I'm Peter Curran. I'm Coveo's general manager for e commerce and happy to be here with you all. Is this advancing? There we go. And I'm Lauren Amundson. I'm with US Foods. I've been with US Foods for nine years, seven years in digital. And if you know anything about what US Foods has been doing, we've been really striving in our technology. So I was fortunate enough to be the very first merchandiser to really build out our strategy, as well as the first product manager. And then in the last year, I'm over as the product owner. All within those seven years has been focused on search and merchandising. So really excited to share the transformation that we've been able to make with Coveo. Awesome. So why don't we start there? Tell us about US Foods. Folks may not know the name or maybe they do. Yeah. Well if you haven't heard of US Foods, I guarantee when you leave here you're gonna see our trucks on the streets. We are a food distribution company. We did just about thirty eight billion in annualized sales. We have over two hundred and fifty thousand different customers. It spans through many different segments. So you have your local customers. Think of your mom and pop shop. And then you have your national customers. Think of your big national chains, as well as hospitality, education, etcetera. And we are located just out of Chicago in Rosemont. Okay. And so if I am a US Foods customer like a restaurant somewhere, how do I buy something from you guys? Yeah. So really exciting. So as I was mentioning with our digital transformation in October of twenty twenty two, we launched our e commerce application called Moxie. It stands for Making Operators Experience Easy. And really, this is a one stop shop for all of our users, whether you're internal employee as well as our customers to really help with product discovery, managing orders, payments, etcetera. So it's a really great application. And fifty percent of our usage is actually through mobile. And when you think of two of our business, we have a lot of sales team out there. And they're not going to want to lug around their laptops. So having everything at the palm of their hands really makes it easy easy for for them. So forty billion dollars, half of it is going through or half of it is through mobile and all of these orders are coming through through Moxie. And this site isn't old, said it's only a couple years in? Yes. So so what was the the transition like going from like a non digital type organization into into what you are now? Yeah. So it's it's been really great. So as I was mentioning with managing the search engine, when I first started, literally my job was fixed search. And as I was investigating how I go about doing that, come to find we only were indexing product descriptions. And if you know anything about B2B and food industry, our product descriptions are very technical. They're very long tailed. So for example, we would get feedback. You know, I'm searching sugar, and I get margherita mix because in the product description, it says sugar free margarita mix. So we were getting a ton and ton of complaints. The other thing too, because we weren't using a modern day search engine, there was no such thing as autocorrect. So we would get a ton of feedback where, you know, for like misspelling something, and and you would land on a null results page. And our customers would get super frustrated. Like, give an example, Worcestershire sauce. I I couldn't even tell you how to spell it. There's many variations of how to how you spell that. So it was really challenging, for our users to to find those products that they that they were looking for. And the other thing too, you could even see on this example, on the screen, speaking of fifty percent using mobile, that that example with rolled, how close the d and the s is. So, you know, when users would do that, they would land on these nulls. And in order for us to make any sort of changes, it either had to go through a code release or we had to make manual manual changes. So it just took a really, really long time for us to really fix fix our our search our search engine. The other piece too is we have a huge, huge catalog that spans across many different types of products. And when users would search something like apples, because we also carry applesauce and apple juice and apple pie, those products would show up. And then again, customers would get frustrated because they're like, I just want a fresh apple. So we really, you know, needed to get a more modern tech, search engine, especially because we, deployed our application in twenty twenty two, and we were still using our legacy search engine. So really excited with the work that we were able to do after conducting an RFP to land land with Coveo. Awesome. So tell us a little bit about sort of what happens when folks can't find what they're looking for and how you go about solving this kind of problem. Yeah. So we have a really great network with customers as well as our seller community where we, you know, talk to them, gather feedback to really figure out what are those pain points for us to accommodate within our roadmap. And I will tell you, you know, as we went about with Coveo, you know, obviously AI being new and trusting it. When we first deployed Coveo, we actually mimicked a lot of the same static rules that we had in our former search engine. And I will tell you, it failed. The relevancy was all over the place. And after we took out those rules and really start to kind of hone in on what really matters to an AI search engine, there's a couple of things. One, your product data. So making sure that you have accurate product data. Two was just trust the AI. So the way AI works with machine learning and all the click behavior, we wanted to make sure that we had all that proper tagging in the database so that those relevant products would would show up. So that was definitely a big a big learning for us. The other thing, too, is with our product data, you know, kind of going back to you know, originally, we were just