Hello. Good morning and good afternoon, everyone, and good evening maybe to some. Welcome to this quarter's new and Coveo webinar for service and workplace. My name is Claudine Ting, and I work on the global marketing team here at Coveo. Today, we'll be focusing on the latest features and enhancements we've developed that can quickly connect your customers and employees to answers and gather insights even across silos, making every digital experience relevant and useful. Our presenters for today are Bonnie Chase, director of product marketing, and Esme Bouchard, product manager. Before we get started, I just have a few housekeeping items to cover quickly. For this webinar, we'll entertain your questions maybe at the halfway mark or at the end of the presentation. However, we encourage you to type your questions on the q and a box and the chat box as we go along. Lastly, today's session is being recorded, and you'll receive your presentation in your in your inbox within the next few days. Now I think we're ready to get started. Bonnie, Esme, take it away. Thank you so much, Claudine. Hello, everyone. It's great to be here today to talk to you about some of the exciting things that we are rolling out with our latest release. And, you know, really to to think back on the year, part of what we've been focusing on is really about accelerating self-service success. So, you know, some of the things that you'll see that that we talk through today are really all about making sure that, customers and whether they're customers or employees, they're able to get the answers faster. Part of accelerating self-service success is, you know, building out these new features, these new capabilities, but also part of it includes, making sure that the technology is easy to deploy. So in everything that we built, we wanna make sure that you're able to start seeing success and start seeing success sooner. So really excited to share with you some of those things that that we have. And with that, I will hand it over to Esme to really tell us about this first exciting capability that we are releasing with this release. Alright. Great. Thanks, Bonnie. Hi, everyone. So just quickly, I'm Esme Bouchard. I'm a product manager here at Coveo, mainly on our service line of business, but I'll be covering both service and workplace. There are some features that, overlap in those, both of those solutions. So to get started with, smart snippets, so what is this? Some of you have maybe heard, of Smart Snippets before. We have been talking about it for a few months. But just as an overview, everyone here is most likely familiar with Google Featured Snippets. So that's certainly something that you've interacted with before. So when you ask a question on Google, we rarely, only get links as a result anymore. We get an actual answer that's been extracted from the content in the web pages that, that has been indexed by Google. So we want to do the same thing and allow you to offer that same really great amazing experience that everyone expects now, to your own customers, partners, and employees. But using your own content and on on all of your digital experiences and channels. So smart smart snippets is a new Coveo machine learning model. It's able to extract answers from any structured content that you already have indexed on the Coveo platform. So content like HTML pages and knowledge articles. And then feature those answers at the very top of search results when a question is entered into a search box. So, you know, let's face it. When people search, what they're looking for is answers. They're not looking for, links to results. Right? So we wanna really shorten that path to get to an answer, and this solution is gonna allow us to do that. So I hope that that explains it quickly in a nutshell. We can see, you know, an animation of what it looks like here on screen as well. Yeah. And this is this is really exciting because this is one of those things that that we've been we've been talking about for a little while. We're really excited to get it in your hands, get started using it. Can you tell us a little bit about what it takes to get started and how they can start leveraging this today? Yep. Absolutely. So, all of our enterprise customers will have access to create these types of models starting on July fifteenth, which is coming up really soon. But until then, if you're interested in getting started with this right now, you can reach out to your customer success manager, and we'll be able to start building this with you and evaluating how to to set up the model, before, before the UI is available. And speaking of which, I'd just like to give you kind of a sneak peek of how easy it is gonna be to set up a model like this. So I'm gonna jump into a demo to show you that real quick. All right. So, I'm here in the Coveo admin console. And so in our machine learning section here, when we click on add a model, which some of you are maybe familiar with, we have the option to create, the different, you know, one of the different ML models that Coveo offers. We give our model a name. And as I mentioned, starting on July fifteenth, you'll have the ability to create a smart snippets model here. And I just, you know, you'll notice this is all clicks, there's no code involved in creating a model like this, it uses content that's already indexed in Coveo. So when I hit next, first step here is we want to, we want you to select, you know, what sources that you have. So, what we want you to point to the content that contains those