Preparing to Win at AI-Powered Product Discovery, with Matt Ezyk, Sr. Dir. of eCommerce Engineering at Hanna Andersson

Transcript

Our transcripts are generated by AI. Please excuse any typos and if you have any specific questions please email info@digitalshelfinstitute.org.

Peter Crosby (00:00):

Welcome to unpacking the Digital Shelf where we explore brand manufacturing in the digital age. Hey everyone. Peter Crosby here from the Digital Shelf Institute. I found our conversation today about the AI transformed future of commerce with Matt Ezyk, senior director of e-Commerce Engineering at Hannah Anderson, both urgent and comforting. Urgent in that he strongly believes that consumer journeys and product discovery will be upended by conversational commerce. And brands need to prepare and comforting in those e-commerce and digital have spent their careers being upended and have built the mindset and skills to manage through it. Matt walked us through his approach to riding this transformation wave, armed with curiosity, discipline, and a test and learn mindset. Matt, welcome to the podcast. We are so excited to have you on.

Matt Ezyk (01:02):

Hello. Thanks. Happy to be here.

Peter Crosby (01:04):

We are going to dive into all things product discovery. I mean, really the way consumers discover product is being fundamentally disrupted by AI and is going to continue to change, but the big change is what does that really mean for brands and what should they do about it? So Matt, first of all, tell us what is changing and what you are watching as a brand at Hannah Anderson.

Matt Ezyk (01:26):

Yeah, I can't really talk specifics of what I'm doing, but just in general for sure. I mean, to me, the biggest shift is that discovery is no longer starting with our site or even in search. It's happening in feed. It's happening in chat. It's increasingly through AI driven assistance. So customers aren't typing, in our case, just kids pajamas into Google anymore. They're asking ai, what are some organic, durable pajamas for toddlers that ship fast, for example? And I'm sure there's many, many other, many other variants of that. So as a brand and any brand, really, you really need to watch how AI is interpreting product data and language because that's the new front door to your brand. So what matters now is not just what you sell, but how well AI understands it. And that changes how we structure our data, our content in even our product naming. It really changes everything.

Lauren Livak Gilbert (02:29):

I think it's really exciting too, right? Because it's naturally how we think about things, right? For the holiday, we had a barbecue, and I didn't go like, okay, I need barbecue sauce. I need this. I was like, okay, we're having a barbecue. What do I need to think about? What do I need to purchase? I need paper plates and paper napkins and forks, and it's moving more towards a natural, conversational way of shopping. And I think that's really exciting. And it must be for a brand too, right? Because that also connects into how you're telling your story, I would assume.

Matt Ezyk (03:03):

Yeah, absolutely. This also shifts the journey from a search and click to an ask and decide to your point. So instead of browsing 10 different websites, a shopper might ask a question, find me, in our case anyway, find me the best gender neutral sleepwear for a 6-year-old, under $50, for example. And then the AI is going to decide what gets shown. So we're seeing this already with Google and Perplexity and Chat GPT, our traditional SEO tactics, just don't cut it there. You have to do things differently. So one example, I guess you could optimize PDPs and feed data, not just for humans, but how AI summarizes attributes. For us, it's terms organic and durable and made to play. That will fundamentally change how often you show up in AI generated product lists. It's not about ranking number one in Google as much as now it's being shifted into being recommended by the assistant that a customer really trusts.

Peter Crosby (04:08):

I mean, it really does change the KPIs for success for so much of it is what are my clicks, what's my traffic? And a lot of that is going to be sort of intermediated.

Matt Ezyk (04:22):

And

Peter Crosby (04:23):

So kind of figuring out how we understand success, I think is going to be an interesting challenge as well in this period.

Matt Ezyk (04:33):

Yeah, absolutely. That's from measuring the KPIs like you mentioned, to going down and looking like how do you plan from this financially, logistically, it's just a major shift that I think a lot of brands may not be really actively looking at right now, and they should.

