Hi, everyone, and thanks for joining the webinar, five quick wins to improve your b two b ecommerce experience. My name is Clara Bollinger, and I work on the marketing team here at, Coveo, and I'm really excited to be a part of today's session. I'm also thrilled to introduce our speakers for today. We have Brian McGlynn, who's the VP commerce at Coveo, and we also have Vincent Bernal joining us, who's the lead solution architect at Coveo. I have a couple of housekeeping items to cover quickly before we get started. First, everyone is in listen only mode. However, we do want to hear from you during today's presentation. So please feel free to send your questions along using the q and a section on your screen. Today's webinar is being recorded, and you'll receive the presentation within twenty four hours of the conclusion of the event. For those of you just joining us, welcome to the webinar, five quick wins to improve your b two b ecommerce experience. Now let's get started. Brian, please take it away. Well, thank you. Really excited to chat with Vince, about some of the the work we've done with our clients and what we're seeing in the the b two b world around ecommerce. And I think everybody from the audience here and what we've seen there, there's no question from time and time again, I get working with, many companies that I've been with and involved with the differences between b to b and b to c. And I I think what we really wanna dispel is the myths, the rumors, and all that that b two b is just trying to emulate b two c. Sure. There are components we see around that. But all said, there there's some very, very fundamental differences that may know very well. And, really, coming out of this, we're hoping that we can just really showcase what we've seen within clients we actively work with today, ones that you can check out in a regular basis, and see where there are ways that we're super serving and our clients are super serving their clients to come on board. So I think we'll start off with a couple of premises that we'll dig into. B two b is complex, and we look at it as well very, very complex, just in in many different aspects. Commerce platforms, whether it be Intershop, SAP, Hybris, commerce tools, Magento and others that have stretched into it, they certainly come through and really evolved on this being that the layer of complexity are much larger from business rules perspective, from the ways that there's fulfillment, the ways of the focus on the buyer in different areas around that. Absolutely. It's just a complex nature when you think about it. And that mirrors the analog world. So you look at whether you're buying an aircraft motor, whether you're buying parts for water filtration systems and all that. Just a myriad of data components, they're very complex. And then we look at as well, we look at just the size and sheer magnitude of it. There is a lot of times on on just the data parts around. So the data can be poor. There's a lot of with interactions and other parts around that. And the biggest thing is we look at it's just the overall component where people are coming in and and shopping. So price lists. For example, in b two b, most clients deal with I have one with ten thousand different catalogs of products that are available in certain geographies and others. These are things that are absolutely just part and parcel of what you need to do in order to make that excellent b to b experience. So it's complex, and it's complicated. And that's the big thing we go into. So there's a lot of manual work that goes in. So what does that mean? Well, one client I was working with has six million SKUs in a legacy mainframe system of which the engineers have to go in who are building the products and update. There's a lot of manual work in making certain data is up to date. There's also as well there's, just the complications that Textacts, as we've seen in many b two b clients, they're using Lotus Notes, they're using other components like that that have gotten the job done in many areas, but still lack a lot of the functionality and and really the expectations. And as for b two c and b two b, there is a big divide as we go into. B two c world, all of us shop on Amazon or Walmart, Wayfair, some of the behemoths that are out there. They certainly not just raise the bar, but they are the Barca what people expect. They expect when you search for something, you find exactly what they're going into. What makes it different in b two b is there's the data structures really are are not necessarily conducive to that. So being able to guess, being able to stem, being able to look at SKUs or part numbers and make sense of that and guide someone to the right item. At the same time, how do you inspire? How do you build a relationship? How do you cultivate that in b two b? That's absolutely an area where it gets complicated, but yet people want that retail like beauty and simplicity that's there. So with that said, I'm gonna pass over to my esteemed colleague, Vince Bernard, who can walk us through really some real life examples on making search intelligent to reduce, manual efforts. So, Vince, you wanna show us a few things? Sure. Thanks, Brian. But first, before going in, actually, you'll see here for each chapter, we're gonna have a title with the subject, which we're gonna dig a little bit deeper with a client example. But just before, since you've worked a lot with b two b overall with different clients and everything, I wanna have your take on that. Why should we make the search intelligent and actually try to reduce manual effort? Is there a reason why it's more important, I would say, in b to b overall? B to b and b to c, here's a major shift in thinking where a lot of b to c, one of the KPIs for a successful site is to look at engagement. So you actually wanna keep people longer on a site. You wanna engage them, have them look at items and potentially upsell. That's a shopper. A buyer professional, someone coming in looking for a particular item, looking for a replacement Barca. Maybe it's a technician that's looking after fifty different coffee stores in a location, trying to find the right kinds of filters and what fits in there, what information. If you look at manual effort, this manual effort from the size of the b two b seller of going in, creating content, curating price list, working with all that information, and trying to put business rules and help people navigate, that's a lot of manual effort that gets involved on the buyer side. So you think about an excellent experience. That technician, they're gonna buy from you because, well, sir certainly, they wanna be competitive. They wanna get through it. If it's piecework chop, they're gonna wanna go in and then get out very quickly, find what they need, order and move on. So a lot of manual efforts are certainly a part around it. Part and parcel to making it work is to search for information or the looking for information. B2C, some catalogs are five thousand SKUs. That's navigable through a taxonomy in many cases. B two b, two million SKUs in many cases. We have some customers where we're talking tens of millions of items that are inside there between SKUs and documentation and support. You need a way to reduce the manual effort of locating that information