Alright. So we'll get started not to lose too much time here. So hello, everyone, and welcome to our new in Coveo twenty twenty one year end review webinar. My name is Jasmine Oraz, and I work on the marketing team here at Coveo. So today's session is a special edition webinar, and we'll look back at some of the most exciting updates to the Coveo platform throughout the year. And we'll also share an exclusive sneak peek into our road map for twenty twenty two. It's obviously gonna be a great session, but before we get started, I do have a few housekeeping items to cover quickly. So for this webinar, we do encourage you to, type your questions out using either the q and a or the chat box in the right hand panel on your screen as we go along. We should have time for a very short q and a at the end of the session, but we can also answer any questions offline if we do run out of time. And as usual, today's session is being recorded, so you can expect to receive a copy of the presentation within the next couple of days. And finally, I'm excited to introduce our speakers. So joining us today from Coveo, we have Steven Rahal, our director of platform product marketing, and Nicolas Bordeaux, our VP of products. So without further ado, I will pass it off to Steven to take us away. Thanks so much, Jasmine, and hello again, everyone. Welcome to the twenty twenty twenty twenty one year in review with Coveo. Again, I'm joined by our special guest, Nicholas Bordelot. Nick serves as vice president of products at Coveo, overseeing platform strategy and developments. Nick, welcome. Thanks for joining us. And I was on mute. Thank you. Thanks for having me. Awesome. And so as Jasmine mentioned, the session today, we designed it really for for designers, for developers, for content authors, and administrators that collaborate on the web properties that form the foundation of your digital experiences. So, really, no matter how you're using Coveo today, whether you're just getting started, with Coveo, you consider yourself a seasoned vet, our goal is really to take a closer look at some of the capabilities that'll help push your digital experiences forward so that you can enable people to do more on their own and what whatever that means for you. If it means buy more, if it means fix faster through self-service, enhance proficiency through enterprise search and and, internal collaboration. We're We're gonna talk about some of the capabilities that can, unlock that potential for you. And so, really, for us, the the first part of the session, we'll explore some of the foundational components of the platform. So we'll look at the data layer and some of the improvements that we're making around how you index content and share your data that you're capturing as signals or as usage analytics data. We'll look at some of the services that power relevance and some of the improvements to not only the rules engine, but, machine learning capabilities, over the course of the year. And we'll look at some of the frameworks for building a a modern UI and some of the improvements there and and how you might leverage those. And then we'll look ahead to the platform road map and vision and some of the key investments the team will be making in the weeks and months ahead. And so with that in mind, Nick just wanna really dive in and start with the user interface. And I know as we look at twenty twenty one, we've introduced a new toolkit for designing and building a search UI. Maybe start for for everyone here today with with headless, what it is, why they need to create a new toolkits, and and what is the difference for those that are using JS UI today versus versus headless? Yeah. So JS UI was built, a long time ago, probably ten years, if not more. The library is pretty powerful, but it's also pretty heavy. And it's beginning to be, to be too heavy. We listened to lots of comments from from developers using our our platform. They wanted to have more flexibility at different level, more performance. So we decided it was time to start from scratch and build something completely new. We also wanted to decouple the, the the the the the the logic of the handling search versus the the the presentation layer. So headless is basically the the the logic. How do you perform search? How do you connect to the KobiO API? Everything is bundled and easy to to to to mix with, with with modern library like like Vue and React and stuff like that. Atomic is a set of components that goes on top of that. It's it's brand new. If you if you're using the JS UI, I think at some point, it will be beneficial for you to to migrate to those libraries. If you wanna have fine fine fine control over the UI, we would add less and build everything on top. If you wanna go faster, if you wanna if you wanna add something, up and running faster, we would add less plus atomic. Now if you're looking at specific ecosystems, and we look at, for example, the the Salesforce ecosystem with another number of our customers are are integrating Coveo into, Maybe talk a little bit about how headless and atomic are applied in in a Salesforce context and maybe plans for for other ecosystems, ecosystem specific adaptations or, more of the generic toolkit that will be applied to all. Yep. When we when we build an integration like we have for