some of those data issues by creating synonyms and things like that, but that was always more of a manual process. With the Concept Search, this really allows us to understand what is related to those terms that are being put in. So the limitation, just to kind of drive the point home here is, if I was to say, you know, "What do I need to bring with me to cook on a camping trip?" If I asked this of kind of your classic Keyword Search-... We're getting one result, which honestly I think is a good kind of feather in HawkSearch's cap in terms of relevancy, 'cause we are stripping out some of the nonsense here that doesn't really help from a classical Keyword Search to identify cooking, camping, and trip, and find those maybe somewhere in the description of this product, and say, "Yeah, this is, this is gonna match what you've put in so far." But we know that there is a much better understanding of what this is if the tool can understand what is the concept behind cooking, what is the concept behind a camping trip, and that's what this Concept Search does. This allows us to take those existing product catalog information that we have, the name of the product, the category of the product, the descriptions of the product, run that through a large language model, and then with our Concept Search, I can ask the same question: What do I need to bring with me to cook on a camping trip? And you can see that we're getting back in much more relevant results, because it is understanding the concept of cooking. It's understanding the concept of a camping trip. Camping trip meaning that you're gonna need to have portable items, travel-size items, and so everything that we're seeing here then as part of our search results are related to those terms. Even though, as we just saw from the Keyword Search perspective, these exact terms did not appear within the product itself, it knows that, well, when we're talking about cooking, we're gonna be looking for things that are related to the kitchen. We're gonna be looking for utensils. We're gonna be looking for things related to those terms, and that when we're thinking about a camping trip, well, those need to be travel sizes. So when we scroll further down here, we've got the, you know, the travel stowaway pots and pans, the travel plates. We've got the food that would go along with this, 'cause again, it understands if you're talking about cooking, you're gonna be looking for food as well. These were all products that were available within the initial set of results that we did from a Keyword Search perspective, but because those exact terms did not appear in the product description in the way that I searched on it, it was not pulling back in those results. With the Concept Search, this really helps from that product discovery aspect of understanding what is the visitor actually looking for and understanding what that means from your product catalog perspective to return in these much more relevant results to answer that question that has been provided from the search box. That food, Jonathan, is what I had no idea that our demo catalog- Yeah. It's not a bad idea. You might want something to actually light a fire with or a fire with. Right? You know. Yeah, it's- Yeah ... it's crazy, 'cause as I mentioned at the top, I've been, I've been doing this for 10 years. I had no idea that we had food in this catalog. It was one of those just kind of crazy things. But let's give you another example, 'cause I think this is something that really kind of gets into understanding even more of what we're talking about here. So let's say, you know, again, we're looking at kind of an outdoor camping, hiking catalog. Maybe I need a winter, you know... I need a winter coat for my six-year-old. Now, what I think is really impressive about this is I've not told it that I'm looking for a girl's coat or a boy's coat. I've not given it a gender. I've just given it a year, but it understands that if we're talking about a six-year-old, well, those are gonna be not women or men. Those are gonna be your kids, your toddlers, your boys, your girls, things like that. The other thing I think is important here is, again, that idea of a Concept Search, understanding I've given you the term coat. You'll notice all of the results that we're seeing here are jackets. In our catalog, we only had things referred to as jackets. Typically, you would have to set up a synonym. You'd have to set up another relationship for that or maybe change the product data to account for that. But with the Concept Search, it understands that, well, you mentioned coat. Something that is related to coats are jackets, are parkas, are windbreakers, things like that, and so these results that we're seeing here, understanding the intent of what a six-year-old is, what that means for our product catalog, and the fact that you said a coat can also mean things like jackets that are a part of our product catalog. Yes, Jonathan, it's almost like you don't really need the synonyms anymore, right? No. No, really. Yeah. Well, let's go ahead and show you then the other end of this, 'cause while as impressive as I think this is, where I think the really... the kind of thing