Good morning, everyone, and also good afternoon for those joining us from Europe today. Welcome to New in Coveo for Commerce. This is the fall edition. My name is Sheerine Reid, part of the Product Marketing team here at Coveo. And I am also joined today by some of my colleagues from the amazing Product team. We have Sebastian Alvarez and Benoit Thibault who are also here today. And they'll be spending the next forty minutes or so diving into the latest and greatest functionality for our commerce product, as well as providing you probably with a few sneak peeks into what's coming next. A quick reminder that everyone is on "listen only" mode due to the large amount of attendees on the webinar today. However, we do really wanna hear from you. And so if you have any questions, please feel free to pop them into the chat, and we have some special time reserved at the end of the session to answer those. And before we get started, I will answer typically one of the most popular questions that always comes in, and that is, yes, the session is being recorded, and you should receive a copy of the recording in your inbox within about twenty four hours or so after the webinar is over. And with that being said, let's get started. Alright. So before I hand it over to to Seb, I wanted to just provide a few data points that kind of reflect what's going on in the market right now. And the first data point is seventy two percent. So we recently released our twenty twenty five holiday report. And for those of you who don't know, every year we commission a research firm to survey a large amount of shoppers. So this year, we did about six thousand shoppers to get a pulse on, you know, the buying behaviors heading into the busiest season of the year for typically, for retailers. And this year, what we saw was a little bit surprising. So the signal was loud and clear. Seventy two percent of shoppers said that they're open to using generative AI or Gen AI for guidance in their shopping experience. And that number actually goes up to eighty five percent if we're just looking at the Gen Z and the Millennial cohorts. Right. So that's the first data point. The next data point is eighty nine percent. So on the B2B side of the market, we see, we saw that eighty nine percent of practitioners, so those running the digital storefronts for their companies, believe that AI will have the biggest impact on customer experience. So that's, like, almost everyone. And that figure comes from a twenty twenty five, master B2B report on the state of B2B commerce. So takeaways here, first, of course, is behaviors are changing and the second is the potential impact is estimated as very strong. And the data point, of course, here is B2B, but I don't think you'll find a retailer or a brand out there that doesn't think that there's going to be a big impact on their digital buying experiences in the short term due to the advances and the changes that are happening in terms of AI. But I wanna keep it real. So I think that the excitement is justified, that but the hype is also massive as well, right? and not all of it leads to tangible outcomes. My next data point is ninety five percent. So last August, an MIT study had everyone talking when it revealed that ninety five percent of Gen AI pilots never made it to adoption. And so you may ask why. Well, if you look at the report, one of the the quotes that I pulled out that kind of summed it up, and I'm gonna read it for you. So it says, "generic tools like ChatGPT excel for individuals because of their flexibility, but they stall in enterprise use since they don't learn from or adapt to workflows." So I would even expand on that and say that they're not really grounded, right, in the organization's context. So not another important call out from that same report was that the successful implementations of Gen AI focused on a single pain point where companies partnered with vendors who deeply understood the domain. So it makes a lot of sense. The study also showed, surprisingly, that with specialized vendors, there were sixty seven percent success rate versus only a third when they, when those same projects were only internal builds led. So for the last data point I wanna offer up to you is seventeen percent. And in B2B, there's quite a game of catch up going on right now. So even the basics aren't covered. And this seventeen percent comes from a Deloitte study that found that seventeen percent of B2B companies are still struggling to make their purchasing easy experiences, and that often includes something as fundamental as product findability. So the potential is real, but there's still work to do. And with that context in mind, I'm gonna pass it over to Seb. He's gonna walk you through some of the latest innovations, and explorations Coveo is doing to help you be successful in your AI adoption. And, of course, we're always laser focused on search and product discovery. So with that, I'll hand it over to Seb and stop my share. Thank you, Sheerine. I'm gonna share my screen. Alright, quick thumbs up if you see my slide. Alright. Perfect. Alright. Let's get started. First topic I wanted to discuss today was AI and Generative Discovery, basically to follow-up Sheerine's intro. And I just wanna highlight that there are several investments that we're making that bring us closer to a future where shopper education and product discovery can coexist in a seamless manner. So our vision basically is to bring shopper education within the search experience where we leverage GenAI to provide valuable answers and guidance, and then using these answers to direct shoppers towards relevant products. While it is possible to deploy Coveo in chatbot like experiences, this is a different approach. So rather than having product discovery and shopper education split across multiple or fragmented interfaces. We believe that relying on interfaces that are already known by shoppers for product discovery will help ensure adoption of these new technologies. So, basically, in a way, we've learned that search or user experience patterns offered by some of the most important platforms or tools that people use every day, such as Google, become expected experiences across the different