Okay. So, welcome, everyone. Today's session is all about analyzing and optimizing and how to conduct a health check, in the theme of self-service. So before we get into our agenda today, just wanted to introduce the team. So, I'm Vanessa. I'm the CX program manager here at Coveo. So I'm mostly you'll see me around more in the Coveo Connect community, which I'll talk about closer towards the end. And our host today is Jazz, who's a senior customer success manager here at Coveo, and Hamid, who is a manager of customer success. So we have bit of customer success and community to kind of title together in that theme of, self-service, success or self-service tools. So on the agenda today, we're really gonna go into the purpose of a health check report, why it's used, what are the metrics that you'll find there, and we'll tie that back to the business outcomes. So, how you can actually define and understand your business outcomes and how that actually relates to the health check reporting. And then we'll go into a short live, demo in the Coville platform. And if we have the time, we'll do a q and a. And just before we get into the, session, I did wanna just mention you can always pop your questions in the chat, sorry, in the q and a or ask a comment in the chat as well. And I'm just about to launch a quick poll to get us started, so you should see it on your screen. So I've just launched it, and you should see it pop up. So we'll just give it about thirty seconds. It's two quick questions just to get an understanding of, how much you use the Coville admin console and your level of understanding of Okay. I think we have almost all the respondents. I'll just give it a few more seconds. Okay. Great. I think we have almost every response. So, Hamid, if you wanna I'll end the poll, and you guys can get started. Awesome. Thank you for that. We will start with Jazz. Talk to us a bit about the health check report, and then I'll take it from I'll take it from there. Probably a lot of you have seen health checks being sent out to you directly by Jazz. She's she's, you know, she released the the the health check initiative, from our end, and she's she, you know, grabs all the details, and she wants to share some pointers with you before I go more. I'll try to make it as as much as as, you know, interactive, and I'll show you where to get things. But, Jess, go for it. Please tell us more about your health check reporting knowledge. Thank you. So, yeah, great point to say AI. You know, I'm pretty sure everyone knows what a health check report is, where you've probably done it with your CSM or me or Hamed at some point, you know, during your, journey with Kaveo. But if we go to the next slide, I do want to talk a little bit more about, you know, what is the purpose of a health check? Why do we do it, and what do we get from it? So first of all, let's take a step back, and look at, you know, what is the ultimate goal for everyone who's on the call today. Right? The ultimate goal for us would be we want to provide the best search experience to our end users. And that's where how do we leverage AI check reports to get to that goal, right, to provide that ultimate best experience? With the health check report, we have three main elements, that you can leverage in order to get to that particular goal. The first one being AI into your user behavior. So a health check report will guide you and give you really the details on how your end users are interacting with your search solution. What are they clicking on? What are they searching? How many clicks do they need to do to get to the result that they want? Is there a pattern? Is it a positive pattern, negative pattern? What is successful? What is less successful? To give you an example, let's say you've been working the last year, right, the whole twelve months. You've worked on different campaigns, on your search solution or your search page for your end users. And you want to understand what was successful, what was, like, successful, which campaign did the best, which one did less, you know, which one was less popular with the end users so you can really understand what they're doing, what you, in terms of effort, where do you need to put the efforts, and then really cater to their needs. Right? Because at the end of the day, all we care about is their experience. So that is a type of information that you can get from, understanding your user behavior, and that ties to optimizing your search experience. So you not only understand what how your user is behaving with your solution, but on top of it, the other information you get from a health check report is, you know, what is the content that is the most con consumed? What articles they're clicking on? Is it articles? Is it blogs? Is it news, articles? What type of things that they're looking at? Is it relevant? Like, let's say you're putting so much energy into pushing one type of content, but is it relevant to your ad users? Is there any con content gap? This means that, hey. My users are looking for x AI z, but it seems like I don't have enough content that exists related to that x y z search. So that's a content gap. How can I improve that? So that's where you take every information that you get related to that type of experience for your search, experience and cater to the users. So you take that and you work on it to remove the content gaps. So that's fairly easy. I think most of you have heard how to do this from us. Fairly easy, but it is available in the health check report. You have, you know, your search relevance. You is it relevant? Do I need to focus more on, you know, topic a? Do I need to focus more on topic b? So that's how you can optimize your search based on the metrics that you see in the health check report. And to give you an example, right, let's take a a college use case. I think pretty much everyone here has been on a Cal College website at some point on in their AI, and we're looking for different programs. And let's say I'm a college, admin, and I'm doing a health check report. And suddenly, I see that, you know, this one program that was really, really not popular is getting more traction. So I see that people are looking for it more. Search are related to that. I see that my gap has an impact on it because users are searching for something related to that particular program, but not finding it. So it ends up in my content gap. That type of information is very valuable for you because, let's say, you always you were focusing on, you know, program a and b because they're always the most popular ones. Right? It's been four years. These are the top programs. So you have your team focusing on promoting those two programs, but the health check