answers that you'd like to feature that you'd like the Smart Snippets to feature. So you'll select a source, one or several sources that contain that structured content. So again, HTML knowledge articles is what is supported right now. Here, important to, just fill out a few configurations. Right? So we wanna know what fields on that content contain, the the possible answers to different questions that your users might have. So document type might be a KB article and the field might be called the body. This will be, a little bit different for for every customer. We have this is just a a step to exclude any CSS selectors that might be, somewhere in the content. We don't want those being featured, so we'll just exclude those if there are any. And that is it. So it's really a four step, very, very simple process. Content's already in your index. At this point, you can review your configuration and hit add model. These are a deep learning model, so they do take a little longer than our conventional models that you have maybe used before. So they can take up to twelve hours to build. So you hit add model, come back the next day, and you can already start testing, your new model. So it's really as simple as that. So really excited, to be, you know, releasing this. It's a really cool feature. Ashley, we did get one question on this one. How are candidates created if there are no explicit clicks to confirm? Yep. So how the model works right now is really it is built on structured content. So what we're able to do is, compare a question with, the headers or kind of the section separators in that structured content and, feature the content that's, that exists under those headers. So that's in a nutshell how it works right now. We are working on a next generation which would use usage analytics as well to be able to extract those answers. Great. I'm also, going to add some questions from the q and a box. One question was great to see the move to click. How effective is the model compared to custom code? So, the compared to custom code, I've I actually haven't heard of, of a customer who has built a custom built, like, a question answering system on Coveo. So hard to say. We've, piloted this model with three of our customers, including ourselves. The animation you're seeing earlier is an actual, you know, demo of the model working on our own content for our own website, and it's working quite well, for for questions, you know, and answers that exist in the content, of course. So I hope that answers. If if, I haven't, please let me know. I'll try to add some detail there. Yeah. It's me. So there's there's a question about how machine learning plays a part in this, you know, more of a search than machine learning. You know, what are they missing? And I just wanted to take a step Barca. And and from my understanding of of how the machine learning works on here, it's actually looking at the page structure. So it's looking at the back end seeing, you know, where are the tags that say this is a header, this is a multistep process, understanding what's where the question lies and where that answer would be and then surfacing that within the search results. So we're not really programming saying for this search, provide this answer or for this this type of question, surface this particular section of the article. We're not preprogramming that. It's more the machine learning, looking at the structure of the document and understanding from that document where the answer resides and surfacing that in the search results. Results. Is that accurate, or would you add more to that? Yeah. And I I think, I think that provides a good part of the answer. As to, the specifics of how the machine learning works for for this particular model, I I'm not our our data science expert, so I apologize for not being able to, have more details on that. But if you wanna follow-up, we could answer maybe in more detail, to help with that answer too. Yeah. And this is leveraging question answering technology. Correct? Mhmm. Yeah. Exactly. Alright. And then when you say knowledge articles, are you referring to Salesforce articles? That's one of the questions. Yep. So, I believe we've tested this with Salesforce knowledge articles as well as ServiceNow, knowledge. Typically, knowledge across different knowledge, you know, solutions do have a structure to them. So typically, something like a a subject, a problem, a resolution, and typically separated in that way with headers. And so that is a good candidate for our model to learn and and extract the answers from. So, my understanding is that it it works on any type of knowledge. To be confirmed, it's been tested on Salesforce and ServiceNow for for now. Great. Thank you. Claudine, any other questions on your end? Yes. We do have a couple of questions here in the q and a box. So, first, how does it determine the search if the search is a question, is it just the question mark at the end? It's not, actually. And, what's interesting is, is that we initially wanted to call this feature question answering because that was the the initial idea. The model no. It's not just a a a question mark. It actually is just, comparing the the the search terms with what we're able to extract in terms of an answer. If we have if we have a a high enough confidence that we have the answer in the content, we'll feature it. If we don't, we just show regular search results. That being said, the reason we stopped we decided to not call it question answering, but rather smart snippets is because it works