Welcome everybody to Coveo Relevance three sixty. I'm Peter Curran. I'm the general manager for commerce here at Coveo. Today, I'm gonna talk about how we can turn intent into profit with AI built for ecommerce. I'm gonna start with a personal story. I'm new to hockey. I was kind of a hockey free agent for a few years, but thankfully, Seattle decided to bring a team. So I'm now a Seattle Kraken hockey fan. This is my wife and a friend of mine and I, all at a, a game together. And I don't know if you remember this date, but I wanna take you back to June twenty first twenty twenty four. And if you remember, that was the day of game six of the Stanley Cup finals, between the Oilers and the Panthers. And on that day, I went to Google, and I had a query for Stanley Cup. If you, can imagine what I was thinking was, oh, I wanna know what time the game is. I wanna know where I can watch it. Maybe I wanna look at some stuff relative to the last game, or maybe some news about players or something like that. But, what I actually got was different. In other words, I searched for Stanley Cup, and I'm thinking something like this, but what I got is this. Of course, that's a Stanley Cup too, but that isn't what I was thinking about. What I what I actually got looked like what you see on the screen here, and, the point here is that relevancy is subjective. What one user has in mind when they're searching is totally different from what another user has in mind. And in this case, I was jarred by this experience of seeing a Stanley Cup. That wasn't the kind I was thinking of, but that's okay. Right? It's not the end of the world. I knew how to find my way to, get the information that I wanted. But let's look at some other examples now in the ecommerce space. Here's an example of a query for men's shorts. And I don't know if you noticed, those aren't shorts at all. Those are short sleeve shirts. And that is, go to goes to say that subjectivity, is inherent in search, but subject subjectivity has limits. So you can see here, for example, that they've got a, shorts story banner up at the top. They clearly know that this user is interested in shorts, yet they're showing me short sleeve shirts. So the important thing to take away from this slide is is really just the idea that if you wanna do personalization, if you wanna do more advanced things in in search, and in product discovery, you really have to have search dialed in. So subjectivity has its limits. Let's look at another example. Here's a sir a search for short sleeve dress shirt, a pretty simple query, and look, the results look pretty good. I might not pick that first shirt. The second one is long sleeve, so not quite, what I'm what I'm looking for, but the others are probably in the vicinity of what I would buy, although they're, more than half off. So, so deeply, deeply discounted. Perhaps not going to have my size when I get there. If I'm buying this shirt because I absolutely want it to fit like a glove and it has to look, exactly right for, say, a meeting somewhere in the southeast, then I might now say I'm gonna sort from high price to low price in order to find, you know, a current short sleeve shirt that works for me. And when I do that, you can see on the Neiman Marcus site, I'm getting long sleeve shirts. So now I'm using the features that they've delivered to me in the website, but I'm not finding the products, that I'm looking for. So subjectivity has limits. And if you decide to go ahead and do recall, even if you expand recall and get the right products up to the top of the page, when users start to use the features on the page, they can find themselves in trouble. In other words, search needs to be both relevant, and it needs to be precise. It needs to bring back really the things that, that the user is looking for and only those things. Look at this example. This is a query for cognac leather sofa. A little bit harder query than men's shorts or short sleeve shirt, and you can see that these results are truly bizarre, like acid controllers and battery chargers. This isn't what I would want at all, and I'm sure this is not what Costco wants at all. It asks the question, so if I'm a Costco shopper and I'm accustomed to buying things like toilet paper there, But I've decided I'm going to buy furniture, and I search for cognac leather sofa, and I can't find it. Do I maybe bounce? Do I go somewhere else to find this sofa? What's funny is if you go off the site and you search against only the Costco site using Google and you have that same query of cognac leather sofa, you could see Google knows exactly what I'm talking about. It finds cognac leather sofas left and right. If I go into, the page, however, I can find that, what's really frustrating about this for Costco is that they've not only called it a leather sofa in the product title, but they've also described it as cognac brown. So all of the keywords are there for me to find a cognac leather sofa, but I couldn't find it. Why is that? The reason has to do with the fact that there are all of these other words in that description. Imagine if I searched for the last word in the copy there, finish. That would bring up, this product if I search for finish. So, but if that word brown finish is indexed, when I search for finish thinking I'm going to get jet dry, I would actually get this sofa, which is not what I would want at all. So you can kind of empathize with why Costco is making the decision to search make searchable some of the content and not other content. But search, you know, in twenty twenty five should be able to do to do much better than that. Let's look at another example. Here, I'm searching for coffee, and I've got a coffee table, a coffee maker, and then finally coffee. So what something actually is really does matter, not just what something is like or what it's used with, an accessory or a finish, for example, on the table being called cognac or the the sofa being called cognac. In this case, coffee is describing the kind of table, and these are not relevant results. So as a user, when I get these kinds of results, what happens is I start to distrust the site. I start to wonder, does this site have what I need? And, a lot of times, unfortunately, people leave. If you think that all of these examples are limited to b to c, you'd be wrong. Examples exist in b to b as well. So here, I've searched for industrial sealant, and it's not as obvious as Stanley Cup. But this is an example of a product that there are actually a lot of different kinds of industrial sealant, And, Zorro has chosen here to give me kind of the same product over and over again in different, in a couple of different variations. If you look at their, I believe, sister brand, Granger, and I search for that same industrial sealant query, you get a much better experience here showing, different kinds of of industrial sealants for pipes and gaskets and thread tape and so on and so forth. So this problem exists both in in b two c and b two b, and you can see that companies are coming up with all kinds of different ways of dealing with these, with these problems. Let's look at this example. This is a really tricky one. This is multivitamin with iron. So, you know, someone who has an iron deficiency, you know, would want to take a multivitamin that has iron in