and as well going through and finding it. So having an intelligent search, one that can use machine learning, one that can look at what persons have put in, whether in understanding someone's, a coffee technician or someone's a barista looking to order beans from a supplier, being able to go through and understand the intentions from that reduces the manual effort if someone's seeking it. Also, the manual effort of if that person can't find it, picking up the phone, calling an organization, and at this point, burning time for the organization that supplies and the person finding it. So that's what's really important when we look at manual efforts. So the KPI of time on-site, we wanna drive that down, and we wanna look at the call deflections. We wanna look at those particular parts to take out the manual items and get people the information they need right away. I think it's, you nailed it you nailed it directly, actually, and that's the example we're gonna share with you, in the next slide. So let's get started with Legrand. Legrand is actually a company that have a variety of brands, but they're and they are selling actually brackets and mounts, and other products, but mostly brackets and mounts for projectors, lightning, etcetera. So when you reach, like, in our website, what's interesting is that they're using all the cool preview features such as query suggestion, automatic relevance tuning. But then I wanna really focus on the DB angle here. So if you reached our website, first of all, you're gonna be able to search like a normal, I would say, human being using words. So you can see here that we have query suggestions suggesting previous queries that were useful for other users, such as a camera ceiling mount or camera mount, etcetera. What I really like about that implementation is that, actually, the same feature that is used to suggest previous queries is also really useful to search actually for SKUs. And and the thing that is super interesting here is that this feature will correct, actually, typos and misspelling or any other kind of mistake you can do in that box. So the same feature that provides a lot of value for our b two c clients is also really useful in b two b where you can use this box as a main navigation artifact on your website to reach different SKUs that you're looking for. If you press enter on the website, you're gonna reach a search page that is driven by Cuveo once again. And then you can see on the left side, for instance, a variety of filters that help let the user actually refine what they're looking for. You can see products with a bunch of information. What happened with Legrand is actually once once they go live, we were tracking with them the traffic and everything that was happening on the website using the usage analytics. And we realized that even with these good features of relevance and and overall machine learning, we had some content gaps. Clients were going on the website, and they were trying to find SKUs that were not in their inventory. They were looking for the SKUs of the product. They wanna bolt in the ceiling, for instance, a Samsung projector. So at that point, what we decided to to build with Le Grand is what you see in the middle here called Mount Finder. If we look at Mount Finder, super interesting approach, we decided to do a reverse search. So instead of looking for a Legrand product, we're gonna look for products that are compatible with theirs. Well, first of all, the user uses a box where you have search features and auto suggestion. So here, for instance, I selected a projector, and then if I I I select that projector, it's gonna filter the next, drop down with the different brands that are offering that type of product. In my case, I'm selecting Samsung. And then after, you can see all the different numbers. And these queues are actually not Leggett on a product. They are Samsung products, for instance. And actually not Leggettont product. They are Samsung products, for instance. And these were the content that we, had identified. So if you proceed to that query, at that point, you're gonna find actually the same Leggettont product. But instead of looking for these SKUs, we're looking for the product that are compatible. Little notes here. It's interesting to see that directly on the result. You see some advanced information such as, order information or or the SKUs number, which is not common, I would say, in b to in BDC, but it's really, really common business in b to b where you really have more information. Well As like Brian mentioned, they are power users. Another cool example here to make search intelligent at scale is done by Acquidi brand. Acquidi brand actually is a holding of different brands that are offering a variety of products. And what they decided to do is to use all these intelligent features that Coveo has to offer. So let's say machine learning ranking, query suggestion, recommendations, and they decided to deploy that at scale. So I'll just shuffle quickly through a few of their brands. You'll notice they have something similar, obviously. Search is really good. And not just search, product listings, spec sheets, and even support, but then they're all using the same stack behind it. So it's a scale infrastructure between FICO and Preveo at that point. And for each brand, you gotta notice that they all have that that really good experience of machine learning search with all the products available, query suggestion, filtering, and so on. So really, overall, an interesting deployment here. As you can see, there is a variety of style, but they're all the same brand under the same umbrella. That closes the first chapter, actually, and now I wanna talk about create easy navigation path. So, Brian, do you have any idea why we wanna focus on navigation path and easiness on b two b? We come back and and we look at so we go in where we might have been before, and we're going through and we're we're looking at reducing friction, reducing the time spent on it. So okay. How do we go through in the electronic world mimic what we've done maybe in a b to b in person sale and other parts around there? And two specific items around creating new paths. We look at upsell, we look at discovery. So these are important parts. So in the sense in case of upscale upsell rather the b two b world, we think about somebody that may be in and if it's a new supplier looking at chemicals for a pool and being able to go and and try something and essentially go in and build more wallet share. There's what may have been a b to b sales consultant, somebody that would go in and say, hey. You know, we sell these items, so we go through from there. So that's an important part around it. The same part around discovery. So this scenario, we're creating new paths. We wanna go in and focus on really what as opposed to just putting a catalog up on the web and saying, hey. Here's a way to do it. Content plays a very intricate role in that particular part. So we think about well, I was with a client of mine in, the life sciences space. And we sat back, and one of the thoughts was, hey. We've got scientists coming in here looking for test equipment. So how do we take that particular search? We're scientists going in looking at test equipment. How do we convert them? How