on any type of query. As long as we think that we have an answer that is close enough to what our user is looking for based on, you know, what they're intending to to search for, then we would we would feature that snippet. Okay. Thank you. So, do you mind maybe we should take a pause on the questions, and then we could, go on with the demo for the meantime. Yep. So I I had finished for this particular demo, but, yeah, if we have time at the end, we could circle back on questions if Okay. If you want. Yeah. Sounds good. So I'll go ahead and and move on to the next, the next feature that we wanted to share. So, let's see. And if you're able to see my screen, you'll see that this this next feature that we're talking about is case assist. Now we have, we have made a few releases including case assist. We did launch this at the end of last year, but this is something that we're going to continue to innovate on. One of the things that we're seeing with our customers and in the market is the need to go beyond case deflection for self-service success. So case assist is really that feature that streamlines the case creation and deflection process. Now, Esme, you've made some enhancements to this in this quarter. Can you tell us a little bit about what those changes are and and how that impacts, those who are interested in case assist? Yeah. Absolutely. And and maybe just to to get started on on that particular topic, you you touched on it, but just for those who aren't familiar, with Case Assist, just a quick recap on this. So as Bonnie mentioned, we released this, at the end of of last year. This is really a solution focused on the case submission process and making sure that that process is smooth for the customer, that they have a good experience. But at the same time, we try to deflect that case if it's possible, and or collect as much information that is accurate and complete as as possible when, when the case is submitted so that when it ends up in the hands of an agent or tech technical support engineer, they have everything they need to solve the case fast. So the the two features available are, in in, you know, under the case assist solution, our document suggestions for case deflection, as well as case classification to help, help users classify their own case as they are creating that case. And that's what you see in the beginning part of this animation here on screen. What's the main thing that's new here this summer and that we're really excited about is a new machine learning model, for for doing even more powerful case classification. So what we're trying to do is based on, the description of a of a case, so of a of a problem that a customer is having, we're able to try to predict how to classify that case. So, you know, what product it's about, what type of case it is. Those are the the common sort of classifications that we ask users to to fill out, and we try to narrow down the values that are possible, for them to choose from and make it easier. So what's what's new is, a new more powerful, case classification model. It's built it is actually a deep learning model, which is new for us as well, a little bit more complex. And, I'll I'll jump into a demo actually to show you how to set that one up, as well if if that's okay with you, Bonnie. Yeah. Let's do it. Okay. Cool. So, again, you know, with this case classification, it is based on machine learning to automatically identify potential potential options, classifications that the customer may choose from, in order to submit their case. Yeah. Exactly. So, again, same as as, what I just showed for smart snippets. This is a a clicks only experience. No code required. We wanted to make it we wanted to make it as easy as possible. We're using content from, the index to, train, this new model. So, again, give our model a name, and you'll notice this new option to create a case classification model specifically. Again, we need to specify what source contains, the cases that will make up the training set for this particular model so that they can learn from. So we'll pick a Salesforce source here, and then we can also narrow down the content from that source, to really point out which cases we really wanna use to train the model. We're recommending using the last six months of cases. You might want a little more, a little less or a custom kind of, fixed date range for those cases, and then you can apply further filtering. A typical case that we've seen is is if you have a field that tells you that the case is closed, we'd only want to learn on closed cases so we could easily create that filter. And that kind of defines our data set to train the model. The next step, we do need to identify which field represents the case ID, and then specify what fields contain the content that we want the model to train on. Typically, this will be, your subject, your description, but, maybe your cases have a few more fields that contain some relevant content like, resolution, details about the problem, the business impact, and things like that. You can include all of those just to reduce the noise that the model has to deal with, and then, declare all the fields that you would like the model to predict. So as I mentioned earlier, what we often see is things like, product, something like product type, you know, that would be a custom field in Salesforce or, the reason, something like case reason. So I'm just declaring two here, but you could have as many fields as you want. The model can can predict, up to, you know, a dozen fields if if