Salesforce, Sitecore, we have others coming from ServiceNow, Adobe, not coming, but we release from, ServiceNow and Adobe. We we like to to wrap all the components that makes it easy to work in that environment, connectivity, and everything. But, also, we want to for users not to have to go back to learn how COVID will work, to be able to build UI. So in in the case of Salesforce, we basically wrap the atomic component into into Quantic. We call it Quantic. It's basically lightning web components. That's a new method of of doing UI inside Salesforce. So within Salesforce, you can drag and drop component to build a search interface, build a user agent console, much much easier for users to do it in the in the context of Salesforce versus having to go back to Coveo to design new ways. I think when when you're evaluating, considering headless and anatomic and important points that, you know, many many are aware of is just the fact that Google puts a premium on page speed. And, because it's such a way of determining a site's quality and understanding user experience. And so, obviously, sites that perform well with this metric are rewarded with high ranking. And so, the performant aspect of of headless and and atomic is, is definitely a a great part of of this toolkit and what it can bring to the experiences that you're creating. Nigga, when we look at, again, search UI, we look at headless, we look at, the atomic UI component library. As you look at other ways to be able to integrate into different systems, one framework, one capability that we've introduced is the in product experience or IPX. And now some customers have taken advantage of this, the single line of code to integrate into different applications. Maybe talk a little bit about how in twenty twenty one and and into twenty twenty two, we're we're extending the IPX framework, what that means and maybe in what scenarios could you think about using IPX for UI integration? Yep. IPX is really if you wanna integrate Coveo inside an application that you own. One of our customers, Xero, they make accounting software. They want to provide Coveo access to all of their customers. So anything help related in when you're inside Xero is done through Coveo or any you can think about Google or whatever the application where you wanna have, help and search directly embedded in there. That's precisely when you wanna use the IPX. And what is new this year, Google Chrome extensions. So, basically, we if, and more I would say more in a context of a workplace scenario where you wanna provide, Coveo search to all of your users, we're typically used to go to the Internet to search against all all the information. Now the Chrome extension lives inside Chrome. It's connected to the Coveo, and you can leverage the context capabilities, of Coveo to provide, to provide content recommendation. In this case, we were editing a document. As you're editing that, if you wanna have complimentary document, maybe maybe spreadsheet or presentation that includes similar topic, you just open the Google not the Google, sorry, the the Coveo, the Coveo Chrome extension. And then you have recommendation right at your fingertips, and you can also perform search, straight from there. Can you define kinda where it appears depending on what what applications, the same thing? Up to you. It's basically up to you. You decide which, which are the domain in which you want to you wanna have the the the little work button. That's also configurable. You can have whatever you want in there. But the small button will appear depending on which domain that you configured, the app to be to to to be rendering. So it it won't be on all of the web, but when you're in a Chrome browser, if if you're in a in a in a tool that your company is using, Coveo can be rendered in there and providing context help. It's interesting for for anyone that's using or thinking about using Coveo in a in a digital workplace context, again, to Nick's point. You often think about enterprise searches traditionally been a destination. It's integrated into your destination site in your Internet or it's a stand alone portal. And so that the whole contextual relevance that the Chrome extension brings so that you have search, more easily accessible as as a user, as an employee, as I navigate, through different applications, as I'm working in the applications that I use on a day to day basis, it can be, a powerful tool to to not only provide access, but, provide a, greater usability in in many cases. And so that's for that first part really looking, Nick, into the UI sides and quickly at some of the enhancements in terms of how you integrate Coveo to different systems and the toolkits and and frameworks that exist. If we shift gears a little bit and talk about surfacing contents within those UIs and think about the whole process of crawling and indexing content. And so there's been, an introduction, the number of new connectors over the course of twenty twenty one, some improvements to existing ones. I think two in particular, I wanted to talk to you about, where we've made the move from, not just supporting indexing a system through the standard space or generic connectors, but providing native connectivity. Maybe the first one with it'll be experienced manner manager and the second with with