do we get them to where they would engage with us? Well, the whole idea was around knowledge. The whole idea was around presenting content that was relevant to their search, relevant to their entire way of going through it without having to force them to navigate through a morass of, in this case, six million SKUs and different components. So this is where new navigation paths can come in where we look at what would be traditionally product detail pages. We're being able to augment that and enrich that to where there's more information that comes in to really educate a user, fulfill that user, give them that opportunity to go in and sample a bit more. The whole idea as well of leads on steroids is another client of mine talked about. The whole idea that you may have, for example, in this case, they were a chemical seller, and they sold a lot of stuff online for people with pools. Well, what if somebody from the New York Barca and recreation department were to come in and test something or buy something? Well, this is a strong signal that you may have a larger account that's coming in from there as well. So by going into the b two b world and looking at ways to help navigate a person through discovery, finding parts, using ways where you can creatively ask questions about their need and match that to content that's related to it, you can create new path to purchase. And, ultimately, the KPI of a customer lifetime value is gonna get boosted. And very, very important when you look at Amazon b two b and others coming into the market to to get to that and really focus on the consultative selling but doing it through, really through through an electronic means. I think it's a it's a great example again. And when you mentioned a scientist and science product, I think that's really the relation we have in b two b. It's not just customers that are buying something. They're usually buyers, people that are experts in their domain. So the next example here is, Ecolab, actually, b two b portal. If you look at it right now, it may look overwhelming for a regular user, but these are no regular users that are using that kind of interface. There are actually people that are going there day after day to do their job. They are buyers. They are experts in sanitation, for instance, and they're they're looking to do their job as quickly as possible and to have less friction as possible to reach the product they're trying to buy, sometimes in a recurring fashion, sometimes new products, so there is also a bit of discovery. Behind this interface, there is a lot of little things that are, really aiming at power users. First thing you're gonna notice here is directly on the result template, there is an add to quick list, button. I'll come back, on it a little bit later, but this is really important features. Same thing here directly on the template. You see the add to cart. Brian was mentioning try to to to think a little bit differently. So people in this interface don't even need to reach the product detail page. They can just search, filter, add to cart purchase, and go away, which is actually the the best experience you can have in this case. A little bit not standard in terms of b two c, but that's exactly what they're looking for. If we look at two other features, actually, you see that on the right side, there's the quick order feature. This is where you are stacking actually products queue directly, and you can press one one one and then add them to the cart directly. This is not even search on that point. It's just relevance in terms of a UI. So there is relevance for machine learning and product, but there is also the relevance of the interface in terms of what the user is trying to do with it. There is also directly embedded and not collapsed or in the mega menu directly exposed on the interface the cart information where you can see what do you have in your cart to perform actually a transaction as fast as possible. This cart will hold, actually, what's in the quick list for the next visit. So this is all looping back to a better experience overall. They also have the other features we mentioned in the beginning of the presentation. So if you look at the top of my menu, you're gonna have query suggestion, at the top, which is the standard review approach of having popular queries and queries that were successful with other users. But they also decided here to use what we call product suggestion in the same box. This is a shortcut in terms of navigation to go as fast as possible and try to reduce friction again. Then if you enter a query, you can see that the interface will change. Filters on the right on the left side are dynamic. So a lot of feature package here to make sure that the user is happy. I'll jump directly to another example here. This is Famous Footwear. We're looking for work shoes. And this one is is really interesting because the interface is clean looking, super simple. But then if I scroll on the bottom, you'll see that first of all, I'm I'm scoped in a store. Right now, the FX Shop, which is the nearest store from my location. I scroll on the page and I select, for instance, a size, which is a variation of a product, and width. Both of them are actually intertwined, I would say. These are the variation of the products. And then I go up. I can select a specific store. And within that interface, I did not even notice, but I navigated through more than three millions of different, combination possible, each variation of each color of each shoe across each store, which is super complex as overall in operation to do with computer science, but then the user is able to do it without even thinking about it. The last example I have on this one is actually FleetPride, and I just want to reiterate what we already said. This one, again, has a suggestion on the home page. But then when you reach the search page itself, there is all combined here, actually, some of the guided experience we were talking with the Get On where we're reverse engineering the query based on the vehicle. We're trying to fit, in this case, a clutch break. So, again, everything to remove friction and to make sure that I can you can see on the left side, I can ship it to the to to the location I'm near actually as fast as possible. So then, again, really cool implementation here. Here, let's reach now chapter number three, make complex catalog simple to explore. And you can see on the image these kind of transistor and resistance, which are little pieces that all looks the same. So overall, Brian, how can we help make complex catalogs and entitlement more simple to browse for the user? That's a good point. I mean, Vince, you bring up where we look at it. And so complex catalogs for b two b, this is absolutely one of the things that is is really it was different. I was giving you an idea of how different a lot of cases are. I think everybody on this call, people who have been in b2b commerce absolutely get it. Others that are certainly new into it and coming in, this is an area where it's been a stumbling block in many cases to coming on board and clearly something that needs to be looked at. So making them simple to explore, there's, first of all, just the sheer magnitude of going through. So we see, for example, once again, hearing about SKUs where