S2E13 Sabrina McPherson - Agentic Commerce

[00:00:00] Teaser

[00:00:00] ​

[00:00:01] Casey Golden: Traditional search is dead. Frankly, it just sucks lately. Sabrina McPherson of Publicis Sapient says the new goal is selection, getting the AI to recommend and pick your product

[00:00:14] Ricardo Belmar: And AI shops differently than we do. People buy on emotion. The machine wants clean, consistent data it can trust. Basically the wall of text that the rest of us close the laptop on

[00:00:24] Casey Golden: Oh, I have so much to say. This episode of Retail Transformers recorded live from RetailClub AI Festival, Sabrina explains how to sell to both audiences

[00:00:36] Ricardo Belmar: And why shoppers aren't ready to hand their wallets to an agent just yet

[00:00:41] Show Intro

[00:00:49] Ricardo Belmar: Welcome to Retail Transformers, candid conversations with the visionaries shaping the future of commerce.

[00:00:54] I'm Ricardo Belmar

[00:00:56] Casey Golden: And I'm Casey Golden.

[00:00:57] Welcome back, Retail Razor fans. This is the [00:01:00] podcast where we cut through the clutter to bring you the sharpest insights from the leaders redefining retail, retail careers, and the future of retailing

[00:01:08] Ricardo Belmar: This is season two, episode 13, and it comes to you recorded live and in person from the RetailClub AI Festival in Huntington Beach, where we were joined by Sabrina McPherson, senior vice president and practice lead for consumer products management consulting at Publicis Sapient.

[00:01:24] Casey Golden: Sabrina has spent more than 20 years helping consumer product companies find new digital value pools and build what it takes to capture them. So of course, we talked agentic commerce

[00:01:36] Ricardo Belmar: Of course. Sabrina says the goal has moved from search to selection. To get recommended and picked by AI, brands and retailers have to share data, and brands have to explain what makes their product different in a language a machine can read. Price and attributes alone just won't get you there

[00:01:53] Casey Golden: Then we get into designing commerce for two audiences at the same time. People who buy [00:02:00] on emotion, that's me, and imagery, and AI agents that want consistent data they can trust. Plus, why a wedding dress and a dress for a wedding are two different things that AI just may not get

[00:02:17] Ricardo Belmar: And with Meta's new Muse agent racking up downloads, we asked how far shoppers will really trust a personal agent to buy for them. Sabrina's answer, trust gets earned one step at a time, and nobody wants an agent selling them the day-old eggs.

[00:02:32] But before we dive in, let me tell you about where the Retail Razor will be next, and that's Retail Agent Con. On Tuesday, November 17th from 11:00 AM to 3:00 PM Eastern, join me and more than 700 of your retail peers at Retail Agent Con, a live virtual event from our good friends at IHL Group with ClearSight AI. It's an honest look at what AI agents are really doing today across merchandising, order management, inventory, and store operations, which ones are real, and what [00:03:00] has to change in your org chart to make them stick. You'll hear from retail leaders like Glenn Ellis of Tractor Supply, Scot Seiger of David's Bridal, and Justin Emig of Artisan Lane Furniture, plus Dr. Janet Sherlock on getting agents and people working together, IHL's Greg Buzek and Jerry Sheldon, and yours truly. Registration is free for retailers at retailagentcon.com.

[00:03:22] That's four hours of real-world insight you can use heading into NRF's big show. Register today.

[00:03:28] Thank you to Retail Agent Con for making this episode of Retail Transformers possible.

[00:03:33] Casey Golden: Now let's welcome Sabrina McPherson of Publicis Sapient to Retail Transformers

[00:03:39] ​

[00:03:39] AI Is Choosing Your Products. Are You Ready?

[00:03:45] Ricardo Belmar: And we're back in the Retail Razor Tent, and this time we're joined by Sabrina McPherson from Publicis Sapient. Sabrina, thanks for joining us.

[00:03:52] Sabrina McPherson: Thank you for having me. Yeah,

[00:03:53] Ricardo Belmar: absolutely. So we're gonna talk, of course, because we're at RetailClub AI Festival we're gonna talk about agentic commerce.[00:04:00]

[00:04:00] What else is there to talk about these days? and one of the things that we really wanted to start by asking you is, there's a lot of consideration around how retailers and, and brands have to align their product data, particularly

[00:04:12] around discovery by all the AI systems. And it...

[00:04:15] there's a lot of data involved, updating a lot of product schemas and things. So I, I guess that there's sort of a joint goal of both the retail and the brands want to be surfaced by AI. So I wanted to ask, what do you think needs to happen first for, for retailers and brands to agree on how to do that?

[00:04:30] Sabrina McPherson: Yeah. No, I hear you. It's, it's a common thing that we're actually in the middle of dealing with with most of our clients.