using product descriptions. Come to find in our PIMS or our product database system, we didn't have any category attributes. So that was another big, huge transformation project for us to be able to work with all the necessary teams to be able to gather all of those category attributes. So that way, when you go on a new search, say, and you have all your facets, it's really critical to have all that data. And then with our taxonomy, we had come to find that our taxonomy was also not customer friendly. So we had a lot of different categories just pushed into one. So to give an example, we had aprons, gloves, and caps all within one category. So someone would search gloves, and they would see aprons and be like, that doesn't make sense. So because of all that learning, we actually did a whole taxonomy transformation project. And kind of going back to what you were your original question about, you know, how did we we had to work with our customers to be able to figure out what is that right taxonomy for us to really help drive that whole product discovery experience, and then really leaning more into that AI. Because as the AI learns, with behaviors and pairing it with your product data, we found a lot of success with that. Yeah. That, number three, like I remember one of the critical things as we got towards launch is is just making sure that all of the behavior telemetry, like that stuff that we're watching about user behavior on the site was accurate and that it really reflected what people were doing. And then that's the sort of the gas in the tank of the of the artificial intelligence to move things around. So cool. So this is all the stuff that you did to to sort of get implemented and get live. So what does it look like today? Yeah. So today, definitely a lot better product discovery. So we had deployed our local customers onto Coveo back in August. And we are getting just great feedback on that, which is awesome. The other piece that I am very happy about is the less manual effort on our dev team. We could speed to market. You know, we see something that we need to tweak really quickly, I could go in there, or folks on my team could go in there and make those changes, which is awesome. Because if you know anything about the development sprint cycle, it could take weeks, if not months, to get a simple, simple change implemented. And then the other piece is just the continuous learning. So I mentioned, we deployed. We went live in August. That's only a few months ago. So you know, the system continues to get smarter as users are using the search People don't want to do boring work, right? They don't want to fix stupid problems that are annoying, right? They want to do interesting work and sort of automating a lot of that stuff and giving them the tools they need is important to making people happy at their job. And so okay. So things are better and that may be good enough for some people, but other people like to measure. So tell us a little bit about some metrics and I I Since we've deployed in in August, we have seen a three percent increase in our conversion rate compared to our previous search engine which has resulted in over one point three million in incremental case sales, which is great. The other piece of it, too, is we saw a huge reduction in our nulls. We were previously around a four percent. And now with Coveo, we are at a one point eight percent. And a lot of that is driven by the autocorrect feature, because like I was mentioning, we had a lot of food. Food is very hard, hard to understand all of the proper spellings for that. So really grateful for that feature to really help with that. So let's talk what's coming next and what else does it look like here? Yeah. So as you heard me mention, so we have our local customers as well as our national customers. So national is the next wave that we will soon be getting rolled out. The other piece is continuing to focus on personalization. So I mentioned a lot about that behavioral data, but how do we take that another step? The great thing about being B2B is that we have a lot of that customer attribute data to really know who that customer is. And we know from our analytics that a Mexican restaurant shops very different than an Italian restaurant. So how can we hone in on those customer attributes to really start tailoring the users' experiences, which then goes into the next thing that we are focusing on is around content discovery. So right now we are focused on when users search, we give them the proper products. But we have a whole robust assortment of just different content. Think of recipes, articles. So how do we incorporate that within the search experience and then also tailor it to those different customer segments? So think of an example where you have a bar and grill, and they're searching beef patties. How do we show them the beef patties, but also curate content to show them how can they make a better smash burger or whatever that may be with all of that robust content that our marketing team has created. And all within a catalog that is custom to them. Like there are certain things they can order and other things that they can't order and each customer has its own price. Right? So all of that stuff needs to be kind of woven together, doesn't it? So not just through normal e commerce but through more conversational type experiences. So great. Well, that's the content that we have and we'd love to hear any questions anybody has. So thank you so much. Very interesting. You made it sound as if it's very easy to do such a thing, but I know it took you several years to put everything in place. My question is to be able to have a good search engine on the platform, you need a good data. So I'm interested to know, like, how much effort do you put into collecting those images, the text, the content that you actually read from? Too much. A lot of effort, actually. And we had uncovered that the way our PIM system was set up, a lot of when a new product would get set up into the system was through spreadsheets. And