Lauren Livak Gilbert (04:53):

So Matt, when we think about agentic ai, so consumers are interacting with an agent who you're asking those questions about buying a gift for a child at age six and looking for a specific requirement. How are you seeing it kind of change in the customer journey? Are they going to the agent first? Are they searching and then talking to the agent? Do you have any examples of maybe how you've seen behavior for your consumer shift because of it?

Matt Ezyk (05:21):

Yeah, you can see customers landing on our site through an AI like chat, GPT, and that's why we're, or really any branch should look into optimizing their PDPs and their feed data. You can even optimize your site to show up better in something like chat, GPT, similar to you would do with SEO, like a robots txt file. So those are something that we focus on. Every branch should focus on that now. And again, like I was mentioning before, it's very quickly moving away from being ranked on Google to being ranked on an assistant that the customer really trusts

Lauren Livak Gilbert (06:10):

And their assistant is recommending much many fewer products. You're going from having hundreds of thousands online on A PDP or on Amazon to maybe three or four. So as a brand, does that scare you? Does that excite you? How are you thinking about targeting your consumer when they might only see one option?

Matt Ezyk (06:32):

Yeah, it's a reality that's coming, I think sooner than a lot of people think. I think I read a stat that by 2026, Gartner was saying that organic Google search is going to decline 25%, and that's just going to continue to happen. It's just going to fundamentally change the way customers discover products. So you're right, there could be a situation where some of our key products or hero products are not discoverable by customers for whatever reason, and there's all sorts of data points that an LLM like chat GPT uses, whether that's search or even now they're pulling data from Reddit. I mean, it's evolving pretty quickly. So yeah, mean, there's definitely a lot of situations where we could see skews all of a sudden just drop off in sales and not know why, and that could be a key indicator that something's different there. So it's definitely changing the dynamic of product discovery for sure.

Peter Crosby (07:40):

And when you think about that, Matt, there is that mix you're talking about, we're not paying enough attention, and it's hard to pay enough attention when still how this all is going to work is unclear. It's not like, oh, here's the roadmap to do this. I just haven't turned to it yet. There really isn't really a clear roadmap. And so I'm imagining that there's some kind of deeper collaboration that needs to be done between the business side, IT slash data slash vendors, I would imagine. Who are your key vendors that you rely on for this kind of stuff? And I'm wondering how you think about that. How do we all stay on top of it together in a way that is actionable? Do you have thoughts around that?

Matt Ezyk (08:35):

Yeah, it's looking at it from a lot of different angles, but you mentioned finance, right? That's

(08:44):

A big part of it too. If you're going to do some experimenting, you have to have KPIs that line up with it, just like any kind of tool or investment that you're going to make. So I like to call it AI readiness, like the not flashy experiments, but really building core infrastructure. And that's down to budget, which I think we'll get into too and how we plan for that. But that includes enriching product data, involving or improving how you structure descriptions, making sure content is semantically rich and machine readable. It also includes creative too, like creative. You can invest in content that answers why a product is a fit, not just what it is, but that's what really the generative AI latches onto. So content, it's interesting, it's like a two-way conversation. What can be returned to a user in a chat response is really the key, not just what looks good on A PDP

Lauren Livak Gilbert (09:48):

And Matt, when you talk about all of these different pieces coming together and how much it's going to change, there's so many different people involved in this internally from an organization standpoint, and I feel like sometimes brands haven't really even gotten SEO, right? And now we have to worry about GEO generative engine optimization. So how are you thinking about the people part of this, and who are you working with? Are you working with it? Are you working with marketing? Are you working with sales? How are you getting people to listen inside your org to be like, Hey, we need to change

Matt Ezyk (10:25):

Being an evangelist for ai, letting everyone know that we're not scared of it, we're embracing it, but doing so tactfully and being calculated with it is really important. There's a lot of noise around all the different things that AI can do, good or bad. So really just tying it into actionable strategies and KPIs that matter to the business is so important. And then just continuously having a test and learn mindset. And we've been doing that in e-comm forever, even before generative ai. So we did that with 10 or 12 years ago with predictive ai. It's been around a long time, and the same type of mindset. We're going to take something and we're going to AB test it, and if it works, great, if not, we go back to the drawing board. So there's a lot that you can do, you can test, but that's the important thing is to really make sure everyone has that experimental, adventurous mindset where we want to push the envelope as much as we can.