Slack. So maybe talk to, the group about the native connectivity for AEM and and how that differs when you compare using the the generic site map connector for indexing content. Yep. Generally speaking, when we build a connector, we build it for a few reason. First, because there's demand, because our customers wants to integrate with that system. When we see repeatability and system being indexed, that's where we think about, about building a native connector. A a native connector is really connected at the API level with, with with the other system, in this case, AEM. And and we'd also did Slack this year. The benefits is that we're gonna get all the meta information. You don't need to expose it to be to be to be to be to be gathered by the connector. It's it's much faster. We also pull out security. If there's changes in the API on both side on Coveo or on the on the whole system, Coveo is responsible of managing the connector. So those are the the the obvious benefits of of using a native connector. Still generic connectors are are pretty pretty powerful too. When you want to index any piece of content, they're a bit a bit longer to configure, but we we were always able to connect to to to any source of information with our set of generic connectors. So there there's a difference between the two. But when we see lots of customer using the same systems, that's when we decide to invest in building those connectors. That's great. Any other improvements to connectivity over the course of twenty twenty one that that you wanna highlight for the group? We did, we did a few improvements on the web and sitemap connector to make them easier to use. We had a lot of feedback from from, from customers of our own support team. So we did a few improvement, but there is much more to come next year. We also improved our our, our SharePoint connector, to make it more performance. SharePoint environment are are getting bigger and bigger. So the the connectors are is much faster to index content. It's also easier to define what you want to index so that you don't end up indexing stuff that is useless. We also worked on, we also worked on ServiceNow. Many connector got, got some love, if I can use the expression this year. And, we we maintained them all to be to be always top and top level and performing. Yeah. The in specifically, the ability to include and exclude content from specific pages with SharePoint sources, I think, is is, powerful capability around connectivity. Yep. And so, again, we we've looked quickly, and as we run through this here in in thirty minutes, we we look at some of the frameworks and tool sets on the UI side, improvements around crawling and indexing contents, and things that you you might be able to leverage that you're not yet leveraging. And so if we move to relevance and move to the whole process of being able to tune relevance, Nick, why don't I start first on the machine learning side? And so, over the course of twenty twenty one, there's a number of different machine learning models that were introduced, and it was really the the first in a series of deep learning models. And so could you you talk a little bit about what deep learning infrastructure unlocks? And then we can step through, two of the models that are running on our deep learning infrastructure today. Yep. Deep learning is a, is a new method to do AI on our side. It it allow us to ingest a larger amount of information to be able to create those model. And also much more granularity in which in what we are processing. It's much more expensive, much more, much more longer to build those model. But in the end, the the performance of those model is is incredible. The the first one we the first one we build this year and we made available for mostly for support cases, I would say, is, is case assist. Case assist is basically a a model that helps, users who wants to report a a a problem, a support ticket. Let's say you're having an issue with your modem, with your Internet provider. You go to the the page where you report the issue. You wanna initiate a support conversation. The model is gonna be helping you to make sure that you're providing the right information to describe the case as as as best as possible. Should you provide the name of the modem that you have? Should you provide the name of the service that you have? Trying to help you to to provide the best description, so that in the on the other side, the agent gets the most relevant information to be able to help you. And as you refine how you describe that case, we're also able to to provide you the best, the best documents that you can self serve and and resolve those issues manually yourself. It's basically the next level. If you're searching in a search in a in a support environment, trying to find your own solution, you'll have a search box to be able to input what you wanna search for. So it's it's good, but it's limited. When you enter in a form and everything is preconfigured, not preconfigured, but preintermine, so we know much more. So we're able to get you to get you much further. And if the content is available, you'll find it this way. Else, the agent on the other side get the a pretty well defined support case to be able to help you. What about on the, question answering side of the equation, introducing smart