unfortunately, it's really hard to scale that. And then two, if you're not checking miss pellings, it could result in really bad data. So the great thing is we have a dedicated product team for product content, and we're in the process of implementing a vendor portal to really enable our vendors to go into that portal and give us that content that's very, very necessary for them to drive it on our commerce site. I think the thing I think about as it relates to data is that people get into this trap of thinking, well, I need to fix my data, then I'll do AI or I'll do search after that. And I think that's a trap because data's, you're never really done, right? Because the products are constantly changing and the needs of the consumer and the things that they're gonna ask about are constantly changing. So data is an evergreen problem that you're constantly working on. What else? Other questions? No? People being shy? Right. Sure. Sure. Are are you doing anything related to, GEO, Generative Engine Optimization? Because this is only when people come to you, to your site, to your application. Do you do anything when they are not on the site? Yes. So if are you talking about, like, acquiring new customers? Like, how do we drive that traffic for them? Can they find new product? Not yet, but soon. Yes. Soon. We will be opening it up to yeah. So typical in B2B for everything to be kind of behind a login and, because they're they're dealing with contract prices and things like that. So there are some things that you can see as a as a just a Joe Public type user and kinda getting their content in shape. You have to keep in mind too with these guys, digital is super new. So they're they're transforming a a forty billion dollar business pretty rapidly here. So alright. Good. I'll be with you. Oh, there's somebody. This may not be a super relevant use case for you, but what about other languages? Yes, it French and German are a big expansion area for us, and so search obviously is very different with intent in those markets. Germans only want the facts, etcetera. How is that really part of your overarching strategy, and how has the tool helped you grow there? Yeah, so localization is something that is currently on our backlog right now for search. So if you think about a lot of our customer bases, because we do have Chinese restaurants and Mexican restaurants, etcetera, and a lot of those owners of those companies don't really quite speak English. So localization is something that we really are it's on our backlog to to do. So, yeah, it does does resonate with with us. And I can say just as because of search provider that, you know, you have to have language specific capabilities. There are components of the tech stack that you can plug in to support German, but really supporting German, really supporting it properly in a modern e commerce experience requires more effort than just being able to return a word when somebody searches for it. And as you get into other languages which have diacritic characters, for example, that are going to appear in facets, these kinds of things tend to break a lot of search platforms. And and so getting that kind of stuff right is it requires a lot of time. There is a lot of of of translation type stuff that is obviated by LLMs are, you know, understand multiple different languages. And so you can ask a question in Japanese and based on English content, still respond back in Japanese but still cite the English phrases that the that the that the content was generated from. So that's super convenient. Like, so that's a much better world than we lived in, you know, five years ago, but still getting those language nuances is it requires extra effort and specialization. Hi there. Thank you so much for today. All of this sounds like the right thing to do, but in any organization, maybe old guard or traditional people, what kind of roles pushed back and how did you deal with that change management? And do they feel better on the other side of it? Because Moxie sounds fantastic. We two cocktails talking about this last night. Yes. Not three. Yeah. No, yeah. So building trust and establishing a solid business case is really what we had to do as we were trying to think of a new search engine. So with US Foods, as you probably saw on the slide, we're one hundred and fifty plus years old. We have a lot of ten year employees that feel like they know this business ins and out. And it's really difficult for them to kind of hone in on our technology. So we've had a lot of challenges with that. But what's been really great is when we deployed we created a, we call them Moxie Champion Group. So it's our like top sellers across the the nation. And they are supposed to be like the pioneers for us to be able to, you know, we share and communicate our technology to them. And they go out, and they share it to their peers. They share it to their customers to really help drive that adoption. But I will tell you, it's challenging. It's very challenging. And that is kind of why, too, we had why we already had rolled out the new search engine to our local base, and we haven't yet done national. Because national, they're very particular about certain things. They hate change. So we have a lot of fun coming up as we go ahead and roll the new engine out to them. Yeah, I think from a technologist standpoint, when we're in that scenario and we do work frequently in B2B where companies are transforming a little bit later than you saw in B2C, the numbers matter. So we build business cases anytime we're working with a prospect. We go back and measure to see if we met the case that we said we would we would meet. And if we're not and sometimes we're not, we have to do some tuning in order to kinda get things to to the place where the business is satisfied. Nothing, sells like like numbers and commitments like that. So that I'd say is another key. Good. Other questions? A little hard to see everybody. No. Alright. Well, it's time for your two cocktails tonight. Thank you guys for coming. Thank you. Thanks. Good job.