that makes this product sing, the Zeus release really sing, is the Image Finder here and the Image Search. Again, baked right into kind of our presentation here. We can go ahead and select into that. It'll give us our nice little Image Search window here. Again, if you're on mobile, you can choose or drop photos. This will go ahead and open up the camera application on your phone or, of course, give you the ability to choose from your library of existing photos. I'm here on a desktop device, so when I click Choose or Drop Photos, it's gonna open up my downloads window here. We can, of course, drag and drop photos from other areas, but just in the interest of time and demonstration, I'm gonna select a photo that I have here already, available. And you can see here, it's just named Winter, so we're not even looking at what is the name of the file. This is using the AI to really analyze what is in this image, like we talked about before. This is not just relying on visually similar results. It is using the AI to analyze this and say, "Well, what do we think actually is in this image?" And so we can see here the AI is saying, "Oh, well, we've got some ski goggles. Oh, we've got a helmet that this person's wearing. Oh, we've got a jacket." It's understanding what belongs in this photo and then using the same AI process to look at our product catalog images and find what matches against that. So it's not just, "Oh, these look visually similar to that," because if you notice, they're not. This is more of a profile shot for this product versus the one that's looking directly at the camera. This is instead using the AI to say, "Hey, we think these are ski goggles in here," and looking at our product catalog and going, "Oh, we think there are ski goggles here as well," and returning that as a match. But where we get really exciting here is that as John said, it's not just looking at only the one result. It's trying to understand everything that's in this, and I always like to joke that, you know, we're not all professional photographers, so maybe what I really was looking for when I took a photo of this gentleman was the helmet that he was wearing. Well, the AI also will understand that. Keep in mind the perspective that we're looking at this helmet in, it's looking directly at the camera. When we get a little bit further down into our results, the AI starts to say, "Okay, well, there are other things in here. Let's start to make suggestions for what else you might have been looking at in this photo." And look at the helmets that are in our product catalog. They're entirely in a profile side shot. Does not look anything at all like what was actually uploaded in the image that we have here, but because the AI is not just looking for what's visually similar but analyzing the image to say, "What is actually in this image?" and then using the same technology to understand what is in our product catalog, it finds those, those matches and says: "Oh, there's a helmet in this, and here's a helmet from our product catalog as well." Even if it's not the front and center thing, even if it is, you know, off to the side, again, perspective's different, lighting's different, it's able to understand what is in that image and what is in our product catalog and return relevant results for that. Because we're doing that kind of analysis of the product catalog images as well, that drives the other really cool thing that we can do here. So maybe I don't have the image, I mean, I don't have the thing in front of me, I don't know exactly what it is, but I can describe it. I can say I'm looking for that book with the butterfly on the cover. Since this is a retail kind of clothing catalog here, I can say, "You know, I'm looking for a hat." And when I go ahead and say I'm looking for a hat, of course, it can understand that and start to pull back in hats. But I can further refine that. I can say, "You know what? I want a hat that is fuzzy." And I'll go ahead and say, "Well, here's our fuzzy hats as part of that." And, you know, I'm looking at this and I'm like: "Oh, you know, it had a... it had that pom on the top there." So I can say, "With a pom on top," and that'll go ahead and pull back in all of the products that we have within our catalog that have a pom on top that are fuzzy. And I can say, "Oh, you know what? I forgot it was in pink." And then the HawkSearch image analyzer looks at our product catalog images, understands all of the terms that we've provided here, and understands that, hey, you're looking for a fuzzy pink hat with a pom on top. Here are all of the fuzzy pink hats with poms on top of them. Hey, John, one thing I've noticed as you're typing, it's just dynamically updating. It's not you have to hit return or anything, right? So it's the instant search feature as well, which we're not really talking about that, but it, you know, it's just instantaneous results that the user's- Yeah ... getting as they're refining their search query, which is super powerful. I know a lot, a lot of our customers love that feature there. We just... as a matter in the comments here, I saw a question around just the image load of their- Mm-hmm ... their