there's three million SKUs that are in there, that's not that's not uncommon. Twelve million SKUs where there may be a case where you have currently five million for sale, twelve million in support, not that difficult to go through. What gets extremely difficult is certainly the the items around it, the derivations around it. There may be a case where we have another dimension and maybe several dimensions of why the products aren't in taxonomy, add another layer of where they're available, add another layer of contractual pricing. So you bring in all these particular parts that make it to where getting it where the end user just needs to ask, okay. I'm looking for a resistor. I'm looking for this particular resistor with this rating, this wattage, this carbon filled, whatever it may be. These are the sorts of things that that, we're really going in, not just looking at, providing facets to navigate, but also providing a means where you can close those questions and go into it. So that's a big piece around it. Because you think about it, what we're really trying to emulate in a lot of cases is someone going in and either asking an expert who has digested, all these millions of SKUs and who understands it to help navigate through or going in and doing a whole fishing expedition where you're looking at it and searching, asking information, asking questions, and ultimately getting frustrated, either leaving or going up to, call someone for a replacement part along those lines. So that those particular items, the vector of complex pricing, price items that are not available in certain catalogs, items that may be in stock, out of stock, items around lead time as opposed to inventory and other parts, these are all things that make it difficult for the end user where they get frustrated and they disappear. And having AI and ML as a means to go through and understand interaction and also being able to look at catalogs and have intelligence about ingesting them, making it visible, these are important. So that's, I think we've been you got some examples of what we've done to make it easy. Yep. Totally. And the first one is actually an international company. So, it's obviously multi language, multi country, multi inventory since each country has different inventory, and that client is actually Formica. Formica had a problem with, overall their product discovery. As the catalog is complex, each color is available through a variety of finish, and these finish are available or not across different countries. So really kind of a and that pressure was actually, up to the user to resolve, I would say. So the user had to make all that logic. So with the new version of the website, we decided to do with them is actually, first of all, to split each country. Since machine learning is taking care of listening to what users are looking for and suggesting it back, each country has its own set of suggestion, recommendation, and ranking. So there's a huge part that was taken care of automatically by the machine. Then the the the really cool thing about that that complex catalog is actually that we decided to people were looking for colors. They were not looking for finish or for products or for SKUs. They were looking for colors. These are designers. These are architects or Barca constructions work, and they were really trying to look at the best finish for something. So what we decided to do is to expose actually a search by color feature, what you exactly see here. So you see at the top in the mega menu, search by color. This is what we see. These are all tiles. We call them swatches. And the problem so you see the sorting option is right now set to most popular. This is ranked by machine learning. So for each query or for each user, we're gonna rank the best result. You can see right now that I when I tested it, yesterday when I finalized the slide deck, actually, I was on the slide, and I was noticing that if you sort by most popular, all the top products are actually white, which is really trendy right now in a home interior decoration. So what they decided to do is for the same UI, actually, they enable a dark mode. So you can compare most easily, and you can have a better view on these colors. Super cool feature. Then if you click on one of these finish or one of these collars, you can reach at that point another interface where you're gonna have more information for a specific series or a specific collar. So, really, the point here was to guide the user through layers of abstraction of the complex catalog to make sure that they're not lost and they're able also to to have a UI that helps them do their job. Another really great example here is Hearts on Fire. Hearts on Fire is selling jewelry, and they had a lot of cool products they wanted to share to share with the rest of the world. Problem is most of the products are almost similar. They're just variations. So what we decided to do with them is actually, enable product grouping. Instead of doing variation, we decided to group the same models of product together. You can see on the UI right now that for each product, each wedding band that is exposed, there are some swatches at the bottom that lets you see all the different variations of the product. If you use the facet to filter, you'll see that right now for rose gold, I'm gonna have all these different models that are rose. And then you can see all the swatches still exposing the other similar products. So within that UI, even if I only have three products, I'm actually showing, like, ten or twelve. If you click on, one of these watches like I've done here for the first one, you'll see that the first one becomes gold. So the user is still exposed to a lot of products and a lot of complexity, but in a really, really clean manner. That's it for chapter number three. Now chapter number four, which I think is extremely important and and relevant, is how to unify the and and why actually unifying the experience. Brian, again, what can you say about the unification of experiences in b2b? I think with b2b, this is really, really what makes or breaks any kind of experiment experience around this. You look at it, b to c world, a lot of it was omnichannel, going through with connecting the storefront into the purchase, buy online, absolutely important. It's been a perfection a lot of that. The b to b world, you gotta look at it a little differently and in a sense of a customer. And a lot of it is, you take, for example, the the importance of a CRM. The b to c world, you'll have customer data platforms, loyalty platforms, and others. When you start getting b to b, a lot of that goes into the the depth of which a customer is defined, not an individual per se, but an organization, which would be a customer with many different roles and components. And you start looking at CRMs and other parts around it, which is really a piece. Well, you you peel that back and you think about an experience. To superserve your customer, you need to know a lot about them. You need to take that all into account when presenting an experience. So for someone logging in to a portal to purchase or repurchase, you need to know about what they own. You need to know about their role. You need to guess about what their role is in context where they're buying. For