needed. And that's it. You get a summary of your configuration. Again, same as with smart snippets as the deep learning model takes a little longer to, to train up to twelve hours. So you hit add model, come back the next day, and you can start, testing, see how how it performs. So it's really as easy as that. So really happy to be able to offer that to you guys, starting now. So all of our enterprise service customers already have access to this, in the platform, and that's been about a week now. And that's it for Adebay. Thanks so much, Esme. So in thinking about, deep learning versus typical machine learning, can you share a little bit about, you know, what what that difference is at a high level? Yeah. At a at a high level. Right now, machine learning is what we, kind of started with, learning from our usage analytics data, so how your users interact with our content, so what they search for, what they click on, pages that they view, to be able to learn from content in the index, which is, you know, a source of a lot of relevant information, very valuable information. We had to move to a more complex type of model, which is deep learning, to be able to process all of the content that's available in the index. It's a lot more data to train on than just searches and clicks, which is what we had done until about, let's say, say, a year ago. And that's what required us to go into the deep learning sphere. Let's see. Great. Alright. So we do have a couple more features to highlight some improvements to. The next one that we're going to talk about is in product experience. So, again, in product experience is something that we launched, prior to today, but we do have some enhancements, that are coming out. So just to to remind you, in product experience is really about bringing knowledge where you need it so that people have access to that when and where they need it. And so it's really, you know, whether you're embedding answers and the search and answers within, within your product, within a website, a chatbot, you know, something like that. You know, we really wanted to make that more accessible. So with this latest enhancement as me, what is new and what can people expect to see? Yeah. Totally. So, I think you said it well. You know, IPX or in product experience is something we released a year ago, and it's something that allows you to really integrate kind of a small Kaleo search page or recommendations into any of your web apps, your cloud product, your website. But to fully configure that in the platform and then insert it into those, you know, apps or websites with just a single line of code. So it's it's to really make it really easy and simple to bring Coveo content into different digital experiences and leverage that content that you've already indexed and configured in Coveo already. What we realized and so what's new at this point is that this is a solution that can be very valuable for our workplace customers. So to bring that content to our customers' employees wherever they're working. So as you can see on screen right now, for example, in Google Drive, instead of having to open up a separate search page or look through different systems, why not bring that content to employees where, they're working? So, what we've added this summer to cater to that specific use case is a Google a Google Chrome extension, which makes it easy to use, Koveo in product experience in here, like, Google Drive, in your intranet, or another employee website that your employees work with. So it's an extension that's available on the Chrome Web Store that you can ask, employees to install. And all you have to do is declare what URLs or website that you want it to be displayed in, and and that'll give, your employees easy and quick access to Kaleo content that might be relevant to them, wherever they work. So that's, that's the main enhancement here. Great. And that, so as you mentioned, downloading through the Chrome Web Store, and it can be configured in the admin, in the admin console. For those who are interested in the end product experience, it's really about, I think reaching out to the customer success manager. Is that correct? Yes. I I think that that's a good path forward just to make sure that it's set up correctly, to do the right testing, but it is available on the Chrome web store. So if you wanted to just test it out yourself as well, that's that's also an option. Okay. Great. And the last update we have is for our service connector. Esme, can you tell us a little bit about what has changed with this connector? Yeah. Sure. So, again, just to to set a bit of context, Coveo for ServiceNow essentially has two parts to it. So our solution has two parts. We have, really a package of prebuilt components, for integrating into a ServiceNow environment. So, to be able to have a Coveo search page or recommendations, right, anywhere in ServiceNow. And we also have, in the second part, a connector to index ServiceNow content in Coveo and surface it, in any of your Coveo interfaces. So this summer, we've, revamped the connectivity piece of Coveo for ServiceNow, and, it's actually what you see here on screen. We've really made improvements to the configuration interface for, when you're creating a ServiceNow source. So you now have access to the full list of objects that are available to index in your ServiceNow org, as well as the fields on each of those objects. So this allows you to get really granular about what you want to index or not in your ServiceNow source. ServiceNow can have a lot