report will allow you to see, hey. There's a trend that is a a new trend is coming up. Things are changing. You never know with Agentic, and I don't know what the other new generation is called. I think alpha. So you never know how they will react, but health check report will guide you into seeing what the trends are looking for looking, at or looking AI, and you can focus your own energy on that. That brings me to my next point. Take data driven decisions. That way you can do with the health check report. So, again, same with the college, you know, kind of use case where you were focusing your team's energy on program a and b, and now you realize, hey. Program y z that was not popular needs more attention, so you can take better decision into splitting possibly your team, having one team member focusing on creating a little more content on that, you know, little program that was not popular. You create more content. You, push a little bit more on promoting it because people are looking for it. So these are all the type of information that you can get from a health check that will allow you to AI, really, your search experience for your end users and, at the end of the day, make them happy. So how do we make them happy is with optimizing the search experience. And if you have business outcomes, Hamid's gonna talk about it. They should be tied together. Right? Those business outcomes and, you know, how you want to improve your search experience. And it is gonna explain a little bit more about why they're important. And if you have, like, some questions, like, feel free to drop them. I'll be in the chat with you guys. Absolutely. Yeah. I was gonna say if there's any question, Vanessa, Jazz, if and if anybody wants to actually, participate, please let us know. We can we can allow you to we can have a discussion. Please do that. And we can answer questions through q and a as well. Perfect. There's none so far, but I'll keep my eye out. Awesome. So, I really wanted to put this in in in our deck and talk about it because it's really important. Sometimes we focus. We say, oh, I want a better customer experience. Perfect. But then what? Right? We really, really, really have to always think about what are our business outcomes in general. Commerce, sometimes it's super straightforward. I care about increasing my con my conversion rate, increasing my add to carts. So I have why do I want this? Because we want the revenue of the company to increase by x percentage. How can we help this? Through conversion rates, through, what you wanna call it, add to carts, through more sales. Those increase. Right? Revenue increase. The company's goal increase. So even though we have different personas on this call, some personas are really more on the technical side of things, hands on. They do small optimizations. Some personas are more on their higher level where let's look at the business AI, see what we are. But I think we really need to work altogether because sometimes, unfortunately, even if I have, let's say, a click through, which I'm gonna talk about it in a minute, a super high click through. But guess what? It's really not positively impacting my business outcome. So it's very important when you perform those health checks after you have, you know, the tactical information in it, take it back to see, okay. It's actually positively or impacting our business outcomes or not. I put here some of the main business outcomes we work with with our customers. If I take let's let me start with the the example that that Jas talked about. For example, AI search. It's called website search for college. Right? Or website search for lawyers or stuff like people search, essentially. If I start with people search, perfect. Our click through rate, if I'm search for if I'm searching for Kathy, I'm getting the result. Perfect. I see there, but my click through rate is not that high. Click through rate means when I click on a result. But I do see my, you know, part of my other metrics that I measure is my conversion rate. So people are actually taking the phone and calling Kathy as a lawyer. Right? And they're booking appointments with Kathy, which is contributing to increased conversion, which is phone, and then from phone conversion probably to become a a client. Right? So it's that's why it's really important in this use case to know that what I'm trying to measure is how can I increase interaction from the site to convert it into an actual lead or a or a client? That's one example of business outcomes when it comes to people search. Website search also, if there's no transactional aspect to it, so if there's no add to cart AI commerce, what are the other examples we we can do? Sometimes when I search for something and then and then there is a place where I can say download brochure, let's say. Right? Or click here to send an email to the representative. So those are all conversions that you probably have in your CRM system, or whatnot, Salesforce, etcetera. These are really important to say, okay. We increase the leads by five percent. Right? This five percent increase in leads with one percent of those five percent converted, we increase the revenue for the customer. That's another example for website. And to be completely transparent with you, there's always gonna be instances where the purpose of the website is just AI have good search. Right? But let's challenge the status quo. I have good search. I have good presence online. What does it mean? Is it helping the business? Really, really important. Now there's an easier use case. It's not easy, but the way to measure it is easy as self serve. That one, we can measure a big portion of it through Coveo. Meaning, irrespective of click through rates, blah blah blah blah, am I having an increase in my case deflection? Right? Are customers able to self serve? Because this is directly linked to cost saving. If I'm able to self serve, I don't open ticket with my agent, with any agent. If the agent, doesn't work on it, they can work on something else, and every case has a cost. If the cost per case is twenty dollars, I was able to explicitly deflect two hundred cases. That's an exact number that I can take, and I can report to to their business saying, look. We're saving money, which adding to our business outcomes. And, also, with agent efficiency, for instance, how long does it take every agent? What's the average case handle time? Is it five hours? Perfect. With Coveo AI panel, which is the side panel that gives you relevant results, they are able to solve their cases in three hours. Two hours case handling time decrease that also cost money because agents are employees. Employees, you know, have salaries and and whatnot. So