Latest / Retail Media Breakfast Club / AI Shopping Agents vs. Retailers: Who Owns Discovery Now?
Transcript
- Have Retailers Lost Discovery to AI for Good? === [00:00:00] Kiri: Welcome back to another episode of The Retail Media Breakfast Club. I'm your host, Kerry Masters, and today I have another snips episode for you where I'm sharing some snippets from one of my favorite podcasts, along with some thoughts that I had while I was listening [00:00:15] to it. Today I'm gonna share some snips of the Unpacking the Digital Shelf podcast with you with one of my e-commerce besties, Lauren Levi Gilbert and Peter Crosby. And they interviewed Natalia Krammer Rich, who is the senior director of [00:00:30] product marketing at Kibo Commerce and natalia came on the podcast to talk about agentic AI and AI shopping agents, and and she laid out something that's been bothering me a little bit over the past year [00:00:45] instead of just adding AI features to existing commerce experiences. [00:00:50] We're actually watching the entire interface layer get replaced, and I think many people are starting to realize that we need to go beyond, [00:01:00] should we build a chat bot and think about what happens when consumers stop interacting with your website, with your app, with your carefully crafted product pages, and instead funnel everything through conversational [00:01:15] LLMs. [00:01:15] This is the pane of glass that retailers have spent decades optimizing, and it might get a new owner. I don't think it's all doom and gloom, and Natalia is able to chime in with some more technical explanations [00:01:30] of what's going on and and what retailers and brands need to consider. [00:01:34] Let's listen. [00:01:35] Natalija: But consumers are moving beyond that. They're moving to commerce anywhere. Meaning like, I wanna be able to speak to my phone and have the Tide pod, you know, pods show up on my doorstep when I need them. [00:01:44] I want [00:01:45] Amazon to tell me when I'm running out of, you know, lotion or whatever it is that I need. So, um, commerce Anywhere is. Pushing the boundary of complexity in terms of how vendors and technologists deliver, uh, commerce, um, use cases. And [00:02:00] so, you know, we do a lot of composable, modular headless because of that. [00:02:04] That's really where the frontier is. And we are leaning into agen commerce again, because of that, it's one way that consumers can simplify how they're [00:02:15] purchasing. Because even though they're ordering from anywhere, or they want it to be delivered to anywhere, from anywhere, they don't really care. It's sort of like the delivery becomes commoditized. [00:02:24] They want one central place to funnel all their questions. And so that's becoming, [00:02:30] um, the pane of glass, uh, agentic is be, is going to become the pane of glass, and that presents a threat to personalization engines. Because if you think about it, personalization engines today typically own the pane of glass. [00:02:42] They own the website, they own the ad, they [00:02:45] own the display layer. Where you get to use predictive to showcase the next best action. With generative, you can't use predictive, you're using generative, right? And so you're using, you can't plug into those personalization engines. So personalization [00:03:00] providers have to rethink how they are providing those services and how they're gonna do that in a sort of distributed technology way where they're connecting to many, many things. [00:03:08] [00:03:09] Kiri: I really think about this topic a lot and I'm glad to hear Natalia express it in this way [00:03:15] because few retailers seem to grasp. What they are about to lose. For the past decade, retailers have optimized every pixel of their digital properties. They've invested millions in personalization [00:03:30] engines, AB testing frameworks, and predictive algorithms, all designed to control what shoppers see at every touch point. [00:03:39] But that control is evaporating because even without fully [00:03:45] agentic AI shopping journeys, we're already seeing so many consumers spend time in the research and consideration stages of their customer journeys. [00:03:55] Doing that inside LLMs now rather than on retailer.com. And [00:04:00] so retailers no longer own that layer in the discovery and consideration stage, the AI does. And so all those carefully merchandised category pages, the [00:04:15] merchandising, the personalized homepage, none of that matters if shoppers never see it. [00:04:21] Don't get me wrong, I'm not down on retailer websites and I don't think it's all doom and gloom. It doesn't have to be [00:04:30] a race to the bottom. It does, however, require retailers to fundamentally rethink their value proposition, especially online. [00:04:39] If your retailer.com site, if its primary value, was being a [00:04:45] place where consumers could research and discover and compare products. Well, LLMs are doing a better job of that right now. I'm sorry to say. So what do you uniquely offer? Is it an [00:05:00] exclusive product assortment? Is it convenient store locations? [00:05:04] Is it a unique experience in a category? Is