channels. So our first milestone towards this vision, that moves us towards this vision, was to deliver generative answers within a search experience for knowledge related content. If you navigate to Coveo's documentation page, for example, this is well, actually, we can actually show that here. Is already what we offered. So if I were to search for "Push API", for example, Coveo will provide a generated answer followed by the relevant results that come from the the generated answer. I can also toggle the feature on and off as, you know, it would be expected. Not every user wants to have this type of experiences, but it's there, and it actually has shown important, engagement just within the documentation. So going back to this, that having been our first milestone, we're now moving towards being able to have the product discovery piece that goes along with this experience which is leveraging the generated answer to start recommending categories. So we focused on categories because we believe it's actually the easiest way to actually direct users or shoppers, towards products or categories of products they may be interested in, while offering specific products itself seemed a little bit more risky when we spoke to customers because in some cases, offering one product is not enough, offering multiple products sometimes leads to the issue where bundles of products that don't really go well together, it becomes a challenge. So as a first milestone, we're recommending categories which which already is a step towards product discovery in a single sort of experience. We'll show you a quick demo on one of our own pages. So, if I were to I think I closed it. Well, there you go. So if I were to search for something as... let me go back here. What do I need to start, to get started surfing? So I'm gonna get here the recommendation or the generative answer, sorry, where it's gonna tell me you can use some well-fitted wetsuits, an appropriate surfboard, protective gear, quality surf wax, which is important if you're surfing, you know about this, sun protection and neoprene gloves. And it's also gonna tell me where it got this information from. So, basically, here, it's leveraging some of the rich content that customers already have on their sites, like blog articles, support documents, and so forth to be able to generate the answer. It's also gonna use the catalog data itself to be able to, anchor the answer to what's actually in catalog. And then it's gonna give me the recommended categories. So rather than recommending a single type of surf wax, it's gonna just point me toward the surf wax category page. It's also gonna tell me about some of the repair kits, which are important as well, the surfboards. And you can see that the results are actually quite relevant talking about cases and bags, water bottles, other types of surfboards. It's a I think it's a good display of this vision of providing a single interface for product discovery. We're now entering early access for this. So we already have a few customers who are expecting to leverage the category recommendations this quarter. So more more updates to come based on the results that we get. Internally we've been testing this for a while, so we're very excited to start rolling this out more broadly. So there are multiple ways to actually deploy these types of experiences. One of these is the Managed GenAI applications, I guess, sort of, yeah, you can call it Managed GenAI applications, which is basically the out of the box experience. So, basically, Coveo will fully manage the experience here either through our UI libraries or what is offered in some of our classic integrations. So the experience is more or less prepackaged, and it will look very much or very similar to this. We obviously offer ways to be able to customize the style and the things that you want to use as all of this is composable enough so that you can, you know, remove some things, add the styles that would be desired. But more or less, this is the skeleton and the framework around it. Coveo basically handles everything from processing the query, returning the generated answer, and then returning the results that go along the generated answer. However, there are ways to actually offer these more customized GenAI applications. We have an Answer API that is going to use the same, basically the same pipeline, which maybe they go through enough detail here, but, basically, Coveo internally is gonna take in the query. We're gonna use our semantic capabilities to retrieve passages that rank high on the relevance based on the query that was asked, then we will pass that through, internally through a ChatGPT, and then we'll produce the answer. We'll return the citations, then we'll return separately the results as well that go with that. That's basically what we were seeing on this first screen. With the Answers API, we are able to return all of that via API. As opposed to having customers engage with our UI libraries and have that prepackaged experience, here the Answer API basically allows for chatbots and whatever experience customers want to have without having to care about having their own LLM or their own chatbot or their yeah, basically, own LLM or version of ChatGPT or any other LLM they like to use, but we will continue to provide the answer in the generative style answers already. However, if the customer wanted to use or if you wanted to use your own, LLM, your own experience, basically, and tweak it as much as you like. The Passage Retrieval API is an interesting offering here because it returns the the passages, actually. Instead of returning the generated answer, it will return the set of passages that and it has basically some ranking weights that allow the LLM to decide how to sort of provide an answer or structure an answer for the initial question. And so, you know, there are multiple ways to include this in different integrations. We already have an MCP server, this can be integrated in Copilot as well as eventually Agentforce, which is, we're moving towards that. But and I'll cover MCP in a minute because I know this is a hot topic, especially for developers if there are anybody more "techie" on the call. This is exciting