all this to say, please always link whatever you find in the health check to a business outcome because sometimes it doesn't really mean if you are excellent in click through rates, zero content gap, not all the time it means I'm contributing to the business outcome. Maybe I need to invest in a different part of the AI, right, or in a different place where, AI, Jan said, the new course. Maybe it gives it gave me that. The general numbers are fine, but this specific course can generate more revenue for us. So let's concentrate on it. Is there any questions or comments around this? Yep. There's one that was a private question, and I'll just read it. Yeah. So AI your example of the brochure clicks, we do have something similar with a download button. Is it possible to track these type of actions in Coveo? Great question. So, yes, we could, but it's not out of the box. So out of the box, Coveo, out of the box, if we're using the Coveo frameworks at Atomic, Anciently JS UI, we usually track the click on the hyperlink, right, which takes you to the article itself. All those extra stuff AI download brochure, download PDF, inquire, it would be a very simple custom click that we would send to Coveo. So we are able to track it, but it needs a bit of technical work where we add it in the result template. That's that that's a good question. If you need more details about it, by the way, let us know. Just ping us, and we can send you some documentation on it. Awesome. Now let me show you a bit of the, how we go to the reports in a second. But before that, we call this the the our actual health check, report template. I will be sending this to you after the call. We'll tell you how we're gonna send them and whatnot. But you can ignore the the the the slides before and maybe the slides after, But this can be your placeholder template as you build that report, and we can we're we're gonna put indicators where to get everything, which I'm gonna show you in a in a bit. Couple of things quickly quickly before I jump into it, like mon monthly consumption. Where do you find this? Right? That's how you build it. Take a screenshot and put it so you can present it to to your teams. We can, you know, you can, we can go through this fast AI I'm gonna show you live. And then what does it mean? What are our four main metrics? And then how were they evolving year after year, quarter over quarter. So we'll make sure we send you this so you can easily fill it out. We'll send it as PDF sorry, as PowerPoint so you can use the same thing and and create it. And they can also use it going forward, frequently whenever you do a health check. Do we have any questions? I saw a few things in the chat. No. It was just me. I was posting a link for documentation. Cool. Cool. We're good. Alright. Now let's do this. If you have access to to your org, which I hope you do. And if you don't, it's fine. Let us know also through chat. We'll we'll make sure you have access, or we or, basically, you can speak to an admin, who owns Coveo from your end. We can invite you. If you have access, it's great because we can do this together, walk through together. Some of you will already know this, so we can walk through through it fast. And if you don't, just, you know, follow my steps. Once you land on your on your on your page, you're gonna see a lot of things going on on the left hand side. You have the sources, fees, the service start part, machine learning, search, etcetera. Today, what we when we populate that health check, we rely on only few places. Right? One, the reporting section, and then, we also rely on the settings under organization to actually view consumption, which I which where where I start. So I'll just give you maybe ten, twenty second to log in to your org if you want and follow through with me. And the org would essentially be platform dot cloud dot caveo dot com, and then it will take you to the org to have access to it. You have production and non prod as well. Alright. So under organization, let's talk about consumption. If I go under settings and under settings, you have the consumption tab under license. I'm using our Barca. Barca is our demo. We call it Barca. It's a we created a an imaginary company, but it has the full platform from a to z, including support, commerce, and website search. Why is it important to go here? First of all, what you see in consumption is a bit different than what I'm gonna show you in the in the reports. Consumption is the pure usage of Coveo irrespective of where it's come when it's coming from, why is it there. It's not analytical per se. It's more of think about it in in the technical terms, the number of APIs that are hitting the Coveo index. As simple as that. It's important because let's let's say we are looking at I'm I'm gonna look at this for as an example, Coveo for Salesforce experience cloud, which essentially, if you have a community, it will be it will be built in a Salesforce experience cloud. It will show you your limit, right, which is the license limit. In this case, we have a hundred k limit, and my usage so far in March is eighteen point eight k. What does this tell me? It tells me I actually have room to grow. Right? I have eighty k more room to grow. And by room to grow, it doesn't mean just having more people go to that specific, search hub or site. It actually means I can deploy more for Salesforce experience cloud, pages and use them otherwise because there's still quite a bit of room. If you are over, it also tells you, listen. We are over, so let's look at it more. Are we over because there's more traffic, which is more conversion, hopefully, more business outcome, which is amazing? Or are we over because there's an issue with the system? There's duplicate queries. Right? So it's a kinda an indication for you to compare with your organic traffic that goes to the site versus the percentage of traffic that are actually performing search. High level rule, if it's a website search, community search, I would say for every one visit to your site, probably there's three or four searches on average. So you can also do the math saying, oh, you know what? I have AI have, like, two hundred thousand visitors to my site per month, but I only have twenty thousand searches. What's wrong? Is the search box not in the right place? So all these kind of things are very, very important to get from from this page that I'm showing you. Now there's another thing I'm gonna highlight very high level because we're really gonna do multiple sessions only on that aspect. Sometimes you are gonna find that you have what we call unassigned queries. What are unassigned queries? Unassigned queries are what you see here. The AI Agentic says queries