snippets in twenty twenty one. Can you talk a little bit about smart snippets, what it is, but also how how, a user of Covalo, administrator of Covalo can think about its application and what scenarios it's best suited Yep. Smart snippet is, so you you see it in the in the in the page here. It's a it's a demo environment. Think about when you when you search for something and instead of having to open a document to be able to understand what you have to do to solve your issues or or to just to ingest the information. So that's exactly what smart snippet is. So if you have structured information on your website, q and a, documentation, troubleshooting documents, anything that is structured that COVID can understand. If you build a model for it, if you add it to your pipeline, then when questions come and then we have a a certain level of confidence that we're able to provide you the exact answer to what you're searching for, then instead of adding a a classic link or result template like like you want to display them, then you're gonna have a a an excerpt of that document where we think we found the answer for you. The the the the the flip side of that is that since you are consuming information in the search page, we don't know if it's helpful. So you also see at the bottom that there is a small widget. We don't expect everyone to click on those. Was it helpful? We think that a few people are gonna be clicking on that, but that's that's a feedback mechanism to help the model be able to understand if people are having success with, with the information that you're providing them. That's, that's one more time, helpful in services scenario, but also any website with FAQ or information structure can be can leverage this model to provide, in context response, to precise questions. So thinking about applications, customer service and supports, employee self-service, potentially within an intranet or portal environments, or even in your web, your dot com properties, if you have elements of frequently asked questions as a part of that experience. Yep. So that's kind of a quick spotlight on on some of the enhancements and advancements as it relates to machine learning. If you look on the rules engine sides, when you're putting in place, business rules or or managing relevancy overall through through query pipelines, there's been kind of significant investments in improving and a concerted effort, really, to improve the ease at which you manage and administer Coveo. Yeah. And that means removing friction, enhancing explainability, making it more intuitive. And so I think it'd be great if you can explain kind of some of the changes that have happened at the beginning of twenty twenty one and also into some of the plans for twenty twenty two in improving how you are managing relevance, how you are managing pipelines. Yep. We started the year. We wanted to provide a better experience for AB testing, instead of having to navigate between the pipelines and the and the UA reports to understand what's the results of those tests. So we decided to integrate both together. We reworked the whole flow to do AB test. But as we as we were releasing that feature, we realized that query pipeline as a whole are kind of complex. So we started a larger exercise to be able to come with, with more, with more help and better experience in that in those pipeline. In in January only, we're gonna be releasing, groups and campaigns that we've been working on for a while now. Groups and campaigns are basically a way to manage a a group of rules inside of a pipeline to be able to activate activate them by either manually as a group or time box them if you want to have promotions or stuff like that around that. So making it easier to manage that part. We're also adding some widgets to let them let you know who did modification and when those modification were done in in in the past. And there's a there's a whole, set of, of other improvement that are planned for later later next year. We can we can we could have a a full session just on query pipeline improvement. Actually, if if there's interest, just let us know in the chat and question, we'll probably we'll be able to schedule such a session. No. That'd be great. Planning out twenty twenty two and what just the potential enhancements there to how you manage pipelines and how you manage rules is is powerful. Now we looked a little bit, really quickly at machine learning and, obviously, the the usage analytics or signal data that the powers ML being able to get access to that. And so there are a range of data export options, or data access options that that you have as administrators. And so I think everyone here that's that's familiar with Coveo maybe they've used, the ability to export via CSV or maybe they've tapped into the API with usage analytics read API, to be able to have access to some of that signal or usage analytics data. One of the new advancements in twenty twenty one, was this ability to have direct access to Snowflake, the database. And so Nick maybe kind of provide a little bit of a a quick overview of of Snowflake, what it is, but also what some of the options are for administrators as they wanna get access to to some of that data. Snowflake is a is a huge data repository where you you can store a structure and unstructured