stuff. But, yeah, just wanted to go over basically some of the ways in which the the sort of GenAI answer approach can be implemented today. We've also done some enhancements to the prompt, basically, management. And here, now you're able to basically establish how you'd like to program the LLM, by basically giving it the prompt. And this is something that had been asked for multiple times. Now you can basically tweak it to provide the tone and the type of answer that you'd like to have. This is all accessible through the Coveo Administration Console and now available. There are docs as well to go along with this. Not much more to say about that. So the Coveo MCP Server. For those of you who are not aware yet or maybe just getting into this world of LLMs and agentic processes, the MCP is basically a surface that allows LLMs to interact with services. Before the MCP protocol existed, basically anytime that you wanted to have an LLM interact with a service such as Coveo, you'd have to basically provide or build a set of boilerplate code that would allow the LLM to programmatically start requesting content via API. So you needed to basically create that code that would interact with the APIs. This led to bespoke implementations where every implementation is basically using its own way of interacting with the APIs. So not only it took a longer time to get to value, to get to be able to call those services, but also it led to everyone sort of doing it differently. And this could lead to issues and and mistakes basically along the way. So what MTP does is it allows companies like Coveo to provide a sort of layer of interaction that allows the LLM to simply call a prompt such as search for products or search for content, pass a query, which is the customer's initial query, and then be able to return results very easily without having to worry about coding a bunch of API requests and stuff like that for the LLM. So this is a major improvement that accelerates basically, the workflows, not just for an LLM, but also for the developers working to integrate these kinds of technologies. So Coveo currently has an MCP server. This MCP server is currently intended for experimentation and POC work, meaning that while it is there and it works, it's not yet ready for production environment. I think there will be links to the GitHub. It's so it's hosted in GitHub now. It can be cloned, then set up as a server. Right now, there are three main services that are available through the Coveo's MCP Server, which is the ability to search for content, the Passage Retrieval side of things, and the Answer API. And, yeah, just basically reiterating here, what's available. So if you feel like it, please feel free to try it. We're currently working to have a production ready version of this. I believe it'll be this quarter or next quarter. And, additionally, we're we're exploring having things like the commerce API be available as part of the the the APIs offered in the MCP. Agentic SEO Landing Pages. So today, if a customer wanted to, appear on Google searches and rank on these, they would need to have this thing called a site map, basically, that is an almost nonhuman-readable file that contains all the URLs that link to product detail pages or category pages or any special pages that users wanna make sure that are indexed and corralled by Google. So if, typically, this mean, like, the pages that are offered in mega menus, again, the product detail pages, and anything else. And that's sort of the standard. Then there are also pages that are a little bit more niche that are a little bit more higher-intent pages, what we call, which lie in the intersection of filters or facets. For example, if I'm searching for jumpsuits and rompers, but I want the short, sort of short jumpsuits. There may not be a category page for it, but we know that we can arrive to that by if I were to filter by both of those things. So if I'm filtering for jumpsuits and for the shorter sizes, then that that's more of, a niche higher-intent page. And there is obviously a way to do this. With Coveo, we can generate listing pages for these kind of intersections and have those actually be crawlable and ranked in by Google or more like crawlable and returned by Google searches. So this is already an an improvement over what a lot of companies are able to do today. But one step further than this is to be able to use some of the queries or keywords that competitors are ranking for that you may not be ranking for. Sometimes you're not ranking for them for good reasons because you may not offer those specific brands of products, but you may offer similar products. In this case, you know, if I'm searching for a Dyson vacuum. My brand may not offer it, but I may offer also some, you know, lower cost alternatives to Dyson. And so being able to still enter in that sort of, like, view market share in in in Google results would be important. So there are tools like SEMRush that can identify a lot of these pages that competitors are ranking for, and we're able to use Coveo New Agentic SEO Page Creation Flow to be able to create those pages as well. So the way that the process works is basically we get the keywords from SEMRush, and we put them through a process of Agentic Voting where we have different agents with different personas programmed into them. So, for example, an optimistic agent, which is a user that is a little bit more deciding for the brand, whereas a particular agent or a confused agent or even generous agent. And then we use those agents to basically expand on the keywords that we've derived from SEMRush. And then we use our semantic search capabilities to return products that match the the keywords. And that doesn't need to be an exact match, that's why we use the semantic search approach to be able to return and expand on those results. Then to make sure that we're not introducing noise or not getting products that are not necessarily relevant, we'll go through the Agentic Voting system again to ensure that, again, vetting those results that are coming back for each of the keywords. And then using the Listing Page API, we're able to start automatically generating those pages. So after the voting stage, then we