Perry Ellis International: Operationalizing AI Across a Portfolio of Brands

[00:00:00] Debra Sabourin: A lot of the time, too, if you're building something with AI, one of the things that we always kind of have to keep in mind is, you know, the why. Why are we building this? Like, if it's just to say that we use AI to do it, it's probably not the best use of AI. Some tasks, they're just not built for it.

[00:00:17] Kailin Noivo: Welcome to another episode of The Ecommerce Toolbox: AI and Retail. Today, we have Debra, who's the director of ecom operations over at Perry Ellis International. So welcome, Debra.

[00:00:29] Debra Sabourin: Thank you so much for having me.

[00:00:30] Kailin Noivo: Cool. I always like to start off by learning a bit more about our guests. So, why don't you talk us through your career journey?

[00:00:38] Debra Sabourin: Sure. I've been in the ecommerce world for a little over 15 years. I have worked mostly with brands, which is sort of an interesting arc when it comes to ecommerce because there are so many retailers, but I have been focused mostly on the brand space. And I started, really, I started at a smaller company called We Got Soccer a long time ago, a local company in Massachusetts and then I started working at Samsonite. That was a very large journey for 9 years, where when I first started with them, they had one website, and by the time I left, we had 6 owned-brand websites and two additional ones that were run off-site, and they had purchased almost 8 or 9 brands. So, that was a really large park there. Then I moved to Bose, where I was in charge of Bose North America ecom sales that included America and Canada. And then, I am now at Perry Ellis, and I've been here for about two and a half years. A lot of operational work, I tend to be very left and right-brained, which means that I have a creative problem-solving sort of a knack. If nobody's thought of it before, I can usually figure out a creative way to come up with an answer.

[00:02:07] Kailin Noivo: For those that don't know what, like Perry Ellis, I think of the brand itself, but it's like a kind of culmination of brands at this point. So, maybe talk a bit about that.

[00:02:17] Debra Sabourin: Sure. Yeah. Perry Ellis International is the parent company of a bunch of different brands owned and licensed. So, internally, we have 6 in-house websites that we run. So, Perry Ellis, Original Penguin, Cubavera, Callaway Apparel, which is a license of Callaway Golf, a golf apparel shop, which is a sort of retailer functionality where we sell multiple different brands, licensed brands and owned brands. And then we have Rafaella, which is our one-woman brand, which is a smaller brand and supports a lot of wholesale. So, it's a lot going on, a lot of different demographics, a lot of target audiences, and so we try to keep our tech stack similar across all of those to kind of help us with all of that juggling.

[00:03:08] Kailin Noivo: You know, that makes sense. And do you guys operate the ecommerce operations for the licensed brands or just the owned brands?

[00:03:17] Debra Sabourin: Both. So, my department is both, or is all of those, so that includes the owned brands and Callaway Apparel, which is a licensed brand.

[00:03:28] Kailin Noivo: Really excited for today's episode. We're gonna dive a lot into how folks are using AI to operationalize, and I think what I would really, really be curious about on your side is how are you doing it, kind of in a parent company, right, that actually operates multiple brands. I speak to a lot of individual brands, but not so much for kind of conglomerates. So, yeah, I mean, you've mentioned that D2C businesses are really special because you can kind of pivot faster than traditional brick and mortar. Once you've kind of built in flexibility, maybe talk to me a bit about what that actually means. How is the D2C business traditionally more flexible? What are the things that need to be true for that to be in place?

[00:04:11] Debra Sabourin: So, I guess for me, when we're talking about operational flexibility, it's basically just a way of saying that the systems that we've built can scale or we've tested them, they're not brittle, and this is traditional basically in my career that we usually operate with a very lean team that has to do sort of, you know, a giant's worth of work. We can't have heavy lead times like we see in traditional retail, so brick and mortar. If you wanna change a promotion that requires a lot more people and a lot more steps, but when it comes to ecom, it's almost like, you know, flipping a switch. But we have to make sure that our systems can handle all of that very quickly, and there isn't a lot of, you know, sync time or downtime with any of that and that our inventory systems are accurate because that would be the worst, especially around holiday time. So, we make sure that we test all of that heading into our really heavy time so that we can pivot when necessary. And we're always testing new things, so we know which levers we can pull.

[00:05:15] Kailin Noivo: It's really cool. I know that you're named in the PEI AI champion this year. Maybe talk to us a bit about honestly, what I'm getting a lot of questions about, and we just did a webinar a couple of weeks ago on how to deploy AI agents in your business, and it seems like every conversation that I'm having, people are looking to deploy AI agents in their business. And, obviously, the answer is different for everyone's business line. But for ecom ops, what are some of the types of AI that you're actually implementing that is not hype, that's actually getting true results? Talk to us a bit about that.