Peter Crosby (11:38):

And that's why I get so excited for the listeners of this podcast, the folks that we work with and the community that the Digital Shelf Institute is because you folks have these skills, because your whole in digital commerce has been, oh, this is how this is going to work now. Or it's a constant both sprint and marathon mindset, which is how do I do the best with what I know and can do right now, get that machine going, which will then change tomorrow. But in the meantime, I also need to be figuring out what this looks like on the longer arc. And I'm just wondering if that's how you feel. It sounds like in some ways, and I'm sure you have partners at your organization, but you've taken on that role of AI comm in chief or something. I don't know how you would describe

Matt Ezyk (12:30):

It,

Peter Crosby (12:31):

And I'm wondering if you feel that way about your career. Does this feel like I was sort of born for this to take this on?

Matt Ezyk (12:43):

Yeah, that's an interesting point. I've always typically embraced new technology and just rolled with it. And I was doing that even when I started working in e-commerce many, many years ago. And not everyone does that. I've definitely done that. I'm sure there's going to be something for the rest of my career that will come out in revolutionary revolutionize our industry again. And I've always had the mindset that I'm open to everything and embrace it, but also being careful and calculated and making sure that what we implement makes sense for the business. So it, it's a catch 22 for sure, but I think everyone should never stop learning and embracing what's out there. And especially with generative ai, it's here, it's growing, it's not going anywhere. There's plenty of tools out there for free that you can test and learn.

(13:43):

I'll give you an example. When I look at product discovery or shopping especially, we're in this new era of generative AI with ag agentic ai, and there's certain situations where people are going to shop no matter what, because they love to shop. It's like an endorphin hit, whether it's buying clothes or shoes. But for me personally, I do not like grocery shopping. So I thought, okay, I have all these tools available to me. How can I automate this and build an agent to do this? Right? And I'm not there yet, but essentially the idea is that I'd go in and I'm using chat, GPT and Instacart for this. So I go in and I say, Hey, I want to build healthy meals Monday through Friday. I want high protein, low carb, moderate fat. I choose my proteins, it builds a meal plan for me. And then I connected it through an API to Instacart, and it sent an ingredient list out of all those meals.

(14:44):

And then I can just pick the store and then it will order it all for me. Now, the next evolution of that is that it'll price shop it and then place all the orders for me that way. So that's the new evolution of it, and that's just one experiment that I do personally in my own time. And I'm constantly experimenting just to learn what it can do, because when the time is right for me to implement that into my business, whether it's here or wherever in the future, I'll have plenty of years and hours of experimenting and learning. I never want to stop learning new tech. So it's exciting for me.

Peter Crosby (15:25):

I'm looking forward to the agent that you write. That is, okay, now it's the weekend. Let's go crazy and see what that shopping list is. Yeah. Oh, the carbs. Exactly, exactly. I respect your carb

Matt Ezyk (15:38):

Load.

Peter Crosby (15:39):

Yes, exactly. I respect your healthy agent, but come on, bring it on. We were talking earlier about earlier about sort of financing this shift and your friends in finance and figuring out how do you budget for these tests and learns for the changes that are coming. How do you think about, because that requires both a, I would imagine some flexibility and agility, which is not necessarily finance's comfort zone, but also I would imagine, again, you were talking about earlier tied to KPIs. How do you think about being able to power that shift with the resources that are necessary to do the learning that you need to do?