move on to generating the listing pages with the products that actually make sense and, obviously, before automating all of that, there is a stage where, merchandisers can actually view this listing the listing pages that would be created as well as setting up different thresholds and different sort of settings that could be considered during the creation flow, and I'll show that in a minute. And then finally well, this is the approval mechanism, basic. So I'll show very quickly what this looks like. So currently in the Coveo Administration Console, you can go to the search pages and access the Listing Page Generator. Which would be a page that looks like this. And in here, you're able to set different things such as the minimum weight score. So this is the score that agents, the agents that we just talked about, would use to basically score how each of the products actually fits the keyword that was initially provided by the agent itself, from SEMRush. And then you can do also, basically set up a threshold for how many products, basically, the minimum number of products that need to be considered to be able to have a listing page actually be valid. So through this tool, I can actually view sort of preview results. So, for example, for the query, "best inflatable kayak for beginners twenty twenty five", I can see that the semantic search, well, these are basically all the queries that have been generated, so expanded by the agent. And then I can see that some of the products that were found for this, along with their the scores that each of the agents provided for this. It's actually an aggregate score from the different agents, and then I'd be able to see the results. In this case, it's only one result, and we'll see that some of these have zero results. Basically, the agent didn't think these were good enough, with our agents with high standards. But if I scroll just a little bit more here, then I'd be able to see that some of these actually rank very well. And after I preview, after I've done some of the vetting across these, I can go ahead and submit this and have these pages be automatically created. So within the Listing Page Manager, and it's also in the Merchandising Hub, I'll be able to go in and look at the pages that have been created and assess, basically, have a bit of a preview, be able to rank, pin, do whatever I want, within this, within the UI that a lot of merchandisers are familiar with already. And, yeah, this is basically the short explanation of this new feature. We already are working on deployment for this feature. And so, it's still considered early access. As you can see, some of the panels are still very early on, but, we expect to grow this into more embedded workflows within the merchandising experience in the near future. Structured Product Query Understanding. So one of the biggest challenges in commerce, biggest challenges of all times, and this is across all vendors, basically, is the long form queries. And, typically, queries that consider multiple keywords, some of which may not be in the catalog data and in other cases, we're leveraging the attributes of different attributes to construct the query. So in this in this one, so Lipitor forty megabyte, sorry not mega, milligram tablets, ninety count bottle Pfizer. So, typically, the way that search works is by word, keyword proximity, for example, where we're gonna be looking at how in the search query, some of these keywords are actually next to each other in titles and descriptions, and we're gonna use that data to retrieve. The challenge here is that, again, we're using things like metrics. We're using, you know, the number of units in the bottle, and this data may not live like this, structured this way in the in the in the catalog. So the Structure Product Query Understanding is gonna try to sort of break down this query and try to assess how this actually fits within the catalog data that we have and in a way also extended through semantic expansion. So for example, we know the drug name is Lipitor, so we're gonna start looking at the, sort of the attributes that would match it. So let's say the the name of the product. Then we can look at other attributes such as strength and actually match the forty milligrams that exist in such an attribute. So we're able to sort of break down the query and place the query parts into the catalog attributes that match it. Then we're also able to expand some of this. So for example, if tablet doesn't exist in the catalog, we can look at alternatives like capsules. And if that does exist, then we are actually able to start getting, very neat results. So this is currently what we're expecting to have a work in prototype and moving towards early adoption very soon. It's still sort of early, although discovery has been going on for a while. We are very confident we're gonna be able to have something very, very ""demoable much more, much sooner. And I think this is it for that. And I'll pass over to Ben No. If you're still around. I am. Yeah. Thank you, Seb. Fantastic. Yeah. There is some control for that, that is required in the back end. So I get to present the Merchandising section and a bit of the announcement on the integration side with the CP after. But let me start with our first feature, and we'll swap between my slide deck to present these features, and we'll demo a few of them in the actual product. So if you like demos, hopefully, the the live demo gods will be on my side today. First off is the Facet Manager. It's it's already live. We introduced this in the last New in Coveo around six months ago, the functionalities were a bit more limited. So if we look at what's new since, so let me bring you in the the Merchandising Hub first. At that stage, we had mostly the default collections ready for search. We've now expanded to listing pages and managing individual listing pages all from the Merchandising Hub, and we'll get to the second dimension of Facet Management, which is the actual deals, the actual individual facets, what new configurations are available. First, let's look at the default facet collection, which is the fallback to all pages if no specific collection exist. If we look into this one, what we'll find is