data, right? They load their data in. We can ingest that, but also the images. Can you guys explain a little bit around getting that image data, getting it into that pipeline that you talked about earlier, John? Yeah, I'll, I'll jump in there, but Jonathan also, follow up too. So what normally is sent when we say image, and we're showing up, obviously, when Jonathan is seeing these results, is the URL. So even in general, forget anything we're talking about here around AI, when people send us data, the image is really the URL. So that pipeline that I was saying before, the way this AI works, this, API, is it says: "Go to that URL, grab the image, and convert it into a vector," and that's what gets stored on the other side. So it's pretty crazy, and we could talk about the philosophy of this. And I like to talk a lot about, Moore's Law. So you're probably familiar with this. Gordon Moore was the leader of Intel for a long time, and he said, "You know, space is going... The whole, circuits and all of the capabilities around chips was gonna double in capacity and half in price every two years." But in this world, whether that picture, you had some fun ones there, Jonathan, of, you know, kids' hats and jackets, et cetera, these could be very large file sizes- Yeah ... 10 MB, 20 MB, 100 MB. It's just going to the URL where it's hosted, and it's just converting that into a set of vectors. So those set of numbers is the same, whether it was a 2 MB file, a 50 MB file, or what have you. What we're storing is the vector result that describes that, so that you can either ask the question, like Jonathan was with the pompoms, or by taking the picture with your phone. That's kind of the mechanics of how all these moving parts work. Okay. We already have that data on hand, right? It's already- It's already in the- In your system. ... in the system. Yes. Yes. Correct. Yes. Good point. Yeah, so you're not having to upload the images to HawkSearch and say: "Oh, here, you know, I need to upload... maybe I've got a catalog of 1 million products," and then you have to then upload all of those to HawkSearch. That is not the way that that works. We're just using that URL path that says, here's where our existing product images already exist. Feel like your product detail pages, wherever that information is already living in your system, you're just telling HawkSearch: "Hey, this is where that image lives." And we use the same information for then how we even return the image results from, like, a Keyword Search. But what's interesting there, and kind of maybe simplify what John was talking about, is we're then taking that URL and saying, "Okay, well, now we can take that endpoint and run that through the AI analysis." So you're not having to upload all those images. We're leveraging the existing information to understand what what is actually in those product images. Yeah. That's one great thing I wanna call out here with this release. We've made it very easy to leverage this technology, right? It's not like we have to go through a huge implementation to make it happen. It's already there, and it's just a matter of toggling things on and off. And also the Rapid UI front end, you know, that's a pre-built front end that we make available out of the box. It's fully responsive, mobile-friendly. You know, so you can even start with that and then just stylize it, right? Yeah. A simple style sheet exercise to tailor it to your brand, so, Yeah, but that's a great point, 'cause everything that we've been showing here on this site is using that Rapid UI, which is a really straightforward, kind of, you know, like 50 lines of code to get up and running on your site, and you get all the kind of functionality we've been showing you here. Yeah, it's awesome. But to that configuration point, maybe we can transition to some of the back-end functionality that we were talking about there with that generative AI. So we're gonna go into our HawkSearch Workbench. Obviously, you can see we've got lots of reporting analytics. There's a ton of features, as Carl was kind of mentioning with his eye chart slide there at the top. There is a lot that HawkSearch can do here. And, you know, we'd love to have a conversation, dive into some of the other functionality that HawkSearch can bring to the table for you. But in the interest of this kind of conversation, we're gonna really focus in on our Hawk AI functionality here, because this is what was a part of our Zeus release and really gets into some of that back-end functionality that we've been wanted to show you guys here on this call. So as I mentioned, I'll touch very briefly. There is an existing AI multipliers. HawkSearch has been working with AI for years now, but it's always been from kind of the improving the Keyword Search functionality. So we're able to leverage the different signals of what your visitors are doing with your search results, your classic Keyword Search results, and improve those by tracking. Are they actually interacting with them? Are they adding those to their shopping cart? Are they converting off of that? Using those signals to improve just the core Keyword Search results. Of course, being able to fully personalize