[00:05:51] Debra Sabourin: Yeah. So, the AI champions is actually a great corporate initiative that Perry Ellis International created internally to have representation in every one of our business departments so that when you say we wanna, you know, tackle AI or we wanna start incorporating AI adoption into the, you know, fabric of the business and figure out where we can operationalize that, you have a stakeholder from every one of the businesses and it's basically the AI champion's job to understand the tools that we have access to and learn all of the new functionality so that way you can take that and roll that into your group. You become the subject matter expert on all of those things. And it was a 6-month process. Every month, we had a mission as an AI champion, like, these are the things you have to create using the new tools that have come out, and then you have to share those out, and just to make sure that it gives you all of these ideas of, wow, this is amazing, this is what I can use this for. When we talk about operationally, both on the frontend of site, and back of house, it helps us analyze our business, things that would take, you know, days to analyze complex datasets from a myriad of different places when it comes to like a Google Analytics or our internal systems. We can now analyze that. Of course, you always have to make sure that you're verifying the information, and that's super important nowadays. There has to be some sort of human in the loop who's verifying, especially when it comes to data, so that they're not hallucinating some random answer. But it also opens up your time to be more strategic and less in the weeds. So, my team has opened up quite a lot of ideas, just being able to get to those data points that much faster. And on the frontend, there are lots of, you know, tools that have come out where with personalization and all of those things, and we're testing, we're testing a couple now. We can't get too in the weeds about what we're doing, but it's all really exciting.

[00:07:59] Kailin Noivo: Wow. That's really cool. So, it sounds like you have kind of, your department has led, like, the consolidation of tools, the standardization across the brands, and now you're looking to really, like, build AI systems that enable you guys and testing tools that enable you guys to really have a bunch of operational leverage. That's really cool. Out of curiosity, for our customers listening or folks listening, do you guys have individual, like, ecommerce managers that manage each brand, or is really everything managed at a global level?

[00:08:34] Debra Sabourin: No. We do have site teams for each brand that roll into me. So, each brand has a site manager and an assistant or an associate manager. So, we're handling lots of different data points for lots of different demos, you know, sort of with the one goal. I mean, the only thing we all have in common is we're selling apparel. But, you know, other than that, there's a lot of different things you have to think about for each business specifically. But keeping everything as far as tech goes kind of streamlined helps that, where we're not having to dig into a totally separate problem on one site that couldn't be tackled on another site.

[00:09:10] Midroll: If you're listening to The Ecommerce Toolbox, you're entitled to a podcast exclusive website audit. Go to noibu.com/podcast-audit for a free scan that uncovers the hidden friction blocking your conversions and shows you where you're leaking revenue.

[00:09:24] Kailin Noivo: Yeah. That makes sense. I'm curious. Internally, when we pushed AI heavily on the team, we found that usage was really high, but we found that sometimes things that should be getting built by a central team was starting to be built regionally, and then, like, multiple people would build the same thing, and then, like, the data says different things. Like, have you run into any of those challenges? And, like, how'd you guys, like, because us compared to you, we're like a software business. So, we do transactions, but maybe one or two, a couple of new customers every few hours. You guys across all your properties are probably doing tens of thousands of transactions every couple of days. So, it's a lot more data. So, it's easier to, like, hallucinate. Have you guys run into that issue? And, like, how have you guys actually solved that?

[00:10:09] Debra Sabourin: So, when it comes to centralized sort of ideations, that's one of the things that the AI champion system did is because you would get into a room with all of these people, you know, once a month and they would be telling you what they were working on, and then you're like, oh, I could totally use that. And so, you would put it in, you know, we get access to different agents, and so you're like, oh, so and so built this already. There's no reason to do that already. We just need to add your dataset into whatever it's pulling. So, it was amazing to be able to do that and then to to now continue that as part of just the basic training. So, like new hires now will be onboarded with all of those tools we've already created, which is great. As far as hallucinating, when it comes to data, I mean, there are all these things that you can do to add into your prompts, but honestly, and I think that's the biggest thing, we get a lot of usage, a lot of adoption, but I think you'll find, and I'm sure this is true across all of the different areas where they're using AI, sometimes it's hard to trust what you're seeing because it happens so fast or it's creating these actions at the end and the trust isn't a 100% there. I either find that the trust isn't a 100% there or people are just trusting blindly. So, I think it's really important to verify whatever you're getting out of it. And that's really just what we try to do. Like, it's telling me this, does it make sense?