Matt Ezyk (16:31):

Yeah, not all brands are in the same situation, but if you're able to, I think it's really important, not just with ai, but just to have an innovation budget where you can experiment. I've really tried to do that throughout my career, wherever I was with whatever brand I was working with, whether it was an advisory capacity or I was on the brand side. And in this instance I was mentioning before, I think you could do a lot with carving out a specific innovation budget for AI readiness because it's rapidly changing. A lot of brands aren't ready for it yet. It doesn't mean you shouldn't do anything with it, so you should experiment. You should build out the core infrastructure. I mean, really what makes AI tick, it's no secret to anyone listening here. It's data. And most companies that I've seen, they have a lot to be desired to clean up their data, get more data, figure out how to activate their data better. So it's really building that AI readiness infrastructure. And to your point with finance, it's building out an experimental or an innovation budget for that too. That's kind of how I approach it.

Lauren Livak Gilbert (17:49):

And do you have an education program internally, right? Because I mean, similar to E-com, not everybody grew up in digital and is comfortable with digital, and now AI comes through and it's an even double click or expansion of that digital, and so people need to become comfortable with it and start using it in their, have you built out an education program? Is that what your organization is thinking about so you can adopt it?

Matt Ezyk (18:15):

No, I haven't. I think it's a good idea. However, it's interesting. I met someone last week who is a professor who teaches courses on ai, and he said to me that no one is an expert in AI because it's changing so much, which I thought was pretty profound coming from someone who is a professor. The other thing too, it's like things could be totally different. I'm preparing for a keynote session at eTail East in August, and I've had to scrap my session a couple times because everything changes. So I think that would be very difficult to do, but basic concepts for sure. And you're seeing companies now start to build out guidelines into use and policy around ai. I think every company is doing that or down the road doing that. So because it's here, it's not going away. So you have to embrace it and adopt it.

Peter Crosby (19:20):

Yeah, I mean, I'll speak for myself. The process of just going in and using it yourself is really the only way to start to understand what the potential power is and also what sort of the risks are. I mean, I think any of us that have used it have seen the hallucination, have seen the desire of AI to want to please you want to give you what you ask for, even if it's stupid, and they'll come up with something, they're never going to say, I'm not going to answer that question. No, they're going to answer it. And then so much of it falls on humans to be able to understand context to review. And so I think the only way I understood that was just by going in and doing it myself. And so I think where we work has done a lot of investment in making sure that everyone is in it and using it, and it's become almost part of the requirements, one of the requirements of our jobs to be in there doing this. And that's something I wonder what that sort of adoption curve is from the early adopters to now kind of the midline. And companies, I believe really need to, I think the days of being able to close your eyes and say, la, la, la, no, it's too risky. We're not going to do it. Is not going to work. No. Yeah. And I'm wondering, I think we probably agree on that, Matt, I would imagine.

Matt Ezyk (21:05):

Yeah, absolutely. It's here. It's not going anywhere. It's being infused in every piece of technology in the e-commerce world. There is, you can't go anywhere or log on LinkedIn or anywhere without hearing something about ai. And which I think is a detriment too, because it's so much that you're like, okay, what does your product really do? Right? You mentioned ai, great, but how does that help me? I think that's the key thing too, is that if you're in a product, whether it's for e-comm or something else, just saying that you use AI is just not going to do anything for you. It really isn't. But tying it to KPIs that the business cares about, that's a whole different game. So if it's an business, if it's increasing conversion, lowering abandonment, increasing a OV, if you're implementing something that can do that, then great. You can go in and do an AB test and just again, just experiment, test, and learn. These are things that brands have been doing since forever. This is just the latest iteration of technology where you can do that.

Lauren Livak Gilbert (22:27):

And Matt, I'd love your thoughts on this. I was thinking about it the other day. I feel like a lot of people on the brand side who you talk about ai, they're like, oh, it's not going to happen for a while, or It'ss just not affecting us right now. And my dispute to that is it's going to start happening more quickly. And I think one of the biggest things that just came out was Google's announcement because

(22:51):

People were still using Google and it wasn't really fully integrated into more of an agent kind of experience. But now that it's fully built into Google, which is the number one search engine that most people are using, if they're not in tech, if they're not in the industry we work in, they're just normal people out there in the world. And so now it's fully integrated, so they might not even know that they're using it. And I feel like that will be an even bigger catalyst because it's in the traditional way of working that most normal humans on the planet are encountering. What are your thoughts on that? That's how I kind of dispute that concern from the brand side.

