Hey. Good morning. I see a few people joining. My name is Evan Thompson. I'm director of customer experience here with Coveo. We're gonna be getting started in just a few minutes, and, I'm just watching people join. Once I see a few more members coming in, we'll kick things off and do some introductions. But while everybody's joining, I'd love to hear a little bit about, where you're from. Just say hello, good morning, and we'll get into some housekeeping, before we kick things off. Alright. We've got the chat going. Feel free to jump in. Let us know where you're from. Just say hello. While we're, again, waiting for people to jump in, just a little bit of housekeeping. We'll do some introductions in a minute, but we've got the QA section there we're gonna be monitoring. Please use that over the chat. So if you have any sort of questions, along the way, we'll be monitoring that and making sure that we're we're responding. And we've got a, I think, a really exciting webinar to share with you guys today. I'm looking forward to digging in. See if we've got a good group of people that we can get started with. Oh, looks like chat is disabled. Wait a minute. We made it happen. And I see hi, Robert. Hi, Cindy. Edward Edward from Ottawa. Not too far from us. Yeah. Yeah. We've got someone in Ontario there. Uh-huh. San Diego. That's far. Alright. Very cool. Okay. Well, I feel like it's, we're ready to go. Hamid, Vanessa, should we say hello and get started with today's session? Alright. Well, I've already quickly introduced myself, so I'm actually gonna pass it over to Hamid who's gonna be leading today's session, to dive in and give us a little bit of an introduction as well as, I'd love Vanessa to say hello, and then we'll get started on the material. Alright, Hamid. Over to you. Everyone. I see some familiar names, which is awesome. Thank you guys for, for joining. My name is Hamid Jiragji. I'm a manager, in the customer success success organization, and I worked especially with customer onboarding before along with Vanessa. I was gonna introduce herself in a second. I'm really looking forward to this session. We we did them more of, well, quite a bit last year, and it was really exciting to get to the engagement, and and, hopefully, it's gonna be, fruitful for you. So please feel free to, you know, ask questions and participate. We're more than happy to help. So, welcome, everyone. Vanessa. Yeah. Thanks, Hamid. As Hamid said, I see a few familiar faces, so that's really awesome to see. And as Amin mentioned, you know, we worked on the onboarding team at Coveo. I'm now a customer success manager, so I'll be here today to moderate the chat and answer any of your questions. I think for now, maybe, Devin, you can confirm. Maybe you can just type them in the actual chat, and I'll answer them there until the q and a gets enabled in a few short moments. Great. That sounds good. Awesome. Alright. I think we can kick things off. I'll take a look and see what's happening with the q and a, but we should be able to resolve that quickly. Vanessa's gonna watch the chat in the meantime. Beautiful. So welcome, everyone. What are we trying to achieve today? So, high level agenda. We have I'm gonna talk about some key concepts. Some of you will be familiar with them. Some of you, not. So we'll keep it super high level as well. I'm gonna talk a bit of what is Coveo relevancy. You know? On day one, you implemented Coveo. What do you expect? Right? And then we're gonna dig a bit deeper into the, platform itself, walk through around the platform, and then I'm gonna focus around best practices. Best practices meaning, okay, when should I use this feature? When should I use that feature? Obviously, there's a lot of things to cover. So today, we're gonna keep it high level. My focus is gonna be on the query pipeline, certain features in the query pipeline and the key concept there. And I'll be toggling between the presentation, but I mainly use my my test org because I really wanted to see, more of a hands on how to implement certain features. And then at the end, we wrap it up. We'll share some resources and and hopefully get you excited for the next session as well. So without further ado, I'm gonna go ahead and get started. Please have any questions, comments. Vanessa will help me answer them, and we can if if if there's a question that needs more explanation, I can take it, and we can answer it, live for everyone. Perfect. So some key concepts. There's a lot of, a lot of things. Once you have Coveo, there's a lot of Cavea specific concepts that might be puzzling for you at the beginning because it's not maybe some industry specific or not. So three main concepts I I I wanna I wanna I wanna tackle today. First of all, you're always gonna hear, okay. So we have this unified index. What is that? That's essentially a deck of air platform where all the different sources we have are indexed. Some sources are native sources as in integrations, out of the box integration. Couple of clicks, you're able to crawl a site map or a web source, right, or YouTube source. Or it could be something very custom where you have to use a REST API, push API, etcetera, to crawl it. But everything once everything is called from different sources, it's gonna live in the content section and and Coveo. So that's what we call a unified index. And then we have the concept of search ops. Think about it as the front end UI. You could have one UI if you have a simple use case as website search, or you could have five different UIs, your commerce site, your website, your agent insight panel, your community, multiple search hubs. All those search hubs will point to the Coveo, instance. Right? And then how do I know which information to display? This is where the concept of the query pipelines comes in. A search hub would point to a query pipeline. In the query pipeline, I'm injecting all the rules that I want from conditions, from filters, to certain machine learning