of content, so we don't want, those sources, you know, kind of, exploding your your your index. So this is gonna allow for a lot more flexibility and also allows for a simpler and more user friendly UI. So, you can go check that out on the platform now. Awesome. Alright. Well, that concludes what we what we are sharing today as far as new and new features and enhancements. Before we go to questions, I do want to let you know that tomorrow, we will be joined by Salesforce to as they present, how they strategize, to create a scalable, support experience. For those of you who don't know, Salesforce is a Coveo customer. So if you're interested in seeing how they deploy us, and their approach to customer support, please join us tomorrow. Now, I think it's time for us to take some questions. Claudine, why don't you go ahead and and kick us off here? Sure. So let's take on some of the questions that were asked earlier, in the previous demos. Let's get started with, you've mentioned piloting it on your own website. I'm assuming this would be smart snippets. How relevant were the answers provided, and was there any benchmarking to show increase in relevancy? Great question. So, right now, these pilots that we mentioned are have been internal pilots. So this isn't something that we've, deployed live anywhere yet. So I just wanted to mention that. Regarding results of the pilot, I wouldn't say we have benchmark, numbers equate yet. However, we have done just some, testing of of common queries that we get. Right? So in Kubeo Analytics, you can take a look at, you know, what are the popular keywords, what are the popular questions that are asked in our search. And, we're actually very, pleased with with the the quality of the answers that are extracted there. So, I mean, I think that's sorry about the generic answer to that, but we will be, you know, as as customer starts to use this and we can analyze the relevancy, we'll be able to, offer more details in that front too. Great. Okay. Bonnie, how about you take this one? Yeah. So, we have Coveo for psych four. Will smart snippets work with our instances of Coveo? Oh, yeah. Great question. So smart snippets is really a platform feature. It's a machine learning model available to, that will be available to all of our enterprise customers and can then be used in any Coveo interface. So we already have a component in the in our search UI framework that that automatically handles displaying results like that. So as long as you're up to date on the JS UI, you can use it in in Sitecore. You could use it in Salesforce, in any type of integration where you're using the JS UI, and it will soon have a component in Coveo headless to be able to handle it automatically, as well. Awesome. Okay. So the next question here is, is this model dependent on view events? So to my knowledge, no. Right now, it's really learning from the content itself. Thank you. Alright. Does Covey will inter integrate with Interact employee portal? Interact so Interact is a an, an Internet solution, I suppose. So I I I suppose that it is. So I I honestly don't see why not. Sometimes some solutions have, strict security requirements when, you know, referencing a third party, APIs or third party solutions that have to be that get embedded into their solution. So, let's let's check and test it, but I don't see why not. So the idea is to be able to insert, the IPX into any into any app or websites. Thank you for that. So next question we have here, is there a limit to how many snippets will be suggested for a search query? So the way it works is there's just one snippet per question can be showed at a time, so we wouldn't display more than one. I hope that answers the question. I hope the person will follow-up if I interpreted wrong, but so you ask a question, you get one one snippet. And then and then below that are the the regular search results. Got it. Alright. And so I think we're we're shifting to case assist now. Does the model learn from documents attached to earlier submitted cases? So for the okay. So, I suppose it's a question about case classification. So the model learns, only from cases and how those cases are classified. And then we can we the model compares a new case that it hasn't seen before to its training set, and it was gonna help predict how to classify that next case. So at at at this time, we're not looking at what's been attached to those cases or what content is related to those cases. So I'd say no on that front because it's, right now not necessarily relevant to a case classification. It's an interesting question though because, that is the the kind of data that we want to look at to help, determine what is the best content to deflect a case, which is something that we're still, that we want to improve, that we wanna innovate in, create new models for to be able to do that better. And the the, you know, the event, which is attaching a piece of content to a case is a very strong signal that that content helps, for that type of case. So we are exploring that too. Great. Now, is there out of box case assist flow slash component near future? I think I missed the beginning of the question. Could you repeat that? Is there out of box case assist? So I think they're asking for an out of the box, component. Out of the box components for case assist? Yes. So, we actually have a cookbook, for building a case assist experience in Salesforce available on GitHub right now. We can follow-up with the link if you want, but, easy to find on our GitHub account. So that's