Matt Ezyk (23:29):

That was what I was alluding to with having scrapped my plan for my keynote go updates. Yeah,

Lauren Livak Gilbert (23:37):

Sorry.

Matt Ezyk (23:38):

It definitely did. So yeah, you're absolutely right. There's a lot of brands out there that they have way bigger fish to fry than AI right now, but still doesn't mean that you shouldn't think about it because like I said, ad nauseum, it is coming. And what I would not be surprised to see is that you would see a brand that is ignoring it and working on other things, and then all of a sudden one day Google Ag agentic version of Google and Chat, GPT and Perplexity and all those are predominantly recommending their competitor, and all of a sudden there's sales drop and they don't know why. I guarantee that's going to happen to a lot of brands and it's going to take them a while to figure it out. So I think as a brand, there's probably not a lot of brands out there that can really go all in with AI experimentation. A brand that comes to mind that I really admire is Elf Beauty.

(24:44):

They push the envelope so hard and test and learn and implement and try everything that's out there. I really admire that. So I usually watch what they do. There's not a lot of brands out there like that, though. Most brands, I still trying to figure out their data and how to do more with it. But I think you can focus on three things when you're preparing your brand, whether it's now or something in the future. I think the biggest thing is doing what you can to make sure that you show up in AI generated search recommendations. Again, that's such a big shift that's happening. So if you're not doing anything with that by 2026, you're just going to be losing to any kind of competition out there that is really actioning on that. Secondly, cleaning and tagging product data to make it AI friendly. There's so much that you can do to change that.

(25:44):

And then to your point earlier, training teams to think about AI in the business, whether it's data or it's UX flows, or just having everyone think about using AI into the entire business, that's just going to be more and more commonplace. So I'm looking how conversational commerce will evolve personally, whether that's through voice chat, embedded AI tools and brands want to be part of the conversation. It's not just an endpoint. So you really have to build cohesion between marketing, product engineering to make all this work. It's not just a tech problem, it's a total mindset shift

Peter Crosby (26:40):

Business. Yeah.

Matt Ezyk (26:41):

That you really need to make. And I think that ultimately the brands that win in this next phase are the ones that don't react to AI like we were mentioning, but design with it in mind. So it's not an overnight, just do it. Right? It's a slow mindset shift that everything that you do, every new project, you need to think about it and eventually you'll get into an AI first place with everything in your business. It's just not going to happen overnight. But you have to start thinking about it now.

Peter Crosby (27:14):

And that's where leadership comes in. I mean, when you talk about Elf Beauty, I'm sure if you trace that mindset and those behaviors, they start at the top.

(27:26):

And people are not only given permission to experiment, they're required to experiment. That's right. And people go to work there because they want to work in that environment. So I think the leadership that you clearly provide out of your personal interest, but also because of the role you sit in is incredibly important. But it has to inspire slash be matched, I would imagine, by the people who run the company or you're really fighting a losing battle. We've said this before and it's easy to say and really hard to do, but these are the moments where you have to look around at where you are and say, is this an organization that's going to embrace this future? And if not, what does that mean for me? I dunno. I wonder if, does that sound like bullshit or does that resonate with you?

Matt Ezyk (28:32):

It definitely does. I think that's probably happened throughout time in any technical career as things have evolved. And I can guarantee you there was times where people were learning about e-commerce in the nineties or the early two thousands, and the company they were at where they were just like, oh, this is not going to take off. We're just going to stick with our stores. And people have been faced with that decision throughout time. So this is just another iteration of that. And I'm sure there's something in the future that has not been invented yet that will come along and we'll be in that change cycle again. So yeah, absolutely. You should at the top have leadership that embraces it. And then as a leader, letting your team know that we have to be calculated and careful and tie business goals into technology. We're not just going to implement technology for the sake of technology, but it has to tie to a business goal. And if that does include something that is AI driven, then great, and then we can test it. And if it doesn't work, we scrap it. Right? I could be excited about technology, but if it's a dud, then so be it. Right?