only. If that's the case and you're walking through with me and you find an assigned query, you can sim and you have an actual search hub here. If you click here, you will be able to attribute it to the right products you bought. Again, because it's demo, we have the whole suite. But in your case, you're probably gonna see one or two of those depending what products you bought. So make sure, for example, in my barca shop, I know for a fact this is an actual my my Caveo, commerce SKU. So once you add it here, it will be attributed to the right queries on top. You AI change it because I'm pretty sure they kept it here like this for a reason. That's one option. On the other hand, if you see something like let me see if we have it here in the search. We don't have that. But if you have an actual search hub name here that says unassigned, we have a bit of a bigger problem. It means even if you change the entitlement metric here, it's not gonna go anywhere because it's we don't have a search hub attributed to those number of x squares or APIs. Why is this not good? Not only because we cannot track it, it's it also means it's not going to analytics. So I can't see the details. I can't create a health check, essentially, if all your if all your unassigned query of all your queries are unassigned. Also, machine learning cannot learn. So really important to work on this, and we can help you if if you have that. Now if one percent, two percent have, unassigned, I wouldn't worry about it as much as if all of them are on a site. Right? So also really important to take that into cons into consideration. And in case for the technical folks on the call, to give you a bit more context where we've seen AI, any sidecore environment, when you create your search experience, there's a place where to have to have to specify the search app. If by any chance you did not plug in the name, it will have it as unassigned, so you have to change the configuration in your conveyor for Sitecore and give every page slash search hub a name. Another example that I've seen it, if you have server side queries, meaning queries that are directly on the server only being API tech calling APIs directly, hard coding APIs to call Coveo, most probably, it's unassigned because we are not configuring get properly. Anyway, I'll if there are any questions or comments, Vanessa, Jas, or we're good? One, I have just to, for everyone, because some people might not see everything that you're seeing on the top of your screen. So all the data, like, you know, total unassigned queries, query for SAP, query for ServiceNow. So I just want to mention that it is normal that you don't see all that information and that you have have less. What you see on your administration console is based on what is included in your package. So it's normal. Yeah. Really good call out. Yeah. This, yeah, unless you unless AI which I hope you use all the Cover Your Suite, you're not gonna see all those here. Awesome. Thank you, Jess, for the clarification. Alright. Now that we looked a bit on consumption and it's gonna help us with, populating the the the first consumption slide that I showed you, it's also important to to report that. Now let me go to the reports where we will start to fill out, fill out more, additional info there. So I already plugged in to the two reports that I'm gonna use. But if you're if you're walking with me, in your org, if you go to add and go that go to add a dashboard from template, let's start with the all interfaces dashboard and just select that template. It will populate a report. So I already have it, so I won't save it. But I have the old interface, report health check for the webinar here. Right? So why am I looking at one card only? What does it tell me? Do I have to copy all this in my report? No. Most often, you don't you not you don't only have one search hub. By search hub, it's a search experience on the front end. Example, you might have regular search, people search, location search. You might even have a block search or whatnot, and each one of those is built on a different UI. Right? Because it has its own machine learning. It has its own filters, etcetera. So when we when you wanna do your health check, very important to decide. Do I want to look at the overall view of all the search? How is search doing? Or do you wanna focus on, let's say, one which has the most searches? Example, my main if if my search in the in the header ends in my main search, I really wanna concentrate on that one first because it has the most traffic, and then I look at the blocks. Right? So or you can do you can do it in a combination, but I I usually we usually recommend look at your search hub with the most traffic, concentrate on this, create your baseline there, and then do the other stuff. How do you know which search hub name or if the search hub are name and search hubs are named properly? Through this report called the all in search interfaces. It shows you all the origin ones, which are pages that have. Here, again, this is a demo, so we have a lot of those. But when you sort by manual search event counts, so the organic searches that people put in the search box, I can really easily see that listen. Our regular website search for barca and our search for the in the sports section have the most manual searches. So I really wanna look more into, this search hub to see what kind of information I can get so I can populate a health check only for this. And then do I do I care about the other ones? Hundred percent. Like these four, the engineering and the shop, which is our our, our demo shop, also have a lot of traffic, but not really care about the support block Agentic stuff. Not really. Right? Because it's again, this isn't test. But if it's if it's not as important, I'm still gonna look at it, but focus on the ones that have, definitely more traffic. And that's what the report, is telling you. An extra tip. When you are here, if you go to edit and then edit this card, there's one really nice thing to understand is where is the search hub? Which pipeline is pointing to? Pipelines is another concept we'll talk about, but it's essentially it's the rules that are put in one place called the pipeline where this search hub applies all the source before giving back to to the to the customer. So I'm just gonna simply add the query pipeline as my dimension. Right? And now let me just make sure I save. When I save, I know that oh, right. So my search port goes to my query AI, which is CMH search sports. So if I wanna do any optimization, I know that I'm gonna have to go to this pipeline if I wanna optimize anything in my search sports. For the shop, it's this one. For the search, we have it on