an example. That's actually the code is is available there, to deploy into your Salesforce environment to set up a a case assist experience. If, if you were looking to do this outside of Salesforce right now, we don't have out of the box components. That is coming, though. We'd like to build it into, you know, our Coveo headless framework as well as, Atomic, which is our UI framework for, any integrations other than in Salesforce. Okay. Thank you for that. Next question. This might be, for, in terms of in the extensions. Is this available as an Edge extension? Okay. Good question. No. It's really a Google Chrome extension right now. And I'm actually not aware if, it's on the road map to consider extensions in in other browsers, but right now it's really a Chrome Chrome extension. Thank you. Alright. Let's see. Would this work on SharePoint? I suppose the question is for IPX. Do we know? I think say, but Yeah. So SharePoint online, I I'm not aware if we've tested it specifically on SharePoint. However, I assume that it is similar to Google Drive, but I am not a hundred percent, sure. Maybe we could follow-up with that person if we have their name, to make sure. Yep. But we can do that. A use case that is is common to be able to have a search in SharePoint. Yeah. Okay. Next question we have, is there any demo, of what the in product experience extension looks like to deploy from a specific URL? I will have to follow-up with that person as well. I'll ask, my colleagues to make sure. I I'm not aware of a demo that we can share. I suppose or expecting, like, like, a video, like, a YouTube video. I think that would be helpful. And if if we have that, we'll send it along. Okay. So we got some clarification on a question. Smart snippets. Are the deep learning models dependent on view events like ER models or just question plus click? Yep. So right now, it actually doesn't learn from any usage analytics data. It's actually only learning from the content the structured content itself. We do, so the planned iterations for this model are are planning to integrate, the interactions with the content and the data that we can extract from there there, sorry, to, kind of add a layer on, on the model. But right now, it's only learning from content. Okay. So that was clarified. Thank you. And, is there any intention to allow more than one snippet? It's a good question. Quite interesting. I don't think I've I've gotten that question before, but it's, an interesting idea. From, from the research that we've done, it seems that the best practice is to have really one answer there. If we wanna share the animation again, I don't know if that's possible. What we do have and offer is we we show the the snippet that is the answer to that person's question. And below it, we have, people also ask, which shows similar queries, or similar searches to what the person searched for. I think, if you if you present that, we should see the oh, there. We see the animation. So below that, we see people also ask. And in this particular case, we just have one other example, but there could be three or four other questions there. And when, when you click on those, it runs the search for those, for those queries as well. So, that might be an interesting kind of alternative to having two featured snippets, and I'd also be curious to hear, you know, the the the use case for having two featured answers for one same question. I think the the idea of having just one featured snippet is to to reduce confusion. So one one question, one answer. That's also, the way, Google does their featured snippets snippets as well. Okay. Can the smart snippets model extract several question answer pairs from a single FAQ like document? Yes. Yes. It can. So it's actually, quite good at at extracting answers from an FAQ document that's structured. Another question that we have here. What kind of language is required to add the extension? To add the Chrome extension? Yes. What kind of language is required? So it's just a it's a an extension that's that's that you add to, to your browser, and that allows you to display, whatever, IPX is associated that you've configured to associate with with your extension. So I don't think it's it has to do with a particular language. It's really kind of a install it and go kind of solution. Okay. Thank you. Alright. Let's see. Regarding the ServiceNow connector, are similar connector upgrades coming for Zendesk? Great question. I'm actually not aware of the roadmap for Zendesk. We'll have to follow-up on that one as well. The so just to maybe to add a a a couple details on that. ServiceNow content as well as Salesforce content, It's, does does require this kind of let's call it a a more advanced, configuration just because Salesforce and ServiceNow orgs are very customizable. Can have, all all sorts of different custom objects in there, custom fields that exist. And so the UI has to be our our UI has to be able to adapt to that and be able to fetch kind of the different configurations for a specific org. I'm not aware if that's the case in Zendesk, but that is the kind of the reasoning of why we we needed that granularity and to be able to show that list of objects that exist in a specific Salesforce, ServiceNow org. And that's why it was built that way, and we want to display that in the UI. I'm actually not aware if if that's something that would be relevant for Zendesk at this point, but, we can follow-up if we have plans for that. Thank you for that. Here's another one. I'm a little confused by hearing the