(29:51):

You go, you test, you take the data, you learn from it, and then you go from there.

Peter Crosby (29:56):

Yeah. I was so struck by your list of three things that you need to think about. That's always great to have three and to underst, and then the part of me as I was listening sort of on behalf of our listeners, I was thinking as you went through them, my first question would be, oh, how do you do that? Each one of the things, it was like, okay, each one of those is something to unpack. And I was wondering if you have resources or one or two that you go to that you trust that help you build out these strategies? Are there substack that you subscribe to or anything like that that you feel comfortable recommending for folks that are listening to the podcast?

Matt Ezyk (30:43):

There's so much out there.

Peter Crosby (30:45):

Yeah, I know that's a problem.

Matt Ezyk (30:48):

You really can find a lot of people on LinkedIn that are talking about things that they're doing. And like I said, it changes so fast that the hype cycle is just really, really tough to follow. So I like to follow along with certain people on there. I can remember, I can't actually, I can't remember the name of the LLM, but there was a Chinese LLM that came out and it was deep seat, deep sea,

(31:24):

And I woke up one day and was, I opened LinkedIn and it was everywhere. I'm like, okay, I got to research this and see what it is. And that's typically how I learned it is if it's someone's talking about it, then I'm like, okay, it seems relevant. I'm going to see if this is BS or if it's something that I need to really look at. Because you never know. That could be the next thing where your customers are asking it for advice on product discovery. You probably can pick up that. That's a big topic for me because I think that's going to catch people by surprise more than anything. But yeah, that's probably to me is the easiest way because there's so many people writing about AI up there. It's hard to really pick one or two

Peter Crosby (32:13):

Anyone.

Matt Ezyk (32:13):

Yeah. But there's definitely some content creators out there on LinkedIn.

Peter Crosby (32:17):

Yeah, I mean, we have a Slack channel that is the AI Slack channel, and that's the richest source for me. It's a similar thing. It's a little bit more filtered. We already have AI experts and then they're coming in with the stuff that really struck them, and so I get to kind of surf off of that. But Lauren, do you have any that you

Lauren Livak Gilbert (32:37):

Yeah, I do the same thing. I go to LinkedIn. I mean, I make it a habit every morning of going on LinkedIn reading at least for 10, 15 minutes, kind of seeing what people are talking about and also using the LinkedIn news on the right hand side. So when you log into LinkedIn, it says LinkedIn News, and it's like the hottest topics, and that's what most people are talking about. So there is one person on LinkedIn who I really like to follow. Her name is Lauren Morganstein Shivan, and Lauren, if I mispronounce her name, I apologize, and please tell me how to pronounce it. But I love her newsletter where she talks about practical applications of AI. So that's one person I definitely follow. And then to Matt's point, just checking on LinkedIn, seeing who's talking about what articles are coming out. That's really my go-to source.

Matt Ezyk (33:22):

Yeah. The other places I look is I'm part of a couple retail AI councils, so I have peers in that group that are talking about different things, and I can pick up on trends and then research from there to see what other people like me are thinking about looking at. That's a really valuable resource for me as well.

Peter Crosby (33:47):

Well, we all need to tell each other what's grabbing our attention and how we're learning because it's happening at such a fast pace. Matt, I just want to thank you so much for coming on and sharing your viewpoint on this and how you are trudging through this crazy world right now. It really helps when folks like you come on and share with our community what's working for them, what best practices are, and you've been awesome at doing that today. Thank you so much.

Matt Ezyk (34:16):

Thank you.

Peter Crosby (34:17):

Thanks again to Matt for sharing his AI strategies. More strategic thinking is always on tap when you become a member at digitalshelfinstitute.org. Thanks for being part of our community.