model, to certain rules, to boost certain content, or, you know, deboast certain content. So keep that in mind. We define index, search hubs, and query pipeline. Now one of the questions that we get quite a bit is why do I wanna use different query pipelines? I partially answered this question in in in the previous slide, but, essentially, I'm gonna have an Internet, which is could be completely different than an external or website search. Right? So the fact that we can use a different pipeline with its own rules, it gives us more more, you know, more aid for us to tailor each journey for each, user base. Also, think about it that way from a solution architecture standpoint. Because those rules need to be maintained. Right? What I'm gonna do if I'm using different pipelines, I can inject rules that I want in specific pipeline without affecting the other ones or without having to add additional conditions to to architecture it. So super high level distinct journeys, solution, architecture, and, obviously, what we call contextual alignment. We get into details a bit later about machine learning, but machine learning learns from from search hubs by default languages and tabs. Right? And I can create a machine learning model specifically for one audience, and then I'm gonna basically attach it to my specific pipeline. If I and and one sentence summary, think about it that way. When we are tailoring to different audiences, we we usually recommend using different pipelines so I can control the journey and so I can monitor the journey and drive the traffic exactly to where I want. Alright. Now let's kick it up a notch. How does Coveo relevancy work? So perfect. But Coveo has machine learning. There's a lot of things going on, but how does it actually work high level? So first of all, we have the algorithm, the ranking weights. Coveo algorithm is essentially everything in Preveo, all the results have a score. Right? And that score is determined by the Coveo out of the box algorithm, which takes into consideration different criteria to give a re result, its score. I'll give you an example. If an item, let's say, was updated yesterday, it's gonna get a bit more scoring than an item which was updated three years ago. Right? If I'm searching for for Vanessa in my Internet and Vanessa's name is in the title, the that specific result will be boosted a bit more when it's just buried within, you know, a site or within a PDF. So this is what we call the out of the box algorithm, and it is mimicked into what we call ranking rates. So Coveo already have default ranking rates, which I'll show you in a bit as well. And those ranking rates will basically influence the algorithm. Between the algorithm and the ranking rate on day one, you're really gonna have good search irrespective of machine learning, irrespective if you have any, manual levels you wanna inject. You're gonna you're gonna have really, really good search on day one before anything else. Now is it enough? Sometimes. When it's not enough, what can we do about it? There's there's two main things. First of all, we have result ranking expressions. Those are manual rules that we inject, right, in order to influence result. One example would be if a certain, let's say a certain author, I have a blog, right, and then I have multiple authors, I want every all the articles that are offered by Devon to to get a bit of a more boosting because that's a business rule. Right? Then I'm gonna inject this into the pipelines, and it will be added to the score within the algorithm on the ranking. So I'm not gonna ignore the ranking weights and the algorithms out of the bars. That's gonna give them a bit of a boost. Now you might ask me, why would I want to do this if I have machine learning? Because machine learning takes a bit to learn. Right? It's gonna start learning the moment we deploy it, but it takes a bit of time for it to really does its work and then goes to its potential. In the time being, I'm gonna give it this extra push by adding manual rules. Usually, one question that you always get as well, how long does it take for machine learning to be actually effective? Depending on the machine learning model, we have couple of them. But if I if I give you the the answer specifically for automatic relevance tuning, which or ART, which basically adds scoring and boosts content based on popularity, based on clicks, based on user journeys, I would say maybe ten thousand searches and clicks. Keyword here is clicks. It's not considered a successful journey unless the end users actually click, so they search and click. So that's why with all these four in place on day one, I'm very confident that Coveo will really give your give your customers a pretty good journey to start with. Now those business rules. Right? So are those manual tweaks that I'm gonna I'm gonna put? What do I do there? Do I just add every what I can think about? Do I just go into UAT and search for everything the single query that's possible and make sure the result I want is on top? Answer is no. What we use what what we usually do, especially at launch or you if you're launching a new a new, UI, we would focus on the top twenty, top fifty queries depending on the use case. Right? I really don't want to put effort and statically boost content or or or play with it if only one person searched for it last year. So we gotta take those twenty, fifty queries during UAT. I'm gonna test all those out. I'm gonna see where the result is living. If it's living in in in, you know, number one, two, three, that's fine. Even number five is fine. But if it's buried in page number three, this is where I want to give everybody for boosting them. So UAT is gonna be really important for for us when we test our queries. Also, acronyms and synonyms. I'll I'll show you some live examples, but acronyms and synonyms are are something that I you probably wanna inject from from day one as well. Example. LD LD starts I'm gonna I'm using the