there are two facets that are pinned at the top, the brand and the price, and then we've configured it so that there are twelve other facets that are defined by relevance. And we can configure some of this, define how many values we want for the automatically selected facets. And the way to pin those facets is simply you can search for all the attributes that are available in the catalog, and you can basically move them over and add them to your collection like that. I won't change the default one. But if we go back to a specific listing page that we would be interested in, I selected one already, in this case, hardware accessories. If we are to, hardware accessories, the chain and and ropes, so very specific ones. This has the default collection for now. We can see our two facets that were pinned at the top, and it's going to select the the most relevant one based on the result set and the attributes that are available. So that makes sense, thickness, length of of ropes and and these accessories, certification items included with those, and your price and different brands. But let's say I'm the manager for this category. I know I've been looking at a bit of the facet behavior and so on, and I know certification is really critical, then I could easily go in the facet manager, set up a new collection specifically for this. So we'll do that. Let's look for hardware. And sorry. That's Just jump through. No. My bad it was this. Let's create a new facet collection right here, and this is where I'm meant to search for my category. So this is the category we're after, hardware accessories, chain, and so on. So I'm going to select that. And select the facets that I want to have at the top. So let's say I still want it to be consistent with the rest of my site. I'll continue to select "ec_brand", I'll move this one over. I still want my promo price, so I'll move this one over as well. And now I said I want also certification to also show up at the top. So I'll say certification, and we'll move these around. I want certification at the bottom. I still want some that are selected by relevance. So I'll say, let's say, I want five more of these with default number of of values to show. I apply this change. I publish it. Now it will appear in my collections that are created. It will also appear if now I go back to the Product Listing Manager, and I inspect this page specifically, I go in facet, I will see that this new collection exists, and I can also modify it from here. Now if we go back to the site, this change I made is going to be reflected right away. So full control and quick and easy to to configure. If we look at the fields now and, for example, prices here, we have these ranges that are available, they're well spread out, the buckets are even and are providing a good diversity of the ranges. But let's say, it's annoying to me that the the ranges are not regular. This is the kind of control I could also do in the Facet Manager, but this is a a field level configuration, right? So if I go in my facets, all fields, it was after the promotional price. I inspect this one. More types are available now than what we presented a few months ago. It depends on the type of your field, obviously. So they're not all available at the same time, but hierarchical facets has been added. And now you get the option to change between numerical slider versus numerical ranges. We also introduced the like I said, the control over the values. There are a few use cases, for that. In this case, let's say I wanted to switch over to custom ranges, I could just simply go, let's see, zero to twenty five, twenty five to fifty and add one range, add range. After, let me go, let's see, with increasing ranges and save that. Apply this, and then go on my page. Let me just refresh this, and I would see my range updated right away. So lot of control. Staying on the on the field configuration side, There are other types of control where managing the values will be very useful. For instance, if we look at another facet that is more of a regular type like the the brand facet, and I look into values, have the option to go in and pin values and decide on exact brands that I would want to show up at at the top. There are two main use cases for that. Like I said, like, promoting a specific value among, like, a lot of existing values. But this is also where for values with very specific sorting, like, let's say, sizes where you have these letters that need to be in a specific order, this is also where you could go in and set this up. So this is all what's new about the Facet Manager. We'll get to what's coming soon to to enrich it with analytics a bit a bit later. Back to the features. Next up is Audiences. This is also already released. It's in early access with a few clients already, and this applies to any of the rules that are created in in CMH. So let me go through a specific page that I prepared. I prepared a nice side to side view so that we simulate a bit what the different experiences would be for the desktop user versus what they would be for the mobile users. And if we go in, we look at this page, and we were to say, I want to create a new rule, but I only want to create it for my mobile experience where the visibility is a bit more limited, there's a bit more scrolling, my top product matters a lot more. How can I affect that while keeping my my desktop experience consistent? So we've made that very easy. Let's go in and just create a rule. Let's say I want to create a boost rule. Let's take something very visual such as the the color. And I'm just saying, okay. I want to boost everything that's that's green in in my category, and let's give it, like, a pretty strong boost so that it's it's obvious. We see which product have changed by how many positions. I'm going to publish that. But first, instead of having it applied to all visitors, I'll go in and edit. There's a few pre-packaged options. So device type is one the login status is also one. In this case, I wanted to change something only for, my device's, mobile. So I'll select that and push this and say green shorts for Mobile. Let's publish this. And if we go on the other side and we refresh that. The rule should be reflected. So