information. We we did the switch, last year to, from Redshift Redshift to Snowflake to store basically all of our user user genetics information. So anything that's happening in a search page, keyword, search click search clicks or whatever is happening that you're logging against against Coveo goes into Snowflake into your precise, your precise, location in Snowflake. And with the reader account, we're providing you access, connect directly. So you consume the information within Coveo, but you can also consume it directly within Snowflake. One of the benefit of doing that is that you can decide to use, whatever BI tools you have, to to to to view the information the way you wanna see it. And also Snowflake allow you to, to connect multiple datasets together. So if you're using other tools internally, maybe to do merchandising or commerce or or or whatever you're you're you're doing and you're collecting information. If you put all those datasets within within Snowflake, then you can consume them together, and you can have a a better view of all your business process, not just, not just search related one. It's a powerful tool if you have data teams or, data savvy administrators to be able to analyze, the results of that data. And so we looked at some of the administrative functions that are that are very much low code or no codes, easy to be able to administrate a lot of point and click interfaces, look specifically at some of the data extraction capabilities with Snowflake, and wanted to, here at the end, kinda switch to enabling developers. So, what have we done and where are we going in terms of enabling developers to build great experiences? And what, what assets, what resources are we, providing them? Yep. So we already talked about headless and atomic. So those to me are are are tools for tools for developers. In the end, they build UI, but that's also that's also tools for development that have been built this year. We obviously build more code sample, more documentation, more more more more API, more API, more SDKs. The biggest the biggest change we did this year is to build a a CLI. So it come in line it come in line tool to be able to interact with Coveo. We are releasing more and more feature to that come in line. So currently, you're able to create a test organization. If you wanna do some some testing on Coveo on a different organization, you can you can create it from there. You can download your current configuration. Basically, if you wanna source control the way you the way you work with Coveo, you can download configuration and reimport them from backup, whatever. So it's it's really a a different way to be able to integrate Coveo in the development process of, Coveo is typically part of an environment. So you have Coveo, you have a a a either a WCM, a CRM. So you have multiple tools and Coveo integrate within that tool. So as a as as that within that toolset, as a developer, you have to you have to merge multiple tools together. And typically, that's done through command line. So now we have the option to do the same thing with Coveo. I think one of the key parts of developing on the platform is the creation of a test or or trial accounts, as you're developing. And so we talked to the to the group here about, what we've been doing around, different orgs, and and the trial organization as well and and what that can equip both those that are new to Coveo, but also those that are that are more experienced, users of. Coveo. Yep. It, so for a while, we've we've been offering test orgs for developers who wants to try Coveo and wants to wants to experiment with it. Test orgs are kind of hidden in the in the training section, or if you're currently an owner, you can create a new test org. I still recommend people to use test orgs if they just want to test some things. This year we added, front and center on the website, a way to try the platform. When we try, when we added that feature, we also added a new mechanism to sign into the trial, but it's available for the whole platform. So you can now sign in as a with a password less login. The the the Metallus Classic, you provide an email, you get an email, you and then you click on the link and then you get access to the platform for four hours. So frictionless, if you don't have an account, which we support, you can use this. You can use the same same mechanism to create a test org, much faster, much easier to get started quickly to, to experiment something within Coveo. That's great. This is kind of our our whirlwind view of twenty twenty one. Right? How how quickly can we run through, some of the new capabilities that you can leverage? As you look, Nick, forward with the team, and planning for twenty twenty two, road mapping for twenty two, what does the next year have in store? What are some of the big priorities, that's Yep. That you're investing in? Yep. So, so on the administration side, we're working heavily on the experience. As I mentioned before, query pipeline, there's a lot of work happening on query pipelines for many reason. Fields and mapping, we also got a lot of feedback around the fact that it was hard to understand how fields and mapping, how do you get content meta outside of Coveo to put that into Coveo. That's also complex. How to use those fields when you wanna build