word LD because it's Luca Doncic. I'm gonna I'm I'm using my team today as an NBA playoffs. So a lot of users might search for LD. LD stands for Luca Doncic, but the system does not know that it stands for that right now. So all I'm gonna do, I'm gonna add manually enter a synonym where whenever a a customer searches for l d, right, we're gonna give them results for leucadantes as well. Right? So that's one of the examples where we use, synonyms and acronyms. I'm also gonna show you live when not to use synths because sometimes we overarchitecture it and try to, you know, add stammings and and add, you know, simple typos. We don't need this because Coveo will take care of typos access. Last but not least, business rules. We have certain business rules in place like the example I gave you about offers. That's just, you know, one example. Another example would be most of your audience is really interested in looking at videos, how to videos. So what I do in that case, I'm gonna boost everything that's, you know, videos on top to make sure I enhance the customer journey, again, until machine learning kicks in and and and helps. So I'll take a quick pause and ask if there's any questions or or comments in chat or QA, Vanessa. We're better now. I'm doing a great job then. I love it. Cool. So now, if I go a bit more detail, so what are those query pipeline rules in specific? How do I inject them or where do I inject them? We can we can we can, Today, we're only gonna focusing on on ranking ranking results and modifying the query. Query parameter overrides, this is a bit on the technical side where I can modify elements other than the search UI itself, enable some some some parameters, enable, you know, partial match, stuff like this. One good there, this is gonna be covered in future sessions. But from a from a ranking ranking results perspective, we'll give you example about tracking expressions in the platform right now. I'm also gonna give you a couple of examples about the source entries, stop words. I'm also gonna touch on triggers, notification triggers, which is a concept you might have heard of, but it's it really it really helps you promoting certain content or sometimes putting some sort of advertisement on top of the results. We we Coveo it in a bit because I know we got some questions about it, offline as well. Alright. So, you know what? Enough from the slides. I have couple of slides left, but I'm just gonna jump jump right into the system, and that's the exciting part. So here we go. I am I'm sharing right now our Coveo double o seven. This is our internal demo. You're gonna find a lot of things in it. Everything is enabled, many sources, etcetera. So what I'm gonna focus about today, I created couple of sources literally in thirty minutes. I created a hosted search page also in thirty minutes, which mean you can do it yourself if you wanna test as well. It's really not complex. I'm gonna focus on this. So this is the Kaveria platform. If you have not, seen it yet, any admin in your org can add you literally by going to members, add a member, and they can add you to your own, cover your org. Typically, you'd have a production and a nonproduction org. I recommend playing with the production with the nonproduction art, obviously, so we don't break anything. And that's your that's your main screen. On the left hand side, you you you'll see it's it's basically, split into a couple of sections. Section number one would be my content. Content, think about fields, think about think about, logs on issues with the index, with the fields, think about sources. If I go here and add a source, remember when I talked about some native, out of the box sources? You can see them all here. Some of them are more complex like push REST API, but some of them are really, really simple. Web source, for instance, if I wanna crawl a website, I'm just gonna click here. I'm gonna put the website name, test. I'm gonna put the website URL. Three next, and you basically crawled the website. But, of course, I'm doing it high level. So maybe if there are certain content in the side that we need to web scrape, we need more configurations. So I'm making it simple, but it could be a bit more more complex. For for easy sites easy site maps, that's that's that's super easy. And you're gonna see all the sources here that that that that are crawled in in your org as well. What I did, as I told you, I crawled two things through the sitemap. I crawled the NBA official page, which is still retrieving content, by the way, and I crawled my favorite team, which is the Dallas Mavericks as well yesterday, just to crawl both from a from a site map perspective. So I have those in my index right now. Perfect. And that's the content side of things. Now if I I can show you one more thing. I know I wasn't planning to to do it, but I may do it here. In the content browser, if you wanna see super high level what's in what's in what's in your in your index and the fields, Let me do that here. If I go to my Dallas map source only because I don't wanna see other stuff, this is kinda super high level view of all the content that I was able to crawl. And if you double click into it, you will see also all the fields that we crawled out of the box without web scraping, without anything. Those fields will help us in a bit with our if if we wanna inject business tools, if we wanna inject ranking weights, etcetera, those are gonna help us, a lot in a bit. So one really nice view you can check is under the content browser and then choose your source. In this case, I chose the source because there's multiple other sources, but in your case, it should be pretty simple, straightforward. Everything is gonna be there, and then you can even search for content. For example, if I do Luca, I'm gonna say I'm gonna find all the content about Luca Doncic as well. So it mimics your your your your UI, but more raw, no no, essentially, there is, you know, no CSS, no no visuals in