if I scroll down here now, my products have changed for this segment of shoppers but on the other side, if I'm to refresh and just do it a couple times, It has not changed for my desktop users. So that's an easy and very visual use case. But like I said, there are other types of audiences that are possible. If you go in, you can customize a few and select more attributes. So device type and visitor status, but what's very flexible and it's already available is targeting parameters that may apply to your URLs, either the current page being visited, so the visitor URL, or the page you're coming from, the referrer URL, which may be for internal navigation of the site, if they transition from one page to to the other, you could target only that audience of shoppers. Or in the case of visitors URLs, it can also be to mirror external campaigns, if you carry those, like, something that's very typical is carrying those UTM tags from external marketing. If it's a social campaign or if it's a email marketing, they get those specific link with with these tags of campaign, you want to mirror those rules on your site. You can easily go in and say, okay, my visitor URL needs to contain, and you input your your specific UTM tag that that you care about, and you push that rule, And only those visitors that land on that existing page on your site will see their experience affected by that. So very flexible, and we'll be looking to extend the the pre-built options as well over time. Next up, Relevance-aware Sponsored Products. Sponsored products is a growing segment and we wanted to make sure that Coveo was very friendly for that and that the control remained in the merchandising in the merchandisers' hands just like other types of merchandising. So what we've introduced that is now in early access, let me now jump over to the other screen. And I targeted this page, so the sandals page on our demo website. Let me actually open it. And I would like to start having sponsored products, on this page. There are two main use cases. The use case at scale is you have this retail media network already configured that may power your in store display screens and all other types of ads, and they offer a way for your advertisers to manage their own promotion, their own bidding, and decide which keywords or which parts of your digital experience they would want to promote their products on. So you have this third party ad engine or some of the ad experience the advertiser experience takes place, and they will want to target parts of your site. But Coveo owns the rendering of all these products. They own the the relevancy of the result set, and they unify the experience with things like the facet, the pagination, and so on. So the question was, "how can we make this more friendly?" So what's new is all the requests going to Coveo can now receive this this list of candidates of sponsored products candidates that would be decided by the third party added. So the way I'll do this in in this demo is just I'll simulate it just like if we had received the list of candidates from that third party. But first, we'll go on that page and give some control to the merchandise. So I'm on on this page. I want to create a a first rule, and now inside pin rules, you will have the option to create sponsored ones. So in this case, the the first step is I'll reserve some positions for the products that will be received at query time. So in this case, I'll say, okay, reserve me three and four. I'll publish this, let's see. Three and four for sponsor. And so I'll see this appear If I go back to my site now, nothing has changed. I'm still not receiving those candidates. But let's say my ad engine started to receive some bidding and have a few candidates. And I'll just scroll down, pick a few that are a bit more visual. Let's see, pink water shoes. I'll say, okay, this is the first product that is sent by my ad engine. What I now see displayed is okay, this product will be in the position that I reserved in the merchandising hub, and it will now receive the the sponsored label because Coveo confirms that this product in this position is the result of what was received at query time. In terms of integration with the rest of the results set, if we're to just start adding more products, let's see. I'll add another one from a different brand. Here, it says brand "Ecco." I'll add this in. So I'll go back up, and now that second position will also be you filled. If I keep doing that and say I add another one from another brand, I add this as well. Now I go back. This is lower in the list. Expect those candidates to come in in the order of priority that they should be displayed, and it will fall back and try to fill up as many positions as possible. And in terms of unified experience, this is where if you start to interact with facets, let's say the the "Ecco" one. It seems to be "Echo" as well. So both of these are still sponsored. But if I start to select, like, the the brand from that third product that we added, I think it was this one here, "Dooney and Bourke." And, or maybe it let's pick another brand. Let's say, let's see. We pick one of those more of a a standard "Birkenstock" style. This should be the brand. Let me add this as a fourth product. So let's say this brand started to also bid on this page, and we go into into Birkenstock now, we'll see okay. Well, the product for that brand still has some visibility at least when the the result set gets more narrowed, and the other are not, no longer relevant to all the organically retrieved product on that page. So it really gets the best of what the ad engine candidates are while respecting the the relevance and the experience powered by Coveo and having all your facets, your pagination, and all that. So really simplifying also the implementation and the integration of those two tools to have the final results set. But let's say, in a different context, you want to get started with this sponsored products operation, but you're not ready with that third party tool, you don't have an ad engine in place yet but you still want to get started and get that relationship going with your advertisers. This is where you could easily go in and modify your rule. And