parts, on engineering, blah blah blah. So very, very important, or else you're not gonna know where to do optimization. Right? Like, where do I add it to source entry? This report tells you that. Once you add it, make sure you save, or when you close it, it will not save that new dimension you added. Awesome. That's the first report. Second report, same thing when you go to add. Well, before I go there, you might already see many reports in your in your org, by the way. It it doesn't thread that it's empty because once once we create an actual org and deliver it to you, you by default, you're gonna have four reports at least. And then you work with it. Other people work with it. People who are no longer at the company work with it. A CSM helped you out at at one point, created a report for you. So there's a bunch of reports. All of them are valuable, but what we did specifically for health checks and to make it easier for you, we created one view for all of those, which you can upload into your org again as we speak or after. So go to add, dashboard from template again. And then if you scroll down to the advanced section, there's a report called search performance report. Select the template. Uh-huh. Why did we talk about search hubs? Because it will the report will ask you, do you want to look at all search hubs, or do you want to look at one search hub? As I said, let's say we're looking specifically on the search search hub, which is the website search for Coveo. I'm gonna say search, add my report, and guess what? Automatically, it will be added as a filter. Now what if I made a mistake, or what if I wanna actually look at all the search hubs, not just the search? Really simple. Either change it here, add more search hubs, or go here, disable. It's gone. Even more, remove. So the option it was at the beginning, it doesn't stick forever. You can change it as, as you need. I already have a report created because they don't let us build a lot of reports because they're being used a lot. So the search performance report, double click on it, and here we go. So this is our master report to use for, health check. A lot of these might be familiar for you from other reports, but it's more incorporated with everything in it. So now, one section that Jas mentioned at the beginning okay. Tell me those insights. What's what's going on in search? Like, how many people are searching? How many, how many searches every person is doing? Right? Where where are the searches is coming from? If you are global, you can we we have it here per country. If you are more local, you can instead of having it country, you can edit the report and change the dimension to city. Right? So all these are very important for me to to to also, to report, especially if you're launching new markets. Okay. We launched a new market, but how is it comparable to to my existing main market? In this case, United States versus Canada. That's one thing. Another really something that I really like here, visits per device. With the with the with the ever increasing adoption of smartphones and tablets, we are seeing a shift. In this case, again, it's demo. But you're gonna start to see a shift that, oh, you know what? Personal computer used to be ninety percent. It's becoming it's becoming more than eighty percent because smartphones are getting ten percent of the traffic now. So am I optimized? Is search optimized to be used on the phone? Is it or do I need to change some stuff in my UI? Right? Also, if we have enough traffic from my smartphone, is it time to think to start thinking about an application? Right? So these are just examples I'm gonna give you the things that we pick up from the house check that can help us take decisions, essentially. This graph here, I really like it because it shows now we're looking per day. I'm I'm only looking at the last, week in March because we have a lot of traffic, so I don't want to be, overwhelming. But you can do it, let's say, per quarter and look at month over month because there's always patterns in any industry. A a college, you're gonna see a surge in searches just before registration when when when when they start to come into college searching for stuff, and then it's gonna drop down in the summer a bit. So monitor those and see where and when you wanna invest more and when you wanna instead of looking at the report every quarter, let me look at a report every week or every two weeks during that peak time to catch any issues or to catch any AI. So that's also really important. Also, if I'm seeing something as an anomaly example, if for any reason I saw, oh, I have more clicks than searches, how is it even possible? Right? So that's another thing that you we it could it could signal an an implementation issue. So always look at the search traffic over time. I'll take a quick pause and ask if there's any questions, Vanessa or Jess, as we're good in chat. Like, maybe you could just show it. Divya asked if you could just repeat how to show the actual, like, where you got the template from to add it. Absolutely. Absolutely. Yeah. So if you go to reports and then you go to add, You're gonna go and add dashboard from template. Then scroll down a bit. You will get this template as advanced, and just hit select template and choose your search hub. Also, if you are doing this and you're not saying, you're not allowed, this is grayed out, then you probably don't have enough permission because permissions in Covello are broken down between admin, analytics viewer, content manager, analytics manager. Each has its own permissions, obviously, for security reasons. An admin can see and do everything. Analytics manager will be able to add reports. Analytics viewer, probably not. So if you are having this issue, let the admin know. If you don't know who the admin, let us know. We can go into the org. We'll look at the members, and then we'll see who's an admin who can add you. And if the admin, by any chance, is no longer there, let us know. We have other ways where where we can where we can add you. Good question. Awesome. Let me go back to the search performance report that I was showing. Again, I'm still on the summary tab. Like, all all literally, we picked up a lot of insights only from the summary tab, and we we're not even halfway through it. Now this is my my best part in the summary tab. So what are the top queries? This is gonna show you what's trending in that time frame. Right? Solar panel, hundred and fifty watts, radio, life jacket, blah, blah, blah. All these have a lot of searches, and you can also see what's what the themes that are domin dominating here. Example, these are the top queries. These are the top keywords. So although solar panel