[00:11:31] Kailin Noivo: Yeah. And I also had a mentor tell me this, and it's like, if you get an AI report and it's, like, 47 pages, like, and you're sending that to people, like, that is not like, that's just lazy, right? Like, that's not, like, how to use AI. It's just creating work for people. It's like, here, I pulled this, like, report, made it missing context, might send you down a rabbit hole, have at her. No. That's really cool. And, like, especially, we're seeing a lot of lean teams, especially if they're on platforms like Shopify and they wanna standardize. We work with folks in this space. We're seeing a lot of AI adoption, especially through systems like Claude, where they're setting up these kinds of guardrails in the business, they're building these custom projects, and then they're running a lot of recurring tasks. Is that something like recurring tasks and things like that? Is that something that you guys have gotten into so that you can help pick up some challenges and help with, like, forecasting? Like, I'm curious, has that been an area that you guys have seen value?

[00:12:28] Debra Sabourin: Yeah. I mean, throughout our organization, there are, you know, from buying, planning, there are tasks that are being automated so that you're not, again, doing these things for hours. If this is something that you do constantly, you can create an automation that will, you know, check your email, and you know, if you get a certain email with a certain subject, create a document, and when you get in that document, this is how you're gonna format it and you know, these are the notes that you're gonna put in based on the email. So, you're not spending all that time doing the copy-paste stuff, which I mean, but it's important because I mean, I know our store team is doing that, you know, for store visits, because all that stuff is really important when it comes to historical accuracy and even tasks. So, you know, send me a reminder or send an email to so and so. These are things that become sort of like your admin to kind of help you so that you can actually focus on the things that are going to move the needle or, you know, affect the bottom line or your top line.

[00:13:30] Kailin Noivo: Very cool. Has there been something that you or your team has tried to automate, but hasn't been easy?

[00:13:35] Debra Sabourin: Yeah. I will say that there are times where you spend so much time prompting or setting it up or doing all of this, and that you're spending days, and honestly, the project probably should have just been built by a human. As much as we wanna AI everything, it's not really built for that. A lot of the times too, if you're building something with AI, one of the things that we always kinda have to keep in mind is, you know, the why. Why are we building this? Like, if it's just to say that we use AI to do it, it's probably not the best use of AI. Some tasks, they're just not built for it. We try to implement AI-driven automation for tasks where there is just a bit of nuance to it, and you require human touch, or the setup was such that it just outweighed the benefit, so we always have to make sure that we're thinking about the why. There are so many other good uses for AI, but we can't AI-ify everything.

[00:14:29] Kailin Noivo: No. That makes a lot of sense. Cool. What are some of the categories that you think AI is helping the most with, like CRO, SEO, like some of those categories? Where are you guys seeing the most adoption?

[00:14:40] Debra Sabourin: Definitely SEO. There's quite a bit of one of the things I will say is that as SEO becomes more important, it's always been important, but now with LLMs and how these tools are crawling your site, having your SEO in a good spot is even more so important. It's sort of a contradictory thing when we get like Meta now is, you know, not really paying attention to the words or the content in your captions. It's looking at the visuals, and you have to have visual diversity and everything. And then the complete opposite is now all these LLMs are looking for the content, they're looking for the text, they're looking for the questions, they're looking for that story, you get so bogged down in all of the visuals that sometimes the SEO takes a back seat. So, now, it's become absolutely clear that SEO is, again, super, super important. And as long as you have that structure, you're doing well. And just remembering that every time you're building something new on your website, like a page for products for us, making sure that your SEO is, you know, top-notch. We're not there fully. There are some things that we're working on, but AI is helping with that. I mean, there's a lot of content that needs to be generated, and some of these things can be automated, which is amazing because you can dig into all kinds of sources where, you know, keywords are concerned, and you know, what questions are being asked by your consumers and how you can tailor that, and there's a lot we can do. So, we're looking at a couple of vendors, and we've worked out some things. Because, again, we're a lean team, so we can't be everything to everyone. And you work with consultants, you can do as much as you can, but sometimes it comes down to, yeah, we can take all of your concerns or all of your suggestions. And at the end of the day, we're just one team, so how can we automate some of this to meet all of those requirements but not kill us in the process?

[00:16:35] Kailin Noivo: I love that. That makes a lot of sense. What are you guys seeing on the GEO front?