example, if someone buy who's buying for an architectural firm, are they buying plumbing? Are they buying electric? Are they buying other components? There's that sort of a relevance around there. There's open support cases. There's return rates. There's documentation, especially when you're dealing with technical quest problems. So being able to go through and and really provide all of those items around a relevant experience is really the secret where you increase engagement, where you're able to go through an increased net promoter score, you think about net customer benefit, and altogether get building a long term relationship. Because, ultimately, b two b buyers are there because they wanna get it done. There's an emotional part that's underlying about what the experience is like. But, ultimately, if their job is there to go in and find good information, find what they want quickly and get in, get out, there's gonna be a preference in that case to go into. And that's where we look at unification of the experience. So if there's the initial part going through before a login, making certain that there is the merchandising aspect, bringing somebody in on the early part of the funnel, is after the login engaging with them, looking at reorders, looking at orders, looking at cases, managing it, and bringing all that information to bear from the myriad of sources, whether it's ERP, CRM, ecommerce, whether it's a documentation system, and then a community, being able to surface items around a community, but making certain that every step of where that user is in their interaction experience, buying, trying, rebuying, having problems, that all that information is brought in a relevant component. And that's where an area where we can bring a third vector where we go in as as employees. So we we've been in the business of Internet Internet search. You think about your employees are the first line in many cases to where when it does warrant having a a manual touch or a manual touch point, that they're enabled with that information that makes sense as well. So we look at unifying an experience. This is an area where b two b, it's, it's extremely obvious when you start to peel it back and you look at, okay, there's a lot of different sources of data, and we wanna make certain it's all at the fingertips. And when you're dealing with millions of documents and items, that it's all relevant. And and that's what we see by unifying an experience. I I think you rounded up pretty well, actually. The, and it puts us, Camille, actually, in a really nice position because, like you said, we've been working with services and workforces for years. So, actually, adding commerce on top of that is was simply the right thing to do. The first example here on this chapter is Rio Grande. So, again, in the jewelry section, but in this case, for tooling. So really, BBB for people building jewelries. In this case, I went on the website and I looked for a query for a product, actually, ASIN. I'm looking to, look for product to dissolve gold, for instance, or or to do my job, simply as a jeweler. The first thing you're gonna see is that we're gonna fall back on the tab products because most the bulk actually of the transactional capabilities of the website is really about looking and buying products. On this specific tab, you can see that we have filters such as rating. The more you drill down in the product, the more facets are appearing, actually, so you're gonna have a lot more information if you go deeper. But the point of the chapter is not really to talk about product, but more the support that helps you do your job. So if we switch tab for the same query asset, actually, you can see now that we have articles and videos to show, for instance, how to mix boric acid with something else or then, again, to use, acid cobalt plating solution, whatever. We have a lot of different tutorials and different informational video to help you do your job. And then again on this specific tab, for this inquiry, you have a different set of filters on your left side that will help you, for instance, understanding the topics such as electroplating or findings or torches, whatsoever. Again, if we switch tab, you can have access to safety data sheets. Manipulating these kinds of corrosive products are are sometimes regulated through, different states. So it was important for them actually to share with users all the different information that they need as a one stop shop to make sure that they're able to, as fast as possible, resolve what they're trying to do. Funny fact, you can see on the top left of the screen that they're also so exposing the different gold, silver, and platinum, metal prices, actually. So it can help the people that are building jewelry actually reach optimal profit by buying the same the right thing at the right moment. Really cool feature. Another example of the same, concept is here for Brother. Brother is selling, printers, accessories, and, is also obviously who never experienced a little bit of problem with the printer. You know? So they have a healthy, I would say, support community and resource to help people unstuck paper from their, printer or help change ink cartridges. So a query such as laser printer will, again, bring first products, and you see that they have a tab called all content. So they're mixing everything together, boosting products. But then if you dig a little bit deeper and you go in other tabs, you'll see that we have supplies, accessories, even support article for specific drivers or specific products. And then at the end, they have resources a little bit on, like, on Rio Grande to let you actually, find the information you need to even convert more. So they are using this for webinars, for user group, and any kind of other, meetings that they wanna promote. Really sweet implementation. Again, broader is, multinational, so multi country, a lot of different SKUs in different languages. And they also have two main clients, which is for home or from offices. So interesting use case. Then we're up to our last chapter, actually, number five, automating recommendation. And this has been really something important in b to c. But then, Brian, what can we do for and why should we automate that kind of experience in b to b? So b to b is definitely the sheer magnitude and size of data that that's out there, this is absolutely where recommendations in b two b make a lot of sense and where it's really necessary if you're thinking about it from there. So take, for example, recommendation. How do you how do you go in and look at similar products? Or how do you build a better relationship with the customer? And and, really, if it's upselling. In in the b two b world, upselling being able to go in and look at wallet shares. So maybe a case of carrier distributor and you're distributing different products. Well, there may be in the case of FleetPride and others. There's in house brands and there's others where you wanna be able to recommend that with some logic to try to get people to try something that's either a replacement for an alternative that may have a better margin or a better customer satisfaction rate to look at. Or we may look at as well other items in the b two c world as retail sets. We think about it in b two b, there's