it. Good. So that's my content. Now you have commerce, service. Those are very specific to your use case. I'm gonna go into it because as I said, I'm gonna really focus on search here, especially the query pipelines and search pages. So under search pages, if you're an admin in the org, you have access to what we call search pages. Search pages, those are history search pages. So this, the end user will not see it in your front end. And if you have Sitecore, whatnot, they won't see it. It's only hosted within COVID. And what's really cool about it, I created this one where I'm gonna show you in two clicks how you create you can create your own hosted page. If you add a search page, there's two ways to do it. There's a simple builder using our atomic, UI You can make classic using JSC UI. If I use the simple one, which is the recommended, clicking on it, gonna give it a name. Let's call it test webinar. I'm gonna add, and then it'll ask me to log in again. And done. I pretty much have a search page right now. It's really as simple. Note, I have all the sources in my org. Very important because I did not put any conditions or any filters. That's really important to note. So in in your case, maybe it's gonna be easier because it's you you you need all your sources in your org. But for me, that's not gonna do it for the demo. Right? Because I have all the sources, so I need to do some configuration in the pipeline. But it's really as simple as that to create a host of service, and I really recommend, all of you to do it so you can test anything you wanna do in in your pipeline. Alright. Let me close this because I already have a page for myself. I'm gonna go back here. Ask me to log in again. Sorry about that. And I'm gonna go back to the w o seven. There you go. So this is my, this is my search page that I that I told you about. Now let's go into the fun stuff. Query pipelines. We talked about, search hubs and query pipelines. So now how how can I link this page, which I created, to my pipeline? In your case, how do you link your web page, your Xonet, your Internet dot com page to a certain pipeline? How do you link your Internet to a certain pipeline? We do this by creating a condition, and the condition would be the search hub name. Search hub name meaning the name of the UI. I call this NBA playoffs. And then when you go into your pipeline, you're gonna create a condition, and the condition will be search hub is NBA playoffs. I already created it, so I'm gonna create it again, but it's it's here. That's the step number one to make sure whatever I see here well, sorry. Whatever I do in my pipeline is reflected on my hosted page or is reflected in your in your really important. Or else, I'm just gonna have it's it's my UI will point to the default query pipeline. So a lot of things going on here. Let's start with the easiest search terms. If I go into my search terms, I have two options, have the source and stop words. Stop words, I would use it, for example, if I am noticing in my search there's a lot of things like and, or, off, out, or maybe some bad words that I don't want to query at all and because they are affecting relevancy. What I would do, I can create a a I can add a stop word and say, system always ignore and. Right? Always ignore also, let's say, oh, etcetera. Does it always affect relevancy those words? Not necessary. We have to rely on our reports to tell us if they're actually affecting, relevancy in any way or another. Right? So in this case, I just added an anore for just an example, but you have to determine from your own use case what kind of support you wanna you wanna have. And sometimes, honestly, you don't even need to have any stuff. Now let's do add the, let's add the thesaurus. Thesaurus synonyms, if I go here and add it, rule, I have several options here. I'm gonna do the the the example that I gave you at the beginning about Luca Donchitz. So in my search page, if I search for Aldi, I do have some results. I have thirty three in English, thirteen in Spanish, and my source is I added, you know, source as a as a FASET, Dallas, and NBA official. Alright. But that's because some people quote we're saying to lose LD blah blah blah. So that's just quoting people saying l d, but they actually mean lookup. To do that, I'm gonna go here. I'm gonna say whenever somebody searches for l d, I want the system to also search for Luca Donchic. Right? And I'm gonna take it up a notch. Actually, I'm gonna say l d seventy seven. His number is, seventy seven. So anybody who searches for this, I want the system to automatically give me the results for all three. Once I add it all, I go here. Now let's look again for LD. That did not work. Give me a second. Include I'm just gonna refresh it. And then here you go. I just need it to refresh the page. When I search for LD right now, look at all the results that I got. Right? Because it's expanded to Luca donkeys. It expanded to Luca seventy seven. So as as you noticed, it was reflected right away in the UI. And this is something that you could do as well, and it will be reflected right away in your front end considering you are doing it in the right pipeline, of course. That's one example of, of synonyms. There are other types of synonyms. The first one I did was more of a two way. So l d, Luca, don't you? And Luca, don't you? There's other ways when I wanted only one way synonyms. So I don't want to expand Luca don't just to LD, only LD to Luca donuts. Sometimes it happens. There are use cases where we might use it. Most of the time, we use just a regular synonym because it doesn't change much. Then this is also interesting. So if, for instance, you have a a brand, right, and then you completely rebrand it, and now whatever you were calling your let's say it was Wilson in this case, And now Wilson is not the official, you know, ball provider for for for the NBA. Now it's actually Nike. What I can do whenever somebody searches for Wilson, because there's