IPX may not be compatible with SharePoint. Is it not compatible from any URL? Right. So my understanding is that yes. However, you know, I'm I'm not familiar with SharePoint. I know that, Salesforce can be tricky, with, their their security. So it would be something that I wanna validate before, you know, right now confirming that that it's supported. Just because I don't have the information, I don't wanna confirm it right off the bat, but we can get that information for you. Okay. And how does authentication take place from Chrome extension to that of Coveo to pull into? Great. These are great questions. I love I love this audience that everybody's really paying attention. So, I'm I'm not sure on, how how we authenticate from, the Chrome extension to Coveo. I'll have to follow-up on that one too. Great. And we do have, we do have a large team working on various pieces of this. So that'll be why Esme doesn't have all the answers for us today. But also, this is why we have these webinars because this is the chance for you to raise your questions and for us to take these questions back to the team. Yeah. Yeah. And it looks like we have one more question, about Snowflake. Do you have any information to provide on when customers can access and leverage this? Snowflake? Yeah. Yep. So, it wasn't something that we covered today. I wouldn't say I'm an expert on our Snowflake road map or what's going on there. But the last I heard is that, in June or this summer anyway, June or July, Snowflake would all customers would be, working on Snowflake. That being said, I would make sure to confirm with, your customer success manager. I believe some customers have a, you know, specific schedule to, follow to make sure that it happens smoothly. Yep. Great. So, what we will do is we'll take the remaining questions and any questions that we weren't able to answer today as far as having a complete answer, and we will follow-up with all of you with those answers and, so, Harine, anything else? Yeah. I think we still have a few more. Just quick question about, smart snippets again. Is there any documentation that you can direct us to which would explain how snippets work under the hood? Yeah. So this, is actually going live, as I mentioned, July fifteenth. So all the documentation will be released at the same time. So how, at at the very minimum, how to set it up, and, a summary of how it works. I'm not aware if there'll be specifically documentation on exactly, like, everything that works under the hood, but, you know, keep an eye out on the on the documentation, which will be coming out the same day. Thank you for that. I saw a question about snippets and IPX as well, which is a good one, an interesting one, I think. Yes. I can take this one for you. So k. Can the snippets be leveraged in an IPX? Yeah. So I'm actually happy to have that question, because I mentioned earlier did I? Yeah. So it's, it's a platform feature. It's a model that can be used by any enterprise customer, and then showed in a Coveo interface. And an in product experience or an IPX is a Coveo interface. So, yes, it would it would work there. What's interesting and and what we want how we wanna develop this is actually to be able to use smart snippets in any kind of Coveo experience, you know, not just, a search experience, but also, for for, recommendations, for, use in a chatbot where, you know, a conversational experience where you're actually expected to ask questions and would be able to be able to extract answers and and reply with, with that answer that we've extracted. And IPX is one of those kind of experiences as well. And speaking of IPX, I did confirm with the product manager for that feature that it is compatible with SharePoint. Awesome. Got our SharePoint answers. Okay. Any other questions? I think we've, we've managed to answer, questions as as they come. We still have a bit more time. Do we need to provide credentials while pulling info using the Chrome extension? Yeah. I think it's similar to the earlier question about authentication. I'm not enough of an expert on the extension to know how that works, at this point in time. Likely that if, if you wanna pull credentials from the system that you're using, like, for example, if you're, I don't know, logged into Google on a Google Drive, and then we could use that to authenticate to Coveo and retrieve documents that only that person has access to. Sounds like something we'd wanna support. I I'm not, sure, however. So let's, follow-up or maybe, have a specific webinar for that. I'm not sure yet. Yep. And I just shared a link to the IPX documentation that we have if you wanna go out and take a look at that. Thank you for that, Bonnie. I have another question here. Any updates on the ART ITD two point o NLP update? Cool. Cool question. So I have a small update. So while that that's something that we're working on, that we would like to release at least, initial kind of beta release before the end of this year. For those who aren't familiar, the this this person has called it ARTI tv two point o, meaning, so a new machine learning model that learns again directly from content to do, to understand similarities between your content and, a case. Right? So what is the content that helps solve the problem that's described in a case? Currently, we do that with using ITD, so intelligent term detection, where we extract what we believe are the most relevant keywords from a case to search for content. And next generation of that is to learn directly from content and have a