components, there's parts, there are related items to what someone's bought that are important to go into. So we always look at where recommendations, traditionally in in the the retail world, a lot of people wanted to manually use these through merchandisers and all that. What we're finding is, really the modern world, automating that using AI and machine learning is is clearly driven better conversions, better retention, just overall better business performance. And in this part as well, we start to look at, once again, average order value or or better yet customer lifetime value. These are things where we can go in whether it's a content perspective or a a buy perspective. At the same time, even going through when someone's browsing or they've actually converted, where they're looking at the catalog online or maybe they're they haven't gone into a session. These are items where making the recommendations, it clearly, there there's a benefit where it's necessary. But when you're dealing with millions of items or thousands of items and you've got very complex data, you absolutely have to get it automated to to really take advantage of of that where especially in the long tail as cases are with b to b catalog, make certain that customers truly get a benefit to see what what business they can give you versus going shopping somewhere else. I think it's it's great. And then again, like you said, the scale is here the reason why we're trying to automate. Trying to do that manually is is impossible. It's as simple as that, especially if you add on top of products, super logical recommendations for for tutorials, etcetera. It just becomes out of control completely. The example here I wanna share is with Beckman Coulter. They are manufacturing and selling some complex, science, products, such as particle counters or gyroscopes. So really interesting products overall, really technical. So then again, you can see here that on the result templates themselves, they have a lot of different information regarding particle sizes, fluid type, calibration options, etcetera. Really, really science and complex search overall. Well done. And over what we realized with them is that each product is actually a series, and each series, has a variety of variants and options. And each one of these options have different support and and manuals, so really, really dense overall catalog. What we've done with them is actually on the first search here, which is a listing page right now. If you click on one of these product, you're gonna reach the product detail page. On this product detail page, obviously, you have a bit of information regarding the product. You can request a quote. These are expensive machines. So you can buy them online, but usually their cycle is a little bit longer where people take the time to go through Procurement and then they they purchase it later on. The beauty of a platform like a relevance platform is to combine actually the information that you have here regarding products and the information of users browsing and using and and making transaction on these products. So we're really able to combine, I would say, how users are interacting with the products and what's the product themselves in turn of the catalog. So let's say in this case, I reached Beckman Coulter. I went on the website, did a few clicks here and there. I found the product that I'm looking for, which is actually this liquid particle pumping system. And then if I just scroll a little bit lower on the page, you'll see that, first of all, I'm gonna have customers also viewed. I think they're really interesting. Obviously, these are but there is another model of the same liquid particle counter. There are some air filters. And in the middle, there is a router package that is something that I might need or not in a centrifuges. So let's just do the YouTube experience of clicking and let us go in that loop of infinite browsing. You know, when you don't need to search anymore or or use the mega menu, you just click on the different products and you're and and you you just go with the flow. So if a user clicks on this, rotor package, for instance, which is a hundred thousand RPM, so quite a machine. Again, you fall in under PDT where you have specification, description. But then if we scroll on the bottom, again, you're gonna find customer also views, related products or parts, And you're also gonna find, if you go, lower in that page, the, technical documentation for each one of these products. So that kind of information and that kind of rich experience is available across every PDP of their ultra large inventory. And this is done by automating, actually, and combining user data with the catalog itself and bringing together these two different entities. That being said, you're really able with that to have a really complete solution that is scalable to larger and and multinational level. That's pretty much it for what we had in terms of examples. So if we go for a conclusion, I just wanna wrap up actually all the different points, the fundamentals to improve your b ecommerce experience. The first one that we discussed at the beginning was to make your search intelligent, reduce friction, and and and make sure that the user it's easy for the user to use everything. Then create easy navigation path. Remember these interfaces that were dedicated for power users actually that want to buy as fast as possible, really, really important. Then make complex catalogs simple to explore by exposing, for instance, similar products, variance and abstracting actually all the complexity of the catalog through a nice navigation flow. Unifying the experience is about bringing support and knowledge articles and even different kind of technical, content with the product to make sure that everything is unified, and then automating recommendation, which is actually the art of scaling and selling upselling with really large catalogs. I wanna do also a small parenthesis on the commerce site assessment report. So if you wanna engage with us, if you enjoy that presentation and would like to know how Coveo would perform in your environment or what can lead you to help your experience, please reach out the, URL that you see here, and we're gonna be more than happy to perform a commerce site assessment. This is actually our experts that will go on your website that will evaluate how well search and overall product discovery work, and then we're gonna provide actually some some path a resolution path for you to to to go further and to have a better experience overall. That being said, I think we're gonna jump to a round of question. Thanks so much, Vincent. This is Clara from Coveo. I'm here just to talk with q and a. If you have any questions that you haven't asked yet, please do so in the q and a, panel, and we have some time to get to it. So please send your questions away. In the meantime, we'll start, with the questions that we already got. Thanks so much, Vincent and Brian. This was such a great presentation. We do have a first question here. So can you tell us, for complex catalog, how are you managing entitlement and product less in Koveo? Sure. So when we started actually investigating b two b, we noticed that most of the people were doing that at at query time, actually. They were