is only here one time, like panel in general, but it's also on top as a keyword. Keyword is a subset of the query. So there's a lot of searches around panels, which means there's probably searches around solar panel that are different, not hundred and fifty watts. It's probably different. What is this telling me? Okay. We need to do an investment there. By investment, doesn't have to be AI need to bring more solar panels. It's more of, is there is there enough content and guidance for customers around solar panels? These AI the kind of insights we get. In addition perfect. So these are the searches, but what are the we call them documents that are clicked on. By documents, it could be a PDF. It could be anything that's clickable. So don't think about it as an actual document. It could be a PDF. It could be a product. We just label them as documents in Coveo because I don't know why. But you can rename rename also this template. You can say top clicks, top clicked articles, top clicked depending what you have. Very important for me. So these articles are doing really, really well. Right? And being clicked on a lot. Something I really like, if you want to know what are the searches that are triggering this click on alarm flare gun, simply click on it, and then you can see that, oh, guess what? This is the top query is flare gun, and all this all the clicks are are are are on that from that one. It could be multiple queries, but in this case, it's just one. So I'm gonna remove remove this one. So in the health check report coming always going back to the original goal, I would always mention top five, top ten depending on the traffic searches. I would present it to stakeholders because if there's something that doesn't make sense, they're gonna say, oh, that's new. I didn't know people search for that. Right? And it's gonna trigger a conversation. It's gonna trigger a decision making either. Last but not least on this page, top sources clicks. So Coveo has federated search, federated place to crawl content. Content could be from, from a commerce source, push source, Salesforce AI or website, all all the out of the box connectors and none of the out of the box connectors we have. What this tells me is which source is used the most and which source is not really. So do I really need to keep it, or do I wanna optimize it? Example, the part search the part source is used the most. So even though I'm on the website search, most of the clicks are on parts. Right? Again, it's now it's telling me, oh, customers who are on this are really more inclined to look for parts. Right? Or more inclined to look for support, maybe. Way less though. Right? So that's also important. And then corporate, not even used at all. So what's the value of having it? Is it just taking real estate and indexed items, or do I need to optimize it? So that's another thing that we can take and incorporate it in our health check. Awesome. Now that we did health check, you have the knowledge AI if you want. So what I'm talking about right now, it gives you details and links to learn more about it. For example, I didn't mention the difference between unique visits and unique user ID. Right? Or click event and search click event and all this stuff and search Agentic. Sorry. You can you have all those details there. And, also, you can go to our community, search for search event, and it will give you all the details. Nonetheless, we talked about the summary quantitative data. Let's talk about the qualitative data that will populate the other portion of our health check. Alright. Coveo has when it comes to search specific AI of the self serve stuff, of the the the commerce, add to carts, and stuff like that, we have these. These are really, really important metrics to tell you if how search is doing. Again, you have to link it to the outcomes, but it gives you an indication. The first metric here is called the visit click through. Meaning, I went into the site. I started to search. Right? I searched for one one the moment you start searching or you load the interface of search, not the main page, really important because this visitor does not mean visitor of the site. It's visitor to Coveo, so interacting with Coveo. Once you start to search, your visit starts. I can do one, two, three, twelve thousand searches during my visit. As long as I click once, we consider it as a successful visit, and you are at a hundred percent visit click through. What does this mean in in this case? It means with every ten visits, we have almost everybody's clicking at least once. Again, unrealistic demo, but that's AI the narrative based on the number of visits visits with the cover, how many clicks we have. Now this is good. It tells me there's interaction, but it's not good enough. I need to know how many how many searches are getting clicks. So let's say we have ten people as I said here. Those ten ten people did a total of hundred searches. One one person did twenty, one person did five, blah blah blah. This number here will give you the average of the search event click through, which means for every hundred searches, we have forty two clicks. Makes sense because every visit, you're gonna have multiple searches, right, for many reasons. Maybe I'm searching for something else, or maybe I didn't find what I want right away. So expected this number to be lower. Now I'm gonna mention one thing before I go into these two guys here. So, what are the benchmarks? Because if you're presenting this, you need some benchmarks. AI, ninety six is very good. Forty two, is it good? We have benchmarks based on our customers, which we can share with you, and it varies. It varies by industry. It varies by use case. I'll give you an example. If we if we're talking about an intranet, your visit click through benchmark, best in class is gonna be very high because everybody who is on the intro on the actual search of the Internet, they're gonna search and they're gonna click and they AI we know exactly the content and we control it to optimize it. Right? While if I am more on a general mom and pop shop site, the odds are AI I'm not gonna click. I'm gonna just look in general and and browse. Right? Or if I'm on commerce, right, this number is gonna be lower because in commerce, there's a lot of browsing going on. And then if I go back to my example about people search, if I can get all the information I want from the result templates, there's no reason for me to click. So this will be low. Thus, our benchmarks or best in class is different. We can help you with that. We have this info. So, you know, if if you have any questions, let us know. We'll let you know the best in class so you can compare properly with with your, with