model kind of understand what is the content that best serves a particular case. So, that's something that we're actually, you know, we have a full team on, right now. We're designing it, trying to figure out how how is how it's gonna work. Is it, going to replace ITD or work alongside it? But, that's something that we're actively working on and running tests currently. So I don't have too much of an update, but we do wanna have something before the end of the year. So something to look forward to, I suppose. Okay. And oh, wait. Thank you so much for thank you so much for the feedback. We really appreciate it. I guess, any, again, any more questions? We still have a bit of time, but we also don't wanna force the time on you. I know everyone, wants some time back as well. I will post the link to the webinar tomorrow in the chat here. So if anyone's interested in attending, there's a link right there. Oh, and, let me, repost this, IPX documentation as well. Hold on for hold on. And here you go, Chris. Oops. Sorry. Nope. Not that document. That is the wrong link. Hold on for that. And this should be the one for the the IPX link. Okay. I think we can all wrap. Oh, there's another there's another question here that just came. Does Coveo query suggest and smart snippets slash people also ask, interact? Does course suggest in smart snippets? Also ask Interact. I'm, I'm not sure. I'm I'm not familiar with Interact. Earlier, I suppose that it was an Internet solution. So if, if you're using Koeo in Interact, I don't see any reason why, query suggestions and and smart snippets wouldn't work there. So, let's say let's let's try it out. I don't I don't right now, I don't see why why it wouldn't work. But does it ask interact? I'm not sure, what you mean by that, Maris. We if Interact has content, we can index it, and we could expose that content as well, you know, alongside all your other content in the search results. So I don't know if that helps to answer the question. I think Maris call clarified it here in the chat and asked, just that they work with each other. Right. Okay. So I'm I'm not, familiar with Interact, but, but I suppose so. And not not much help. But, I would say yes. If if you're using Kubera there already, all all the Kubera features should work, including query suggestions and smart snippets. That's awesome. Okay. Okay. I think we're ready to wrap up. Bonnie, Esme, any any parting messages before we, we get things going? No. I would just say, you know, really excited about these features. And if you have any questions, want more information, you know, reach out to your customer success manager. We'll be following up with you after this to to share more information with you. Yep. I'd say I was I was expecting, you know, Smart Snippets to be, exciting and people to ask questions, but I'm really happy with how many questions we got. Thanks to everybody for interacting and clearly, you know, listening intently. I'm happy to see that this is of interest to you guys as well. So As as we speaking of smart snippets, there's one more question here because it's very exciting. Curry suggests helps smart snip snippets or people also ask get better? Sorry. So will query suggest will query suggestions get better thanks to smart snippets? I would assume so. Miers, can you can you can you, clarify that? Let me see the question. QS helps Smart Snippets people ask. Okay. Yeah. That's actually an interesting question. So, query suggestions are there to help users search for something that, that will lead to success you know, a successful outcome, meaning, you know, finding the results that you need. So those suggestions are based on successful sessions, other users' successful sessions. So if people are using, are asking questions and the smart snippet is helping them, we are picking up on signals like, when the feature is is you know, we we feature the answer and you have to, click on show more to see the full answer. We are gonna be picking up on that signal to know whether, that smart snippet was useful or not. We will be learning from it. And so if, our machine learning is able to determine that that was a successful outcome, those questions that were asked that led to that answer would, show up more often in query suggestions. So, long answer to say, yes. It it would just because of how they they interact together. This is still, you know, it's still a search and click kind of scenario. Just instead of showing results, we're extracting an answer from it. Awesome. Okay. Alright. We okay. Thank you so much for confirming, Mires. We appreciate that. Okay. I think we can wrap up. Thank you so much, Esme, Bonnie. The rapid fire of questions was amazing. Yeah. That was great. We actually appreciate it. If you think we're, like, getting overwhelmed with questions, we're not. The more questions, the better because we're learning more from you as we go. Thank you so much for attending today. As we've mentioned, we will be sending a recording after the presentation. And, also, at the end, there will be a survey that will pop up at the end of, at your screen. So please feel free to share your feedback and, share your contact details as well if there's anything that you'd like us to follow-up on. Again, on behalf of myself, Bonnie, Esme, and the rest of the Coveo team, thank you so much for attending today's new and Coveo webinar. We hope to see you again at tomorrow's event, the Salesforce event, and at our upcoming events this summer. Have a great rest of your day, everybody. Bye. Thanks. Bye.