resolving entitlement and filtering out, for instance, products or or things you're not entitled to see. We decided to take a different approach in Quivu, and we baked directly in the index these concepts. So when you're querying, you're gonna query, for instance, with a store ID or a warehouse ID or, for instance, your own client ID. And with that, we're gonna return only results that you're entitled to see. So for entitlements, it's baked in directly in our index. And then for, complex product listing, we also now have data structure built directly in search. So you can search for products, variants, or even stores. So we kind of uncoupled these concepts, and they're baked directly in our cloud infrastructure. Great. Thanks so much, Faiza. We have another question here. So you talked a lot about, data during your the presentation. So what data is required to enable all the advanced ML features? So overall, you that for to enable search, you obviously need a catalog and and data on your products. But ML is mostly done on usage of the website and on these catalogs themselves. So there is tracking that needs to be placed. I would say that if you use, like, all the examples that were presented today, they were all using our own JavaScript framework to do the interface. This, framework is tracking analytics, so you don't really need to do much more than just installing the the framework itself. But then the data we need really for tracking commerce is the basic commerce events transactions, such as product views, add to cart, purchase events, and then searches and clicks. These are analytics events that we're sending to Coveo, and Coveo is now able to bind that to user session and journeys, and they're able after that to understand what's skew or what's relevant or not for a given user. Overall, classic tracking in commerce sent to Google Analytics so we can do machine learning out of it. Great. Thank you. A question that came in. I'm not sure if, Brian or Vincent want to try this one. What is the range of expected revenue increase a b to b customer should expect? Sure. So I can take a jump at me. It varies an interesting part. We've seen conversions, and I kid you not, there are some that have been over two hundred percent. And a lot of times, it's really every case is unique. Some have been in the five percent, ten percent range. So we target a few things. There's a conversion increase, and then we can measure a lifetime value increase, and that's an important part as well. So we think about wallet share, what's in there. It it really gets the benchmark of where where we start. So this is where we work with our partners and and and go in or we work with our site assessments to say, okay. Where's our benchmark? Where are we at right now? But these items, we've we've we have a whole business valuation team that we work with at Koveo where there may be a particular customer or a a case where we need to look at a business case that we can sit down and look at where it is so we can make some assumptions and benchmarks around that. So whether it's a a conversion, I think even more so we've seen as well as cost to serve. If we can drop down the cost to serve where it may be as opposed to a sixty dollar per order processing by the time of a person going through, faxing, or getting on the phone where we can deflect a call, reduce cost, we can go in and we can look at a myriad of different areas around that. But, it it's a case with any benchmark. It's really it really comes into where it starts. But ten percent, I think there's certainly areas around it, but I kid you not, there are customers who seem to two hundred percent. One of the challenges with it where we spend a lot of r and d in our company's attribution around that particular part, we wanna be credible. There's a lot of discreditable information out there to say you were magically gonna give everyone a twenty, twenty five percent bump on the revenue. It'd be irresponsible for me as a business leader to say that. What I can say is that we work very we have metrics in our own dashboard to look at search influenced purchases. So can we truly track to see not based on, okay, you increased your marketing budget for x and therefore you're driving more at the top of your funnel. We're now looking at it to where we can say, okay. To what degree has a search, has a recommendation, has an AI driven PDP or PLP driven conversion or driven a purchase? And then we can also look at the velocity based on that customer over time to see have they repeated and have they come in and have we seen a larger trend in what they're purchasing through growth around that. So some very important metrics that we've seen along those lines that are out there from that part. But, again, I do like where the question comes from and we're more than happy to go through on, on Orca on that as well. So, Clara, back back to you. A lot of questions coming in. Yeah. Thanks so much. Such a great answer. We do have some other questions that came in while you were entering, so I'm gonna hop right on to the next one. How many non catalog data sources can Coveo connect to, and can you give me some examples? Well, Shannon, on this one, actually, you have a limit in Coveo cloud. That is increaseable, and it's all about licenses. So, theoretically, you can index as many content as you'd like. That's my that's my short answer. Then there there's business and entitlement, but overall, technically, there's no really limit. The connectors that we have are built in through a few solutions, such as YouTube, Twitter, Dropbox, Drive, etcetera. But then we have generic connector that can index everything. We're talking here about web connector and sitemap connectors. So if you have existing documentation, technical PDFs, or anything that is somewhere outside in the world, we can index them using our generic connector. These connectors are flexible and are able to index almost everything that is on the web. Then if we go in a little bit more complex use cases, if you have, for instance, a server or a database that have full of technical documentation or support cases or whatever the solution you're trying to index, we have dedicated generic rest connector or even on premises database crawler that can pull the content in Coveo Cloud and make it searchable. So in terms of different sources, I would say there's no really limit. The hard limit would mostly be in terms of maximum documents you can have in Coveo. Right now, our largest b two d client combining support and commerce and website search is at about eighty million documents. So there's plenty of space in Coveo to index a lot of different items. Additionally to that, I would say that we have built in solutions for most of the, largest, CRM that are out there. So if you want to provide a unified experience using support, website, and commerce, you can also use these connectors that are for Microsoft Dynamics, ServiceNow, or Salesforce that will bring, actually, case deflection and support directly to your interface. So if you need more information on that, please contact us, and we'll be more than happy to actually, help you, find the right the right entitlement for your needs. Back to you, Clara. Thank you. Other question for you guys. Do you envision the amount of