instead of saying, I only want to reserve position, you could also say, well, I also want to define some candidates manually. Let's say we go in here and I want more slippers or more sandals. Let me search for a few very visual sandals. Search for sandals here. I'll publish that. And now I go to my website. If those are retrieved. It seems it may not be letting me. Try another one. Let's say I want those green slippers to be sponsored instead. And take another grey one. I'll publish this. And we'll change our type. We'll see slippers in three and four for sponsored. So now those are done manually instead, and I get my two manually defined sponsored candidates in my result set. So it's a quicker way to get started, and this is available as a configuration for any of the pages globally or locally to a page. So it gives complete control to the merchandiser to say, okay, on some of these pages, I want more control or I need to take action for some of the advertisers manually, so you just go in and select those and let some of the other pages or some of the other positions to be more dynamic and decided by the third party ad engine. So staying on the same team, Spotlight Content is also something that is new. This one is not released yet, but we're getting pretty close. So I can already do a quick demo of that. So if we stay on this screen, I've prepared the a different page, in this case, the the GPS page on on a different site, the engineering side where we have a bunch of blogs and other content that merchandisers may want to redirect the shoppers to for some of the categories and so on. So let's go on that page and, again, create a ranking rule. We'll see that what's introduced is a new type of rule at the bottom here, "Spotlight content", where you'll have to define a few attributes of your assets. Let's say I want to first inject a a content piece, a campaign in the first position. I've prepared two assets, so I just need the image link. You'd probably go in your Digital Asset Manager tool where you have all these marketing assets prepared with their image link, a title, maybe a few attributes of these assets. So let me just paste that in. And right away, I see it appear, it's it's loaded from that pasted link. I'll say, I want to apply the same on my mobile experience. And the second thing I need is I need a link on my page, on my website where I want to redirect those two. So in this case, find your catch, fish finder, buying guide. I want the help, I want to help my users on this page with the buying guide, so I'll just publish that. Publish now, say "buying guide fish finder." And I'll publish this. Now if I go on my site and I refresh and excuse the ratio of my image, it's a bit stretched in this case but easily, I could change my asset and see that reflected on the site. Let's say I now wanted to add a different type of content. I want to add another one which would be, let's say, more aligned on brand awareness or promote a specific brand just like we did with the sponsored products and redirect them to a brand page. This is also something typical that I could now easily do from the UI. So let me pick a second asset. Like this one. So Darmin, our fake version of Garmin, some other GPS. Again, I want the same experience for mobile, and I want their brand page that exists already on the engineering website. So I'll publish this as well. I'll say "promoted Darmin", and I'll publish. I'll see this second rule being saved, if I go back to my page now, I have these two assets that are injected among the result set. I'm not breaking my pagination, I'm not breaking my facets and so on. It's really unified with the experience, and it redirects to the pages I wanted. In this case, it's the brand page. So tying back to the feature that we saw right before in audiences, if, for instance, I wanted to start to customize this page for people that come from this click, then I could easily add an audience on my landing page and say, all those that come from this previous page where they they clicked on my campaign, I want to make sure that it's these new products that are promoted and maybe not my best sellers, for instance, because that's what the the campaign was hinting about, let's say an AI powered fish finder. So I want to make sure that my AI powered ones would now be at the top there. So this is for the combo of Sponsored Products and Spotlight Content. Next, a bit of hygiene, Flexible Targeting is now available in the product, it's in progressive rollout over client orgs, where we're enabling some of the details, for different clients right now. But, basically, what it does is it allows to target groups of pages either by selecting multi values, a bit like you see there, or pieces of, like, parts of their URL or parts of their name so that you can easily target a a group of pages or a parent page and all the children pages. And same for searches where you will target a a single term or you will target an exact query with the full expression. So both are now available. It was available also in the platform before, but, yeah, a bit of hygiene bringing this back in the Merchandising Hub as well. Also, on the hygiene side, a few other details that are now under deployment is things like copying rules. So if you have different properties, different sites, but a lot of the rules are the same, then you only need to create it for one of these sites, and you can easily copy those over. Either one rule or a set of rule and copy that so that it's really quick. So overall, a bit of hygiene in terms of scalable rule management in the merchandising hub. Now in the category of the coming soon stuff, we went over the Facet Manager. I mentioned it, what is the the missing piece still is the performance part so we're coming with these dashboards where you can inspect how each of the facets perform. If you took action and pinned a few of those, you can review the performance, make sure, or segment the experience, and make sure that on specific pages, some of the the facets perform as expected. So you will get global reporting, but you can also, it's very interactive. You can narrow it down to to target specific segment of your site or or specific pages. Also in the coming