your system. But if I give you a bracket of of of best in class, the physical click through, I would say, goes from fifty up to sixty five percent, seventy percent. The seventy plus percent would be more Internet stuff like that. The lower the lower stuff would be more website use cases. The search event click through could go from thirty, thirty five up to fifty as well. Why fifty? Think about it this way. Whenever someone goes on the AI, they're gonna search at least two times before a click, which is fifty percent. Right? So that's why fifty is is still a very good number. But, again, we can share all those details based on LOB and industry for you for you. I will let you know if yours your industry is very specific and we don't have enough statistical information. We'll tell you the average, but it's not statistically significant for you to use it as benchmark. But at least we say, we have ten customers. The best is this and the worst is this so you can know where you stand. The following two metrics, also important. Average click rank. Average click rank tells us an early indication how is relevancy working and how is machine learning working. So means when I search for something, am I clicking on the result that's number one, two, three, four, five, or ten, or am I going to the third page? Because the third page would be probably the click rank would be twenty seven. Right? So this is a good indication that relevance is good and machine learning is boosting content slowly but surely to the top. So, ideally, the less this number is, the better. Again, there's benchmarks, which we will share with you. Commerce, the average click rank is way higher because people browse, and and even the result templates are not lines. They're usually, grids grid view, something like that. So you have maybe five in one line. So the the moment you go to the third line, you're already at fifteen. Right? So that's really important. And, now the most important, in my opinion, is the content gap. Content gap, what does it mean? It means I search for something, and I got zero results found. No no in search. Content gaps sometimes are justified as in somebody really just goes on the AI and just clicks on whatever. Like, on the keyboard, hits enter, there's no content. But if it's a legit search, and I know for a fact it should have content in my system, that's not good. Right? So I need to either create the content or maybe there's an issue. The content is not going to cover you. Right? Or maybe I just rebranded. Right? I had an I had a product that's named that, you know, I we used to everybody used to call it a b c. Now people are calling it y y z. So but people are still searching for a b c. So in that case, I can do something about it in the in the query AI, and I can say, listen. Whenever somebody search for ABC, add a thesaurus entry to also search for AI. I solved my problem. So this is very important to report. In this case, we really want it to be less than five percent. Right, irrespective of industry. Less than ten percent is okay. It's less than five percent is ideal. Those five percent would be the super bogus stuff that machine learning or anything did you mean nothing can catch them. Right? This report obviously does not tell you what those content gaps is, but we have uploaded this here, which is content gap and query performance, which shows you the exact user queries with no content. And if they had a facet clicked. Because sometimes if there's a facet clicked, it's not really a content gap, but I'm looking for a Samsung TV, and I actually have the Apple brand clicked. So I'm not gonna find results. Right? So these are the AI of things you can you guys can can can look at as well. I'll take another pause and see if everything goes good with chat or if anybody have any questions. I think we're good. You have ten minutes left. Beautiful. Thank you. Now let's look at the last tab that I that wanna cover and the last tab that we use for the health check. All the other tabs are very helpful, by the way. But for the health check specifically, we use summary health check, obviously, content gap, and machine learning. Why is it important? Kaverio's power or Kaverio's differentiator is the the it's the the the the the capabilities in machine learning in AI. There's many AI, models that we have depending on the use case. For this specific health check, I'm gonna concentrate on that most the top two famous ones. I AI mention it, but the the third one. The top two most famous ones are one, automatic relevance tuning, where machine learning learns from successful user journeys, and it will add scoring to certain items to push them more on top. Thus, they get more clicks because, it's more relevant to the customer. How machine learning works, that's a whole science. You know? We can we can send you some documentation about it if you're not familiar. But in the highest level, it will take whatever it's famous based on the similar user behavior, similar old signals, and boost content on top. So now okay. That's great. But how do we measure if machine learning is helping outside of those four big metrics? Couple of things. Number one, we wanna know, is it is machine learning learning a lot and influencing a lot? This here, this box here tells you the percentage of visits with a click on ART. The higher the number, the better. It means the more machine learning is learning. Right? In this case, in the last four days, what this box is telling me that fifty percent of visits had at least one click on ART, which is automatic relevance tuning. By click on ART, we mean a result that's boosted by ART. Extremely healthy. Right? And this if if your content is a bit static, doesn't change much, this number will be way higher because I know we don't update the content as much. So whatever machine learning learned, it it learned it and it just AI it. But if your content changes, this might fluctuate a bit. Right? It might go slower. But at one point, you're gonna see that most of the sir of the visits have a click on ART, and that's really, really important. And it shows you how machine learning is impacting, is impacting the user journey. Now remember the click rank we talked about? It was five point five in the earlier screen. This these two these two here show me that okay. So if machine learning did not learn, did not learn yet a query, where is it where is the result that's clicked on? It's at seven point seven, the average click train. While when it actually learned it, makes sense, it's at two point eight. So that's a direct ROI saying the results are higher on top. The relevant result are higher on top because that's machine learning, thus getting clicks because of