UI or or refine or whatever. The a lot of investment on the on that front. As usual, we continue the investment on the dev frameworks. On the data science side, we are investing a lot in the same experience. How do how how do you how do we make it easy for people to understand what's happening into Coveo, how to create those models, which one should be created, how to configure them, how to optimize them. So there's a a whole new team, on the ML experience side that is, that is revamping that section. Another big investment on that area is on data validation. If you don't get the right data into Coveo, then the model learns on that data, and then and then the models are as good as they can. But if we we're actually, we're investing a lot to make sure that the data data we're getting in is is as best as we as we can. And on the global infrastructure, so ISO certification, working on that toward the toward q one, and also reliability. We're working on most probably a green green, environment, to to to alleviate the fact that AWS can can be down sometime like we've seen recently. And region, we're releasing more and more feature to the whole set of regions so that the whole capability of Coveo are available in all of the all the regions where Coveo is available. That's a that's a condensed view of the of next year roadmap. Yeah. I think the whole theme here today for everyone is is gonna condensed view of what happened in twenty twenty one and what are we looking forward to in in twenty twenty two and and hopefully raising awareness for capabilities that maybe you've heard about or or haven't heard about and and might think about tapping into as you move into the new year. And so, Nick, as you were talking a little bit about, headless, and talking about the developer tool sets that that, have been built, the question in the chat around a customer using Coby out for Sitecore and wondering if they'll get a performance increase in faster page load, if they go with headless. Any any kind of guidance that you can provide, for those that are running their website on Sitecore today? Yeah. Definitely. So one thing we did to improve the performance on-site core is to get rid of the of the reverse proxy that was, available inside actually, that was, mandatory inside Sitecore. If you don't use security, you don't you don't need that reverse proxy. That's so that's already improving performance. If you need something much more performant, you can go Barca. You can you can decide to use Atomic right away. And if you want something even more performant and more control more fine control on the look and feel and the and the behaviors, you can go to headless. So they're not integrated into into the query for site core integration package, but they can be easily deployed manually just like you would deploy, headless and atomic any any any website. Another quick question here as we run close to time. A question around IPX is any specific features or enhancements planned for twenty twenty two that that you can share, at this time? Anything, that you can elaborate in terms of how it's gonna be iterated on? I don't have a pre I I know there's there's investment made to make the, to make the integration easier. Today, configuring IPX is is one big page, and you have to know what you're doing. And there there's no real feedback about how it's gonna behave. So there's improvement on that front. And for precise feature about IPX, I I could dig down a little bit and and get back to the person who asked the question. Yeah. We can definitely share directly, with with that individual as well. And so, again, goal today was really to provide a quick summary of of what's happened in twenty twenty one and look forward quickly into twenty twenty two and some of the things that we're most excited about. Camille has obviously been the business of improving relevance for over fifteen years. And so, being architected to enable every business to deliver relevant experiences no matter how much you've invested in, in data science or infrastructure, being able to empower every organization to create relevant experiences. And so, Nick, appreciate you running through some of the highlights from a platform standpoint, what happened in twenty twenty one, looking forward a little bit in twenty twenty two. And as you look for opportunities to be able to learn more about Coveo, to upskill yourself on Coveo on search, access self paced learning, earn certifications, join live events, take a look at the level up community, a great resource for education, and tooling to help you get the most out of your investment. Again, Nick, appreciate you providing this quick run through. Appreciate everyone attending today, and Jasmine for hosting, and I'll turn it back to you, Jasmine. Thanks, Steven. So, yep, we're kind of running to the top of the hour here. So, thank you, Nick and Steven, for presenting today. And just as a quick reminder to everyone, like we mentioned earlier, you will receive, today's presentation in your inbox so you can share it with your other colleagues or anyone who might have missed it. And on that note, hopefully, everyone has a great holiday season, and we hope you're as excited as us for these twenty twenty two, product updates. So thanks everyone, and have a great day. Thanks. Bye. Bye.