it. Then again, somebody with a little kid would probably not want a multivitamin that has iron in it because iron's poisonous if it's taken in large amounts, especially for a a, you know, a small person like a child. So if I search multivitamin with iron, I'm not really thinking too hard about this. When I see that first product, I see that that's a multivitamin. Nature Made is my brand. I might add this to cart and buy it. But if you look closely, you can see that this Nature Made multivitamin for her fifty plus tablets with no iron is exactly the opposite of what I asked for. It is a multivitamin, but it has no iron in it. So keyword search has real limitations, and you can see it left, right, and center all over the Internet. Let's look at another example. Here's a site that does a pretty good job of bringing back what I ask it to bring Barca, but there's a more subtle problem here that I wanna double click on today. This search for women's sweaters is showing women's sweaters. And if I sort from high price to low price or low price to high price, I'm getting sweaters, and they're all women's sweaters. So so that's right. The problem is in that little brown badge that you see in the lower left hand corner of each photograph where it says clearance, all of these products are on clearance. And so the question is, if your search platform is exclusively focused on revenue per visit, it's really only trying to get the customer to convert, and it doesn't care what they convert on, then is that not the road to ruin? At the end of the day, businesses are in it to make profits. And so a search platform, a product discovery platform should be really good at not just bringing back relevant products, but being bringing back products that are relevant to the company, that's offering them for sale. Right? The the products that are full price or fuller price but still have a high propensity to convert. Sometimes people will talk to you like there is only one kind of product that matters to you as a consumer, as if that red sweater is the only one I would consider buying, which is nonsense. It's important for retailers to show users a breadth, you know, a breadth of their assortment and some diversity, not just all the clearance items. So at Coveo, we believe that experience is today's competitive frontline, and you're either embracing AI to fix these kinds of problems or you're going to be competing against it. The leading companies are are solving these kinds of problems with artificial intelligence. And make no mistake about it. It's an arms race, in this space. Our goal is to build a profitable and transparent AI product discovery platform, and we do this, with four kind of main themes in mind. The first is artificial intelligence for shoppers and buyers. So that's what I've been talking about, up until now in the presentation, the different kinds of artificial intelligence solutions to these really common problems. This is our main focus. The second, though, is AI for merchandisers. Some other companies in the space will kind of tell you that merchandising is going to go away. But what they do is they mistake merchandising for fixing broken search. It could be true that AI is going to make search work much more naturally where it understands my intent in the future, and that just means more real merchandising opportunity for businesses. So we believe in building AI for merchandisers that help, make them more powerful, more able to execute on, their business strategy. Third, AI to drive superior business results, and so this is the turning to profit angle in my, presentation title and what I was talking about a few slides ago was showing clearance items. This is where you can start as a business to introduce what your priorities are for a consumer interaction. And then lastly, being composable. So we wanna be able to integrate with your tech ecosystem, you know, whatever that is. At Coveo, we, always try to ground our understanding of the features that we're building and the capabilities that we roll out in terms of the basic equation of all commerce, which is traffic times conversion times AOV equals revenue. And we're always trying to think about how you can keep that revenue. In other words, how the products that people purchase can remain with them and not get returned. We can drive more, traffic using agentic workflows that build landing pages for you. We can convert that traffic with one to one personalization. We can optimize, sorting for products that are profitable and help you take bigger orders as well as recommend complimentary products and upsell products, so that you have a a a larger AOV and, again, help you keep that revenue and trying to do that all at a lower cost, with less operational overhead. We have six basic capabilities that we bring to market. First is we have great search that's very precise. So we can understand the semantic intent of a user's query and bring back relevant products, and then also not bring back too many irrelevant products. In other words, be very precise about what we bring back. Those things that we do bring back, we personalize, so we push the most relevant products up to the top of the screen. Third, we bring more traffic to the site, so we help you really expand your site map without making you look like a link farm. We have an agentic workflow that handles that for you and, make sure that those products that go on those landing pages are hyper precise to that intent. This drives more traffic to the site. And what's great, it's not just any traffic. It's very high intent, typically very long tail SEO traffic coming to the site. Third fourth is conversational engagement. So if you're ready to go into more Gen AI and conversational capabilities or searching with images, things that are more a little bit more advanced, we have all of those capabilities. We have an amazing suite of recommendations that have full business user controls, and then finally, a merchandising interface that, again, allows you to really act like a merchandiser and not just fix broken search. And the last thing that we're really focused on over the last couple of months and into this year is is this thing I've talked about a couple of times, which is ranking for profit optimization. So what I'm not talking about here is simply sorting the results by gross margin high to gross margin low. I'm talking about blending this concept of relevancy, blending the idea of what's attractive to a consumer to buy with the things that are profitable for a business to sell. And we can take that profit at any form. You can pass us some number of of dollars and cents. You can sort of abstract profit into a number, pass that to us, and then we can kind of put that into our learning to rank machine learning models and optimize sort. And if you think that sort doesn't matter, just look at TikTok. Right? TikTok is all about ranking optimization. That's why it's so addictive. We intend to build a platform that is just as responsive to a user's interests as TikTok is. So, again, I'll leave you with that call to action. You need to embrace AI or you're going to compete against it. We wanna arm you with the greatest tools at Coveo and give your shoppers amazing experiences. Thanks for your time today. I hope you enjoy the rest of the webinar. Bye.