those results as well here. Really wanna make sure that you guys are aware that even when it comes to the kind of classic Keyword Search functionality, we do have a lot of AI functionality that's helping out improving those results... but we're obviously looking towards the future here, and that's gonna be things like our Content Assistant, which is what we were showing you earlier, where you can get that kind of prompt approach to, "Hey, I need a paragraph about this." This is right here within our tool. It is an extension of that existing content items thing within HawkSearch. So when I click add a new content item here, I wanna drop in some rich text on a landing page, I need a paragraph about something. Well, I can go over to our content item section here. Say, I wanna use that custom HTML functionality, that kind of classic WYSIWYG, you know, text editor that you have here, and now there is a new button here where we can generate that, with a prompt-based approach here. So I can say, you know, "I need a paragraph about the importance of light bulbs," let's say. And I work in the B2B industry, got a nice light bulbs category page that I'm looking for here. Can ask it to go ahead and generate that information, and here we can see now we've got that prompt that's already been generated for us, talking about how important light bulbs are, you know, how important they are to kind of your everyday life. I can just go ahead and insert this right into the tool here, publish that, we're good to go. One of the other cool things, maybe I'm like, "You know what? This is just a little too formal. Maybe I need..." You know, we're a hip and cool lighting company. I want something that's a little bit more in light with our business proposition here. So I can change the tone of this and say, "Maybe I want this more informative, maybe more persuasive, more formal." I always like the casual example here, 'cause we can, "Hey, you know, make that paragraph about the importance of light bulbs more casual," and we get, you know, "Yo, let me drop some knowledge about light bulbs and why they're so darn important." So obviously, changing the tone drastically there. But if I like this, I can say, "You know what? Go ahead and insert that," drops that right back into the tool here. I can save this, get that published on our landing page, have that show up when someone navigates to our light bulb or light bulbs category. Really kind of streamline that functionality there and, and really help, you know, cut down the, the time that you need to spend in the tool here. But there's additional functionality here as well. Maybe we've already got this existing content and I want to improve that. Well, there is now this dropdown here, where I can have the AI automatically change the existing content's tone. I have it proofread that, enhance it, summarize it, paraphrase it. So it's not just new content that the generative AI is available for, we can leverage this on your existing content as well. And kind of building on top of that, we thought, well, You know, kind of prompt-based approach is so kind of neat. Where else can we leverage this within the tool? And again, that was where we thought about that synonym generation. So again, if we go back into our Hawk AI menu here, we do have that, its own dedicated section here for synonym generation. Kind of walk you through what that is, but again, it's an extension of our existing synonyms. So if I click Get Started here, this takes us directly to our Keyword Search section, our synonyms section, and then I can go ahead and click Add Synonym here, and now we have our kind of classic synonym functionality within HawkSearch, but now we have this new AI synonym generator. So if I give it that search term for continuing on the light bulb theme, let's say I'm looking for a flashlight. Well, when I put that term in there, we're gonna go ahead and leverage the generative AI to understand, well, what are the things that are related to this? This is using the same kind of large language model. Like Carl said, you know, using the Concept Search, synonyms are really kind of a thing that you don't need for the Concept Search. But supporting the classic Keyword Search functionality, moving forward with that, you can leverage that large language model functionality to start filling in the gaps that you might have with your Keyword Search functionality here. Say, oh, you know, well, we wanna make sure that we've got... We're capturing someone wasn't put it in space, or if someone was to, of course, use the, the British version for torch versus flashlight, or maybe they're referring- they're getting very technical and they want a handheld light. I can just select these from the list here, from that AI generation, save it, and we're good to go within this functionality here. So really a very cool kind of extension of the existing HawkSearch synonym functionality by leveraging that generative AI. Really, I just think it's a really neat kind of extension of the tool and really showcasing kind of how we're moving forward within HawkSearch by not just focusing on kind of the exciting stuff, the Concept Search and the Image Search, but