nothing or I rebrand it, I'm gonna change the query right away to Nike. And that's could be very, very helpful, as I said, for rebranding or for legal issues if you're not allowed to use specific brand specific names in your search, we can actually, do that as well. And then, one question we always get, what's the difference between match terms exactly and the other ones? This one is case sensitive. So l d will only be replaced when it's l d. It's not gonna replace when it's l d and small l and small d. G. Everything I'm showing you, you can inject the condition in it. I'm not gonna go there, but you can actually say only apply the synonym in certain places. For example, only in my Internet or only in my XRNET or only in my when the user is Spanish. Right? Or when the only when the locale is English or Spanish accent. I'm gonna take another pause and ask if there's any ones that just came in kinda privately to the chat as well that we wanna jump in, and there's a couple ones that just came in kinda privately to the chat as well that we could maybe take a pause and answer. Sure. Perfect. Yeah. So I answered, yeah, the one with the q and a. And then the other one we have is, should we adjust ranking leads if we find that most of the results are very relevant? When do we use those versus the other types of rules? Great question. So ranking rates, which I promised I'm gonna go over a bit. Thank you for the good reminder. So under in the play pipeline, under advanced, if you go to ranking weights and then add a rule, those ranking, ranking factors that I talked about, remember the the out of the box algorithms, this is how we see them. So they are item last modification date, keyword frequency, so how many times the word Vanessa is in a in a document, keyword in the title, keyword in concept. Concept is just a linguistic algorithm of extracts to create a field. For example, in in our example, then the playoffs, it's gonna be, let's say, in in a field because it's, in the concept field because it's it's what the the system determined that it's linguistically relevant. Here in summary, if some, pages have a summary, if I'm seeing the keyword, I'm searching the summary, they're gonna get more a bit of more boosting. And then last but not least, keywords pre proximity, which means if I'm searching for a three, four words, if the three, four words are next to each other in the document, we'll get more. When do I touch this? Essentially, never. Unless there's couple of specific use cases. If you have a commerce use case, I know for a fact that item last modification is not as relevant because I might be searching for, Jordan shoes. Doesn't matter if they were they were, you know, they were or it they were in entered in the index or if they were launched yesterday or ten years ago. I'm searching specifically for Jordan Chavez eleven, whatever. That in that case, we play a bit with the item last modification there, let's say, as an example, and make it less less relevant by moving it towards this site. But I wouldn't touch it, go live. After go live, maybe one week, two weeks, a month, I'm gonna go into my reports, and I will see if there are certain patterns, very low click through rate, very low, engagement on certain queries. And then I'm gonna take an educated decision saying, What I'm seeing is a lot of searches are literally keywords in the title. And if the keyword or if the if the keyword or the query is exactly as the title, I want it to have way more boosting than anything else. And in that case, I'm gonna add it like this, add it all, and then I will test it up. Note, I don't have to just do it more of a big bang. Just add it, increase it, and add it all. We have what we call AB testing where I can say, send twenty percent of traffic to the, the the pipeline where I made the keyword and button more relevant, keep the eighty percent as is, and then the system will actually tell us, did our click through rate get better? Or maybe nothing changed. And in that case, probably there's an issue, somewhere else. Hopefully, this answers the question. If you have any follow-up to that, please put it in q and a as well. Well. And there was another question you said, Vanessa? Yeah. So more around, like, the pipelines. So we were talking about the query pipelines. Okay. Just, hold on. Let me get it. Mhmm. So when should we have multiple pipelines versus just one? So maybe just more of like a practical example. Sure. Good question. Two completely different use cases cannot have the same pipeline. And I'll tell you I'll I'll show you an example here. So this pipeline that I created, if if I I leave it left it as empty with no conditions, then my hosted search page is not gonna be relevant because it's gonna have all the content in my my test org, which is everybody uses that. Right? So to tailor it to you guys and myself as the audience of the theme as MBA, I went into filters, and I said, my constant query, I want to only see content from these two sources that I crawled, which are Dallas Mass and NBA official page. So think about it that way. The audience is very specific. Use one pipeline. I need to put certain rules. When it's a when it serves a completely different purpose, then I would create new pipelines. Now, conversely, if I when would I use the same pipeline in that case? If you have one website, multilingual. Right? Have one in English, one in Spanish, etcetera, based on the locale. But but all the rules, all the machine learning, models there and even the sources, I just want them to be essentially the same. I would use the same pipeline. The only thing I would do in my filters, I would dynamically grab the locale and say whenever the locale is e n, show me the sources that are English only. Whenever the record is, show me the my students SP. Show me the sources that are Spanish only. And then everything else here will still apply. And I don't need to differentiate them, especially if the the the distinction