it the ease of doing business with you as a retailer? Retailers who figured this out and double down on that [00:05:15] messaging . Are going to stand out, but it requires honest introspection about what you're actually good at, beyond just having a website. Did you [00:05:30] know that leading retail media networks drive 85% of their ads through mid and long tail advertisers? [00:05:39] Kiri Masters: Miracle Ads provides full funnel ad formats tailored to both one P and three P [00:05:45] advertisers leveraging unique AI capabilities that provide unprecedented levels of relevance and engagement. Retailers who want to capture ad spend from the long tail of three P Marketplace sellers use miracle [00:06:00] ads in their tech stack. [00:06:01] Learn more@miracle.com. That's M-I-R-A-K l.com. [00:06:09] Kiri: Let's listen into the next part of the conversation when Natalia talks about some of the [00:06:15] specific. Technical foundations that retailer dot coms need to have. [00:06:19] Natalija: Yes, it is. And I'll, I'll leave you with this. I think one of the backings, you know, backings of the technology that people that often gets un sort of unspoken is, you know, let's say you have a [00:06:30] shopper agent and you have synonym search on your site, which is the most often kind of type of search that you could have. [00:06:37] Um. You ask it a question, as an lm, they can understand your question very well, but then it has to go into [00:06:45] the commerce engine and search for your product. And if it's backed by synonym and keyword search and the search is not gonna be very good. And so one of the things we've done is we, 'cause there's like a whole search spectrum, so you have. [00:06:56] No search. Some people literally have no search. Maybe they feel like they don't need it. You have [00:07:00] very simple synonym and keyword search. Then you have the middle of the road, which is what's called like AI vector search, and then you have overly personalized search. So like the Algolia, the um, constructors of the world. [00:07:12] They're like, you don't even have to tell it what you're [00:07:15] looking for. You, you say something, it's like, actually, based on your profile, here's what we think you're looking for. And it's probably more accurate than what you said to us. 'cause consumer, you don't actually know what you're looking for. So that's like the high end search. [00:07:26] But that can be, you know, uh, it can cost a lot of money to get [00:07:30] there. And so in the middle of the road, AI vector search is kind of like. Taking all your product attributes and merging them into a vectorized space. So like you're putting them into, um, you're giving, each attribute has a dimension and each [00:07:45] product therefore has a unique vector. [00:07:47] And so when you're searching for your term, le let's say I search dress shirt, um, you know, a synonym or a keyword switch would. Look up a, you know, red dress shirt. It would look up maybe red dresses, as well as red dress shirts. [00:08:00] It would bring up both of those options for you because both of them have those key words. [00:08:04] But in vector space, it would know the difference between red dress and red dress shirt and it would be able to, um. Take out red dresses. Um, and so [00:08:15] basically my, my point is if you're not using AI vector search or higher, you're putting lipstick on a pig if you're launching shopper agents. [00:08:22] [00:08:24] Kiri: And then the host of the podcast asked Natalia for her recommendations on what brands and retailers should do. [00:08:30] She gave some practical tips, but then I love where she landed on this as some overall advice. [00:08:36] Natalija: You know, we're entering an era where com there's also a lot of competitive proliferation in retail. [00:08:43] Like American retailers [00:08:45] are, have a lot of pressures from international players, from marketplaces, uh, from cheap, you know, like the she and the teams of the world. So there's not a lot of. You know, [00:09:00] there's a lot of threats out there and the only way forward is to innovate on how you're delivering goods and services to your customers and, um, how you're communicating your brand, the power of your brand to your customers. [00:09:11] And so there's a lot of different tactics by think leaning [00:09:15] into digital innovation and experimenting using these technologies so you can figure it out earlier, other than later is super paramount I think. [00:09:22] [00:09:23] Kiri: I couldn't have said it better myself. I think Natalia is, is spot on here. The future is [00:09:30] changing right in front of us and um, the way that we have shopped in the past is going to change the retailers and brands who lean in with curiosity and look at where things are shifting to get messy. [00:09:44] Try out some [00:09:45] of these experiences yourself. Think about what is your unique value proposition as a retailer when the discovery layer changes? [00:09:55] Many more thoughts on this to come. I hope you enjoyed this snippet of the [00:10:00] Unpacking the Digital Shelf podcast. Thanks for tuning in and I'll catch you tomorrow. [00:10:04]