also understanding that, you know, kind of the here and now, people are still gonna be looking to, you know, do Keyword Searches, and we wanna make sure that we're taking advantage of all of the AI functionality that we have available to us to really help deliver for our customers. John, Carl, anything you guys would add to the functionality that we're talking about here, as we start to talk about wrapping this up? Just a quick comment on... Yeah, so that I know we're showing keywords separate from concept, right? But we do have the ability to have that be a combined experience, right? So they just enter in one value into that query, you know, and, and providing for a unified result set- Mm-hmm ... right. So that's something that we can incorporate as well. You know, so- Yeah ... I know we're filling in our demo environment. But I just wanted to clarify, that's something that can be made available as well, right? Yeah, Carl, good, good catch there, because for obviously, for demonstration purposes, we wanna be able to show you the difference between kind of your classic Keyword Search and what would a Concept Search look like. And so we're breaking those out into two kind of very distinct experiences here. But, and John, I think you were talking about this earlier, this is all gonna kind of be one endpoint then, that you can hit that one request, and we can return back in kind of all of that functionality for you. Yeah, and Jonathan and I have a side bet going on, like, well, and these will all be options, 'cause I personally would never say, "Nope, you have to do it this way. Mm-hmm. It really will come out, like, what is the best user experience here? You know, maybe what we're kind of, not clumsily, but just kind of very crudely saying is the Concept Search. Maybe that's, "Didn't find what you want?" or something like that. Yeah. Well, there's a lot of creativity here. I do like the idea of just giving options and, you know, figuring out what makes the most sense. Even the irony in some of Jonathan, our other side bet, is, you know, uploading image is wonderful and that's cool. Yeah ... but maybe it is the Image Finder that really catches on. So we don't know, obviously. That's where innovation and all the creativity will come out. But I'm curious, even, and we love feedback and, you know- Yeah ... people have been testing this, et cetera. We'll see. We'll see what is the best user experience. Yeah. That's a great point, and I know we do support A/B testing, right? So, like, if someone's using- Yeah ... a A/B Testing tool, you can have two different types of search experiences and see which one works better for your audience, and- Totally ... have that be the winning experience, right? So- Yeah. Yeah. ... yeah. Yeah? Another question came in just in regards to those images. Again, I think you mentioned it earlier, John, just the real-time aspect of our API, so being able to pump data in, including images, in real- time. So if you have a product update, that image changes, you wanna take a new picture of the image, you load it in, call our API, boom, it can update in real- time, right? We're not waiting for a nightly job or anything like that to run. Right. There. Yeah, and maybe just to spend one second on that, let's just imagine the simplest model you could imagine. I have a product catalog, and it only has four fields. It has a title, a description, a price, and a picture. I know that's very simple, but let's just say... You're already a customer, and you already send us this data. As you just said, John, Carl, the picture changed. You know, what can I do here? You would just literally call a single update in a JSON request using our API with that image. And by the way, it could be the same URL, maybe it's a different URL. The trigger for generating this AI stuff is just that you re-index that one item in this case, and the text-the title is gonna be stored as it always was. Excuse me. The description is gonna be stored as it always was. The price is gonna be stored as it always was. But the image is gonna have two actions. It's still gonna store the URL. I mean, that's always gonna stay there. But then it's gonna call an endpoint and says, "Hey, grab that image from that web server and create a vector for it, and put that in another field called vector field." So moments later, you do a search, same thing, describing the butterfly or whatever it is, and that result will come back in that next time someone searches. Yeah. Very cool. Just the butterfly thing, I was just thinking, like, logos, right? Like, taking a picture- Yeah ... of a product with a logo on it, seeing other, I don't know, Adidas products or whatever, right? So depending on the catalog, of course, but that's certainly- Oh ... a common use case, searching for brand, right? You know, that's simplified- Mm-hmm ... just by taking a picture of that, you know, a product with that brand on it, so. Yeah. Yeah. It's so funny. That wouldn't come up in the example we were talking about before, Carl, but it is a good point. Most likely, your data has a field for brand, most likely. Yeah. But let's say you didn't. Yeah. But if all of the images have that logo on