between Spanish and English is not much, and you probably want to show maybe some overlapping content. So think about distinct user journeys, distinct use cases, completely commerce versus serve service and support, etcetera. Hopefully, this answer you. We can share also after the call some best. We have we wrote some some blog articles about best practices when to use different, pipelines. We can also share it with you. And thank you for your questions. Right. Anything else, Vanessa, you want me to Coveo, or you have it all covered for now? Beautiful. So Okay. We were talking about the, search terms, etcetera. So now I wanna add one more thing to search terms. When should you not use synonyms, as or as the service entries? One example, if I search for I searched for foundation, but I butchered it. I forgot the a. No results for foundation. Query was automatically corrected through foundation. I'm not gonna use the source entries for very simple typos because Kuva will take care of it two ways. Number one, through this, what we call the did you mean feature, could be did you mean automatic or did you mean regular, which I'll I'll talk about it in a bit. So let's not over architect it. Coveo will take care of these kind of things. Unless it's a type which is completely butchered, like, even the system cannot get it. In that case, I would add it. Another thing, I don't have it enabled here because, remember, I told you machine learning makes takes time to learn, and I created this today. I'm gonna show you on actually on our docs page, which is an amazing resource, by the way. Everything I'm talking about, throw a search here. You're gonna find it. I I guarantee you. So if I search for pipeline, let's say here. My I I wanna search for pipeline, but I also butchered it really, really bad. Look at all these guys here. Those are query suggestions empowered by machine learning. So this is my first line of defense anyway. I don't need to add disorders and this for for for this kind of sentence. And even if I completely insist and say maybe, let me do this and search for it. Look at that. That's my did you mean feature, which is similar to the automatic feature we saw. The difference in case you're curious. Difference is because in Coveo docs, we are crawling some, blogs and stuff like that, this actually in in those eighteen things, eighteen results we found, there's actually pipeline. So that's why the system will not decide not to show you the results. Maybe you are actually intentionally did a pipeline and then blog so you can find your blog or something like that. Right? So it's not gonna automatically correct it. It will give you a suggestion. And the other case here in my page, it automatically corrected because there were no results at all for foundation without any. So that's just a bit of a differentiation between did you mean automatic and did you mean regular? Now the question that came through, how do I enable it? You can enable it through the org in in the query parameters. You can also enable it from the front end. You can basically, statically say enable the domain is true. You can do it two ways, and it's it's available for both for actually all our, UI frameworks. JS UI, atomic Coveo happens. Alright. So we set when we not when we should not use the source. And please now let's go to more of a ranking, a deboosting and boosting, we're talking about. If I go under, result ranking and I go into under, adder rule, have two options. I have ranking expressions and featured results. Featured results. Remember when I talked about scoring? Every result has a score. Featured result, I use it when I wanna pin a result on top when a condition is met irrespective of everything else. The way the system does it behind the scenes, it's gonna give that specific result you choose a million points. So it's gonna override machine learning. It's gonna override all the other manuals you rules you put in place, and it's literally gonna get a million points addition. When would I use it? I will show you here. Let's say I'm searching for where is the parking. Right? That's really not not giving me what I want. There's FAQs. There's old content. That's not good. So I want to have the parking page pinned on top whenever there's a search that contains the word park. How do I do that? And I go here. I'm gonna go into featured results. I will choose the item I wanna feature. The item I wanna feature was parking. So let me try to find it in the index. I'm gonna select an item here. I'm gonna say, maybe maps parking. Oh, there you go. Found it. So they actually have an official parking page for their parking. I will select this item and add it. Now if I add a rule with no query, it's gonna be pinned on top for every query, which I don't think I want. So I'm gonna say whenever the query contains parking. I want this page to be featured. So let's add a little and see if it works. You're gonna get prompt into what do you wanna call it, and then you can add some user notes. Best practice is to to to say what to put the date when you added it and the purpose just for other users because because other admins can also see this. For the sake of simplicity, I'll keep it as is. And I add the rule. Let's see if it works. Where is that parking? And let me just like the other time. There you go. I need it to refresh, and look at that. It's now on top pinned whenever there's a word parking. Let me see if it works with parking only. It does. Parking AAC. It does. AAC is the American Airlines Center, which is where the Mavericks play. So and that's essentially how I did with three clicks. I was able to feature an item. Super easy. Keep in mind one important thing. It will always be on top until you remove the featured results. So you wanna be careful. If you're not sure, I wouldn't use it. That's the featured results. I'll also take a quick pause, see if there's anything we wanna answer or anything that came through. Yeah, I can continue. Perfect. I'll