machine learning. Difference is massive. Between seven point eight and two point eight, this one is number three. This one is number eight. Right? So that's a very important aspect to look at when it comes to ART. Chain switching gears. Numbers here are not gonna make sense, obviously, because it's a demo. The same concept applies to query suggestions. Query suggestions, some people call it type ahead, powered by machine learning. It's not static. It's not based on a rule you put in. It's when I start to type very similar to go to to to Google. When you start to type in Coveo, c o v, it will show you Coveo machine learning, Coveo CRGA, cover, etcetera, cover letters, stuff like this. So all these recommendations under the search box are query suggestions. What the report tells us, are end users more inclined to click on a, to click on a result if it was actually chosen chosen as as a query suggestion or not. So, again, a direct, ROI on on on query suggestion. And we have the same concept of the average click rank. In this case, again, demo, it the number is amazing, but it's not fully we can't fully rely on it. But it's saying when it's a query suggestion, it's an actual number one that's being clicked on, while when it's not, it's five point five. Right? So that's another really good thing to look at, and you can see the queries that are suggested with what are they, etcetera, and you can report over this. The last machine learning model, I'll do it in one second. Recommendations content recommendations. We have another report for that, but I know it's not used a lot. So we we we have it different we can do a different session on it. But I wanna also report on recommendation, meaning if somebody looked at this article, there's this article too. Is it getting clicks? This recommendation model is powered by COVID as well. We have a question. Yeah. Go ahead, Jas. Can we add a section for RDA here, or this tab machine learning is only showing ART and, query suggest model Great question. The default template on the ART and query suggestions, CRGA, recommendations, DNE, all these AI of stuff have to be a separate report. Many reasons because we we measure c r the signals that we measure for the AI are completely different. Right? So it won't fit here unless we make it extremely customizable, and it's gonna take more effort. So we will share links to CRGA templates, which you can add that are specific to to CRGA. Same thing, by the way, for smart snippets, which is, another one that we can look at. Alright. Any other questions or or or comments? AI think You're good. Now, so that's it for the health check. Again, we'll send all these details. I hope hope it was really helpful to walk through. Please add the dashboard performance, search performance and relevance, And Vanessa will tell us now a bit how to be able to help find stuff yourself. There's really exciting things going on in the community we'll wanna talk to you about, and then we'll wrap it up. Perfect. Thanks. And I know Jazz left a comment as well. If you don't see machine learning in your Cobio platform page, you can always reach out to Hamid and Jazz using that email. So that's maybe just, copy paste that one. Alright. So in terms of, you know, where can I find these resources that we spoke about today and just in general, if you don't have access to our community, that's really our hub of where we're, sharing all these resources? We have, so I'm sharing my screen now, and, basically, this is my personalized, home page. So if you're in there, some of the new features, which I'll show you, in the last few minutes here, if you're commenting and posting in the groups, for example, you'll get points, and trophies. So there's a lot of fun, new AI updates there. If you wanted to search up, let's say, you know, Hamid just talked about, ART, you can do a search, and you'll get taken to, our knowledge base, so our technical documentations, any training. So if you wanted to learn more about reporting, for example, you can find all of, this content here. So the most exciting update that we wanted to share today is within our Coveo Connect groups. So we've just launched this new, customer success community group, and we'll be posting, all the assets that we went over today. And you can ask any other questions that you had in here. And just in general moving forward, we're really excited to, connect with you guys here. And I'll actually go ahead and just paste the link to the group in the chat so you guys can get to it. And we have a whole bunch of other groups as well that are more, line of business specific, so more tailored content. If you're really in service and support, you'll see a lot of content related, to that, for example, versus website or workplace and so on. And lastly, last exciting update. I know we only have two minutes. But under the resources page, once you're actually logged in to Coveo Connect, you'll see this new page called product announcements, announcements, which basically is a page dedicated to sharing any communications around, product deprecations, even early access programs, open betas, new releases. So it's really great if you do log in and you're looking for product information or communications. We have it here to complement any emails you'll get, or any direct outreach. So it doesn't replace anything. It's just an additional place for you to actually see it all in one place. So that's really all I've had today on the community. If you don't have access to Coveo Connect, please reach out to us, or you can reach out to your delegated admin on the account, and they can provide you access. It only takes a minute or so to to grant it. So, please do that if if, if possible. If not, just reach out to us. We're happy to help. And I think we can last our oops. Sorry. Launch our last poll. It's about that AI. So it'll just take a second. And we'll just give it a minute or so. While while everyone is answering, really thank you everybody for the, engagement. Really good questions. Absolutely. And it's really good to to read some of the familiar names. They're looking forward to doing more of those, and we're really here to help. So, these are for you. Right? So sometimes you're gonna feel, oh, I know this. Sometimes you're gonna feel I learned something new. We're more than happy to to work with you on, you know, specific topics and whatnot. And both Vanessa, myself, and Jazz are are very happy to collaborate. So any ideas, throw them in the user group, and we'll make sure we take them into consideration. And, everything that we talked about today will be available in the CS community group as well. Perfect. Yeah. So thank you, everyone. Everyone. Thanks, everyone. Bye. Bye.