soon section is sort of enhanced visibility over the ranking. This is an ongoing project. There's already very good ranking visibility in the Merchandising Hub over these different layers of of AI that will explain the the final ranking of your products both on search pages and also on listing pages and to stay up to date with the the progress of the machine learning models, basically, we'll release updates to that. The one you're seeing here is around the listing page optimizer that Seb mentioned earlier, and that's it's been released for a while now. Wanted to make sure that the ranking visibility reflected that. So that that is is ongoing work that should come online in Q4 to really be able to isolate or explain what signals that are picked up from the behavior of the shopper explain the ranking that you're seeing, with things like, yeah, the trends and clicks, the trends in purchases and any of these behavioral interactions, make sure that we have enough transparency over that so that you can confirm the the business sense or the business logic of the ranking that's powered by the AI. Last bit is on, like I said, more of the integration side. This is for the SAP community. A few improvements that especially affect the very large scale analogs. So performance of the the indexing, there's a few things that we we increase some limits and how we handle the different batches of updates that are received and how we queue them. It's a bit technical, but overall, the performance has been, really, improved for that and also around stock availability control. Basically, you can configure property that allows customers to set a threshold for when products are unavailable, so it's more granular control. And improve indexing reliability with retry logic. So the retry logic is also new if something happened while indexing, then automatically without the trigger by human, it will just retry a few times. Thanks, Ben. Alright. I know we're at the top of the hour. I did promise a Q&A portion, so we will go ahead with that. If you can stick with us, then that is great. If you need to leave, we understand. So let's get down to it very quickly here. One question that did come in is access to this deck, and, absolutely, we will send this out to you. We also have a web page that's always up and available on our site called the "what's new" page, and, links to documentation and everything are on there. And it includes not just commerce, but all of our different use cases, lines of business. Alright. So, Ben, quick question for you with audiences specifically. Someone wanted to know if had a logged in user, can they customize to target high value customers or specific clients? So, I guess, using a specific client ID, if that's available or coming soon. Yeah. Well, like I said, we want to bring more prepackaged options or more granular options to the audience flow. In the meantime, the, like I mentioned, there's a few URL options including the the current page URL. This can always be leveraged. It doesn't need to reflect perfectly what the display URL on your storefront is, it's really the URL that is sent to Coveo, so you could enrich this URL with additional information that is used specifically for audience targeting, more in the transition until the audience type that you want is is native or available as a prepackaged option. Okay. So kind of a workaround, you can use the URL to to do that. Okay. Got it. And a couple of questions that came in around, I have frog in my throat today. Around language support, Seb, if you want to specifically around GenAI. I did answer in the chat, but it's good to outline again which languages are supported by our Generative Answering functionality. Seb's still here? Maybe he had to go. Oh, he's coming. There he is. I was about to hop off. I believe it's the languages are listed actually in the documentation, there's a lot of them. I think all the main ones supported, as well most LLM models. I can provide the link to the documentation so that you can take a look. Okay. Great. Even though some may say "beta", they have been trialed and tested by some of our customers. So take that with a grain of salt. They're probably actually quite ready to be used, as is. And then another question that came in, Seb, for you, was when will the feature to automatically generate the category filter pages, so those are the Agentic SEO landing pages they were talking about, available within Coveo. I did answer it's available now to early adopters to contact your customer success manager. That is correct. Right? Yep. That's right. That's right. And then one last question is around the flexible targeting, Ben, for you. Does the customer have to do anything to get access to it? Because you said it was in the process of rollout, so there was a question around that. Yeah. To get a bit more details, it's actually because we bundled that change with the change that we did to the the API that is can be used externally to create or keep your listing pages in in sync with your pin or with another system where you actually create those pages. So the reason why we didn't rolled it out right away everywhere is we want to make sure that all the clients that have jobs that automatically synchronize those pages and migrate their, changes to the new API first be before we roll it out on them. So, yeah, again, CSMs, they've they've been in touch with you, I'm sure, about this, already. Okay. So they can be proactive in reaching out to their CSM if they wanna make sure that they roll that out more quickly. Yep. Excellent. Alright. And with that, it's a wrap for today's content. Thank you. If you're interested in any of those reports I talked about at the beginning, I did leave some links in the chat, but they're also available on our website. And if you wanna dive into Agentic Commerce, and I think, Ben, you had flashed it on the screen previously, we do have a master class coming up on November twelfth at ten AM eastern time, and that will be given by Jason Hein. And that, again, you can sign up directly on our website. So there you go. Thanks. And, thanks to everyone for joining us, and your engagement is appreciated and your questions. And have a good rest of your day.