continue in that case. Actually, the I see a question, sir. So what if I want to show that a result is featured? That rule does not show a featured badge automatically. Great question. So you're not gonna have an automatic badge that says this is featured. Really good question. But we have a a component which is a badge, and that that badge will show up here. You just need to put it in your front end, and it will show you that this is featured. That's actually great best practice because I don't wanna force stuff on customers kinda like an ad. I just want to label it properly that okay. I I wanna fee we are featuring the result because we believe that it's that it's, it's gonna be relevant for you. Great question. Thank you for answering. Nice. Now let's go here again, and I'm gonna show you another feature here. This one is a bit more complex, but super powerful. Featured result, I said, give you a million points, gonna be on top all the time. Ranking expressions. What if I want to boost a group of items? Right? And I don't wanna boost them by a million points like a featured result, but I wanna boost it by a tiny bit. One example would be, I want all the content that was published last week to have a bit more boosting all the time. The way I can do it, prerequisites are we need to have that feed that I'm boosting on and the index. Obviously, that's like one one or I cannot I cannot boost it. Once I do that, I'm gonna search in this case, let's let me search for date. Just the that's just generic date. I'm gonna say whenever the date since last week. So whenever a result has a date, which is a week old, I want to boost everything by, let's say, a hundred seventy five points. Once I add it all, I'm gonna name it Content. Add a and that's it. Now in my front end, Same thing. Everything that is, weeks old will get an additional boost. Key point here. I did not boost it by a lot. Meaning, machine learning, the out of the box other than my other business tools will still apply. Right? I just gave it a bit a bit of a boost. And this is something that we would do usually when we go live to implement certain business tools. In this case, it's date. It could be pretty much anything else. It could be if if the author is, as I said, author of the blog is the owner of the company. If you are a law firm, whenever the actual, author is the CEO or the major the department, I can I can do that? And you think you can think outside of the box here. Any fee you have in the index, we can boost on it. Can also do the opposite. We can deboost. Right? I don't want to remove the content from the index, but I wanna deboost it. Example, if you are in commerce and anything with a low margin, maybe I wanna deboost it a bit, right, to make sure it doesn't show up, and I keep the the the the other items that have a bigger margin for me more on top. I showed you how to do it in the what we call the simple filter, but if you are really interested, we have an advanced mode where you can actually write any expression you want. It doesn't have to be one expression. You can use and. This is the query, conveyor conveyor query syntax, which we could also share with you as well. So once you're comfortable, you can also do that as well. Alright. Last but not least, I wanna cover under advanced triggers. Triggers are really nice concept. Similar to what we did here with a bit a bit different serves, different purpose. When I add the rule, notification triggers. So what are triggers? Triggers are something that's literally triggered in the front end when a certain condition is met. We have different types of triggers. We have one which called notify. Notify, it's simply gonna display a message over the results, not on top of the results. The message is gonna be here, not part of the results. And this message, the difference between this and featured result, it does not need to be in the index. It could be any message you're putting. It could be there's a webinar happening. Right? Or it could be, it could be literally a notification about what are the steps we taken to to to counter the COVID spread COVID nineteen spread. Right? To do it, there's we can make it complex. We can we can decorate it. We can also share after the call some examples, or it could be as simple as something like this. If I I was thinking about an example to give, and let's say, users are looking for phone number. Right? Phone number to call the the the ticketing the the ticketing office. What I can do, I can say this is the phone number for the ticketing office. And, of course, I don't want it to show anytime or no no matter what the query is. So I'm gonna say here add a condition saying whenever the query contains phone, Show me that ticket. I'm gonna add a condition. Before I add the rule, I wanna see what do I get right now when I do phone in my UI. Oh, getting bunch of stuff going on. Two six two forty six, two forty eight fourteen. That's not bad at all. I really don't see don't see what I what I want. So if I go here again and now this is my message, and I add the rule. Then I do the same search here. There you go. This is the phone number of the ticketing office. And that's as simple as that. It looks ugly because this is all out of the box, the framework, and we did not decorate it. You can decorate it. You can add, hyperlinks. You can add, picture as well. And what's really important, we can also track if we have a hyperlink here and there's clicks on it, we can track it, and the analytics will come to compare so we can analyze it as well. Those are this is my notification, trigger. Wanna show you another tool, which is the last one I'm gonna show you today. It's called redirect. I'm on the edge in this one personally because sometimes if I'm direct redirected right away, maybe it's not a great customer experience, but it could be helpful in certain reasons. So or certain ways. Let me see if we have, let's say, carriers, and b a. Right? They have a careers