[00:16:40] Debra Sabourin: It's still hard. It's hard to tell. I will say that there is a lot of confusion regarding GEO, AEO and what exactly is required. You have a lot of schools of thought where you have to completely make a whole different, you know, catalog where your product data is completely structured in a different way so that all of these tools can be reading them, and then, there's other schools of thought where if you just have really robust SEO that the LLMs are smart enough to go parse through that data and actually, you know, pull out the intent and it doesn't need to be this whole big thing. I think the truth is somewhere in the middle, where it's great to have the parsed down versions, but you really should be working it out on what the consumer sees as well.

[00:17:22] Kailin Noivo: Yeah. It's interesting. It's a tricky one. Like no one knows. No one knows.

[00:17:26] Debra Sabourin: It's still, and you get everybody that tells you 100% this is gonna help, but there's no way. It hasn't been time-tested enough to even know that, because I'll tell you, this has taken off a lot faster than people thought it would. Even so, I feel like every new technological adoption happens at a rate that's so much faster than the last one, because we were talking about like, you know, mobile commerce is gonna be, you know, 50%, but it's gonna take 5 years, and that didn't happen. It took a lot less than that. And now they're talking about the AI adoption is gonna take, you know, 2 years. I think it's gonna be less than that because I, as a person who tends to be a skeptic on all of these things, am using these tools faster than I thought I would. And the way I search now has changed. I'm not looking for keywords. Every search bar I see, I ask it a very long-tail question. So, I think that the adoption of these things just since April, even since April of this year, has, like, skyrocketed, and you know, the algorithms are always changing. So, it's very interesting and exciting what's to come, I think.

[00:18:25] Kailin Noivo: No. I agree. It's really, really exciting, and what we're seeing on our side is that the traffic is small but growing, and conversion is high.

[00:18:34] Debra Sabourin: That's the thing. You may not be talking to a lot of people. There aren't a lot of people in the room, but every one of those people wants to buy something from you. So, the conversion rate is amazing. So, even if you can get just a few more of those people, you can really, like, scale that up.

[00:18:48] Kailin Noivo: No. I'm fully aligned. It's very interesting. As we look to wrap up, what do you think about, like, storefront traffic? Does AEO or, like, does, like, Discovery kind of get sent to the search engines eventually? How is that in fashion? Like, I see somewhat of an argument for light bulbs, but I don't know if I do for clothing. Like, yeah, like, how are you guys thinking about that?

[00:19:14] Debra Sabourin: It's funny that you ask that. I mean, we're still struggling with this because one thing that and it's not, you know, just an us problem because, you know, every vendor we have, we kind of ask, "What's the buzz with your other consumers? Are they seeing any of these things?" And we are seeing an effect on direct traffic, people who are coming directly to us. And there's an argument to be made that, like these AI summaries and these AI overviews, are kind of interrupting that journey. So, now when someone would find you in search, they're not even getting to the search results anymore because these AI summaries have kind of like taken over and they're kind of getting the information that they want from that or they're moving, they're suggesting another product that's not you and it's moving on, and you're not getting that traffic, but it was an interesting drop. But how do you overcome that? That's the kind of where we're at. We are seeing that there is a change in the pattern of how people are getting to our site. And I think the problem is, is it a true traffic problem? Are we losing it, or are they just finding us through other means? It's because, then you have the bots, and you've got all that. So, traffic has become a complete mess at this point. But, yeah, that's kind of where we're at.

[00:20:35] Kailin Noivo: No. It makes sense. I mean, it's pretty in line with what I'm hearing within our kind of customer base. It is a bit of we'll call it a question mark, if we're all question marks.

[00:20:42] Debra Sabourin: Yeah. And I think that's the most frustrating thing because you're at the mercy of these algorithms and these tools and these features. You can spend all of this money on gorgeous marketing and all of these customer journeys, and in the end, you get blocked by an AI summary and none of that works. So, now what do you do? Like, how are you recouping that? And do you just trust that they're finding you through other means? Or is there something else you need to do, but nobody has that, like, playbook yet? And this is gonna be a very interesting holiday season for sure.

[00:21:15] Kailin Noivo: That's for sure. As we're looking to wrap, Debra, thanks so much for the great chat. Really appreciate your time.

[00:21:20] Debra Sabourin: No. Thank you so much for having me. This was fun.

[00:21:23] Outro: The Ecommerce Toolbox: AI and Retail is brought to you by Noibu. To find out more about Noibu and how we unify error monitoring, site performance, and experience analytics to uncover growth opportunities and skyrocket your revenue, visit www.noibu.com. That's n-o-i-b-u.com. And then make sure to search for The Ecommerce Toolbox: AI and Retail on Apple Podcasts, Spotify, or anywhere else podcasts are found, and click subscribe so you don't miss out on any future episodes. On behalf of the team here at Noibu, thanks for listening.