time client staff spend doing merchandising for crust and upsell relationships can be reduced or removed with Coveo product recommendations? You know, it's an interesting thing. Big question. Big question. Actually, we had a a, we we'd had an event, and we had some white papers as well. We've talked about one of our customers, actually, Acuity Brands talked about that, about the benefits around it. So short answer is absolutely. That one client that had sixteen people that were focused on looking at search logs, looking at behavior, looking at all those items to set up and do merchandising but allows to do a lot of the mundane things we were able to eliminate and and and automate that. So that was a part that was important. We we were absolutely able to do, on reducing the staff that were necessary to do it. So what it meant was they could focus on being more strategic. They could focus on, okay, what are we gonna do as far as, increasing data to where we can look at, what basically, what's what's converting better? Can we look at specials? Can we look at margin data getting that particular item? What can we do to further the signals that are put in there? Because, ultimately, when you're dealing with thousands and millions of different interactions, you've gotta learn from that and be dynamic around that. So when we look at business cases, those are things we take into account. Are there savings from the current way? And the other is even unplugging legacy tech. We see a lot of stuff that's on prem that people are caring for and feeding that maybe millions of dollars a year that are going in. So by going to a pure cloud solution and a cloud scenario, we're in a position where we can go in and reduce a lot of with a manual cost. So it's a cost reduction. Absolutely. The interesting part is look at a customer like QE brands cited, not too long ago when they were talking to, to some folks. It was a case where there's an increase because the machine learning and AI, when properly used with data, properly used with, the collective intelligence of an organization, can outperform what a lot of the manual rules are because you really can't guess what people are gonna do. They're gonna interact. Things are gonna change. Personas are gonna change. In order to meet that, there's, there's a benefit. So it's twofold. There's absolutely a benefit of higher lift, higher conversion, better customer lifetime value, and you're also gonna see a reduction in cost around that as well. Great. Thanks so much. We have other questions, to answer. So, bear with me, Brian Vaisant. What are some challenges that customers face that hold Barca search or recommendations initiatives? I think the one of the big ones, we'll go back to a customer story, b2b specifically data, where it lives, the volume of it. We've clearly shown a search is absolutely I mean, the technology is more than search. It's a platform. It's an AI customer platform. It's absolutely necessary to to provide that experience. Clients get that. When they see it, they see it in action. They understand that, yes, they're having a brain on your system and a and a brain that puts relevant information at relevant points. There's no question that the benefits there. Biggest stumbling block, a lot of it is, okay, our data is in a mainframe. Our data is in Oracle database. Our data is in an access database. How do we what kind of a project do we need to go into it? Do we need to get engineers to go through and start reviewing all the spec and start to sort it? Realistically, there's two things. You can go through and spend years going and tagging and looking at data. Better part is let machine learning go in and look at it from there. So we spent a lot of r and d money on our side in looking at not just the problem of interacting with the user, but interacting with the user created content to make sense of it and ingest it. So by looking at using AI and ML to go in and understand the data better, we we helped accelerate a customer get onboard and online in literally, what was a twelve week project dealing with legacy data that they hadn't yet put into their PIM. They actually put their PIM project after the search and recommendation project because we can get identical results from a customer perspective. And I'm not gonna dispute the the value of PIM. Absolutely, it's value of PIM, but what we're able to do is improve the customer experience within an an eight to twelve week window by indexing content, in this case, a Microsoft SQL database, and putting it in to drive the customer experience from a support, from a purchase, from a service, and all those aspects while the broader organization of data can take place on a separate program. So you're not wasting time from a, a go to market perspective, And it really takes that object the objection a lot of our customers face off the table saying that we don't have to go through a massive data cleansing prospect before we get value. We can get value right away. And and that's been what I've seen really been the the top item about pause. Are we ready for it yet? And our answer is you probably are a lot more, and we've we understand this. We've invested a lot more to help clients, get over that hurdle to to where they can get value with their customers. Thanks, Brian. One last question I see here. Is Coveo available in multiple regions, and does it support multiple languages? On that front completely, actually. So we rolled out earlier this year, our platform in different region to cover data residency problem and not problems, but redundancy requirements, I would say. So now the platform is completely deployable as the data residency fully hosted platform in Europe and in Asia Pacific as well, additionally, to our original North American platform. So this is really for data hosting and speed. It's the same Coveo cloud infrastructure. You don't need to install Coveo is a SaaS platform, so you don't need to install or or to or to deploy whatsoever. Then in terms of languages, interesting subjects. So Cuvelo was a website and enterprise search before, so we have, obviously, baked in directly the index and the relevance support for many, many different languages. I think we can support up to thirty different languages in the same implementation ranging from French to English to Chinese to a different, to to to all different languages. What I mean by support is that, actually, you're gonna have full search, but also stemming, highlighting, did you mean. All these advanced features actually will be totally, accessible in different languages. Just bear in mind that each document in the different language will inflate the number of documents that you have in your index. But still, like we said, we're good for up to eighty millions, so it shouldn't be a problem at all. So short answer, yes. We're totally compatible for multiple regions, multiple language, and really high scale, deployments overall. That's it for our. Thanks so much, Versa. We're now out of time, so I think now will be a great time to wrap up the session. For those of you who had a question that was not answered during today's session, we'll follow-up with you directly. On behalf of Brian, Vaiza, and myself, I'd like to thank you for attending today's webinar. Have a great day, everyone.