page. Beautiful. Let me copy this. And if I go back here so I'm gonna say whenever somebody on my page queries any query that contains careers? Jobs be multiple as well? Highly conditioned. First of all, before doing anything, let me search for carriers here to see what we're gonna see. Carriers. Okay. Perfect. So we have we see a couple of things here. If I go here and add this rule, Let's see what's gonna happen if I search for all jobs, maybe. Here you go. I was redirected right away to the Coveo page. As I said, I'm a bit on the edge to sometimes customers, if they search for something, they wanna actually see the results. Sometimes they just wanna be redirected, but that's something also available at your fingertips to explore and see if it makes sense for, for your end users as well. There are other stuff here like execute and query. Execute is pretty interesting because it's gonna execute a JavaScript function in the front end, meaning you can change the whole color of the website when a certain condition is met. You can have an actual pop up on the page, and the pop up would trigger by JS function saying, you've won the log or something like that. So you can pretty much do do that, as well. As you could see, we covered barely, I'd say, fifty no. Not even forty percent of what you can do in the pipeline, but those are the ones that are typical, to be honest. Like, there are more deep stuff, but those are the typicals that will help you solve a problem based on, the analytics you're getting. Really important. Any rule that you wanna put in place outside of the business rules that you're hundred percent sure about, Go back to reports. Go to your search details report. See where we where we have gaps and action upon it. A lot of times, we try we we assume that this is what our end users want, but it's not true. So let's see what the analytics are telling us in order to to execute them. We can have an analytics session, by the way, as well. So I won't do any teasers right now, but some one of the future sessions will be about analytics, and we'll dig deep, link those analytics to any work that we're gonna do in the pipeline. And I think I have couple of slides. I'm not gonna go through them because I covered them live, but we will send them, in the email as well, post webinar for you guys. I think there's a couple slides at the end that just discusses some of the resources there. We wanted to wrap that part up. As we're just getting that sorted, I'm looking I don't see any additional questions coming in. Nothing, Vanessa. I think everybody, if you have anything else that you'd like to ask, obviously, feel free to jump in there now. You know, as Hamid is kind of wrapping things up here, one of the things that, you know, we wanted to do is better understand, you know, a little bit from you what are some of the challenges or or issues that you may be experiencing that we should be knowing about. So I have a quick poll here I just like to launch. And as I said, it really is more about just getting a better understanding from, from our end users. You know, what are the challenges? What should we be aware of? You know, where can we jump in and help? Whether that's from a customer experience side or customer success side of things. So I'll just give everyone a few minutes here to just go ahead and and respond. Let us know. There is a chance to just add in kind of the other piece if we're missing anything that I didn't prepopulate. Don't, don't be shy. Let us know. And while you're doing that, as Hamid kinda quickly mentioned, you know, we do have a couple more sessions coming up. So we're really excited to start to dig back in, get a little bit deeper. We're looking at these, you know, foundational sessions as helping you build that foundation. Some really, you know, good reminders for people that have maybe been with us for a while. But, also, if you're new, just getting started, I think these are the perfect, sessions to dive into. So if you haven't already done and done it, there are two more. One on, June sixth that, as Hamid has mentioned, mastering those analytics. So all of the wonderful things that he's just shown you, how are you gonna dig in, report on it, better understand, you know, if you're successful and how Cabello is having an impact. And then the last is really diving into the the the big stuff here in relevance tuning. So I think this one, you know, Hamid is gonna be really expanding on some of the things that you've seen there. And and I think this one is really, again, where we're gonna get our sleeves up a little bit, and learn a little bit about, you know, actually how we're gonna be doing this and not just all of this in theory. So I'll just end that poll. Thanks, everybody for participating. Anybody that joined us today is obviously going to be getting, the follow-up email that will include these resources here, that, Hamid was just showing, a recording as well as the slides. I will add a link in to redirect you to any of the other registration pages, as I said, in case you've missed them. And we wanna make sure that you understand how to self serve as much as possible. There's a lot of great resources. Our Coveo Connect community, there is level up for your, self serve learning opportunities there. We have our technical support. There's blog articles for best practices. So we're gonna make sure that all of this is bundled up in our our thank you email. Thank you for joining, and we really hope to see you guys in the next few sessions. And we're excited to dig in and figure out what's next for everybody and how we can really drive some impact for you. Thanks, Hamid. Thanks, Vanessa. I don't see any other questions here. So it looks like we can wrap up. We're about, fifty six minutes, almost to the full hour mark. Perfect. Thank you, everyone. Thank you, everyone. Looking forward to to, working with you on the future sessions. Yeah. Great. Forward to those. Thanks, team. Bye. Thanks, everyone.