the picture... like, guess what? You search for that brand name or you take a picture, as you were just saying, it will return all of those. That's why it's gonna be very interesting to see how these things, you know, work out. The metadata is always gonna be critical. We're certainly not saying move away from that, but the AI can fill a lot of gaps- Yeah ... that might not be in your data. And you know, we deal with people who are challenged with, "We get data feeds from multiple sources," and- Yeah ... you know, "it's inconsistent." And, you know, Jonathan's been living with that for quite some time, and this helps smooth out a lot of those challenges. Well, just even looking here, like, using the descriptive term of fuzzy, you know, would should I have used wool? Should I have used knit? Should I have used, like... You know, the way that the product information might have been stored, it can be inconsistent, can not be uniform, and so really, it's about kind of eliminating the wall that your visitors would run to when they would use terminology that didn't exactly match with what your product catalog had. And that can be that, you know, multiple vendors referred to it different ways, or just that you didn't think of how a visitor might actually think to search for this. And so it really kind of breaks down those walls, and that's kind of the idea of, this isn't search anymore. This is product discovery. This is really kind of exposing out what your product catalog really is and getting out of the way of letting your visitors kind of find what they're looking for, using the terminology that they're expecting, either through taking a photo of it, describing the photo of it, using kind of that concept approach of, "These are the things I'm looking for," and letting that understand this is what your product catalog actually should be. Yeah, super important stuff, and I think, Carl, to your point there, that idea of that this isn't a major process that needs to run in the background, and that this isn't something that, "Oh, I made a change, now I need to wait till tomorrow so this is reflected." No, that... Because this is such an extension of how HawkSearch is already working to drive Keyword Search functionality, we can do those kind of deltas, those real-time updates for your existing product catalog. This is just an extension of that, functionality. So when you make a change to something that already exists in your catalog, using our, our existing functionality to make those real-time updates, well, that then talks to the next part of the process here to update, well, here are the vectors that we do now for the Image Search or for the Concept Search. And, you know, you're not waiting on that kind of, delayed response, then. This is not a, not a, not a problem within HawkSearch. Jonathan, can we be expecting you to be wearing a pink, fuzzy hat with HawkSearch on? I found what I was looking for here. I hope. I got my I heart ear flaps hat for girls. There you go. It'll cover my headphones nicely for the next one. Okay ... I put it on. Good, good. Nice. Real quick, I was gonna pick up off of that fuzzy keyword, and I have a little fuzzy guy here, so I- Oh, is that it? That would definitely come up in the search. I nicknamed him Hawkeye, and- There we go. ... I'm just gonna make an offer to everybody on the webinar here today. If you follow up with us, have a demo, you become a customer, we'll send you a swag pack with a Hawkeye in it. So, Excellent ... that's my little gift to you for joining here today. There we go. I like it. I, I don't even have one of those, and- I know ...... I've been here 10 years. He's coming to get you. It's gonna be a job. But no, if you do wanna go ahead and get started with us, I'll pass this back to Victoria. I know she's gonna talk about the QR code, and she's gonna wrap everything up here. But just to kind of start that conversation, I really do wanna thank Carl and John, and thank everyone for joining here today. Hopefully you got some good information here. There is that QR code. There is the email address there, hello@hawksearch.com, and scan either the QR code, email us at that address to book a demo, get more information on all the AI functions that we covered today, talk about all the other features that, you know, Carl ran through at the very top of the call as well. So we'd love to have a conversation with you. But, Victoria, I'm gonna pass it back to you. You can, you can- All right ... wrap things up here. Yeah, and I, I did see a couple of comments and questions. If you haven't reached out to us already, this is a perfect email address to reach out, especially if you want that swag pack with the Hawkeye stuffy or if you want us to set up a customized demo, so one that's really specific to your business use case, your industry. That is essentially what that QR code will help you start to set up. But I do wanna thank you all, John, Jonathan, Carl. This was a phenomenal webinar, and we will be sending out the recording to everyone who registered, everyone who joined us today. And yeah, we will see you guys all next time. Thank you. Excellent. Thanks, everyone. Thank you. Bye. Take care. Bye.