[00:04:35] Search Is Dead. Selection Is the New Game

[00:04:35] Sabrina McPherson: I would say, first off though, the, honestly, the goal has changed, right? So it used to be always about search, right? Right. Right.

[00:04:41] And

[00:04:42] now it's selection. So it's what do you have to do to get your product recommended and selected by the machine, by the AI? And so the whole conversation has shifted, right?

[00:04:53] Originally, it was traditional search. Obviously, we all know that's dead. But it went to AEO to disaster.

[00:04:58] Casey Golden: Have you tried lately?

[00:04:59] Ricardo Belmar: Yeah.

[00:04:59] Sabrina McPherson: [00:05:00] I just, it's a nightmare. And anyone who's-- It was funny 'cause I, I was talking about this a year ago, and people were still in denial that traditional search is dead.

[00:05:06] Yeah. Here we are a year later- and we're.. .and we're not even talking about search. We're talking about selection. Mm-hmm. And so I think what's happening now is you have brands and retailers discussing, "Okay, let's share data. Let's talk about the data that we can build to move us towards selection," because we're both at the same, same agenda.

[00:05:23] Both sides of the equation want a purchase, right? Mm-hmm. to be selecting the right product that allows there not to be a return. And so I think what, what that means is really for brands, the first stage of just having the basics done around the product data layer just becomes more increasingly important, right?

[00:05:41] Right. Right. Not just to get searched or discovered. It's really because the-- you really need the AI to understand the differentiation of your product versus other products. And then you need the data from the retailer to inform what the human is searching for, the needs, what are they really clicking on, all of that jazz, right?

[00:05:59] Yeah. So it's [00:06:00] about the combination of the enough information behind the products from the brand side to inform sort of a differentiation. It's not just price. It's not just attributes anymore, right? It's a lot more that's required- Right ... to move, to be able to get to a selection.

[00:06:15] Ricardo Belmar: And, and there's a sense there's the more detail, the better.

[00:06:17] Sabrina McPherson: Yes. Yes. Absolutely. And it's not just about have the detail. It's also consistent data, that detail that basically the AI can trust, right?

[00:06:26] Casey Golden: And

[00:06:26] Why AI Can't Tell Ralph Lauren Labels Apart

[00:06:30] Casey Golden: context, like so much context around the brand that like, a lot... ChatGPT does not know the difference between Ralph Lauren Purple Label, Blue Label, Right ... double RL. Like, it has no under- concept of the house.

[00:06:39] Ricardo Belmar:

[00:06:39] Casey Golden: There's that... it's not on the website, right? It's definitely, not on reddit.

[00:06:45] Sabrina McPherson: I know. No, absolutely, and it's- it's usually scrapping from the same places, right? Yeah. A- and I think it's this thing, too, around it's not, the- it's not just marketers that need to- brand marketers that need to be speaking to the retailers anymore.

[00:06:57] Like, this is about commerce execution. [00:07:00] Right. Right? Yeah. Like, so it's- Yeah ... it's not that marketing isn't involved, of course, but getting those products sort of data layer in place that- so that the AI can recommend or pick you, that's an ex- that's commerce execution, next level now.

[00:07:12] Casey Golden: You know, GEO and AEO and like mach- your, your site being machine-readable, it's like the lowest hanging we have right now.

[00:07:20] Yeah. it's not launching a new channel necessarily. Just some basics. Yeah. And I can say like how many brands that I've done analysis on that haven't done anything yet, and we're less than, we're less than s- we're like 30, 45 days from holiday.

[00:07:38] 30 Days to Holiday and Brands Haven't Started

[00:07:38] Sabrina McPherson: I know.

[00:07:39] Casey Golden: You know, and, we don't have a commerce solution for the GEO and AEO and all of this new sales channel. We're usually just plug into a TikTok shop, or now there's an Instagram shop, right? Like it was a simple channel integration, and with this we need like a full tech stack.

[00:07:55] Sabrina McPherson: That's right.

[00:07:56] Casey Golden: Like a new stack of software that does this because have a soup to nuts [00:08:00] yet.

[00:08:00] Yeah. And so a lot of people, brands that I'm speaking to, they don't wanna sign seven contracts ...when everything's so new and half ...the stuff everybody's looking at is, pretty much vaporware. Well, and it's all like- changing- ... point solutions... every week.

[00:08:12] Sabrina McPherson: Right? It's

[00:08:13] it's like if you want a specific thing, go buy this technology.

[00:08:17] Ricardo Belmar: Right ...

[00:08:17] Sabrina McPherson: that's not... the problem. It's not

[00:08:18] Casey Golden: of people are waiting because they don't wanna get into those contracts. They don't know which vendor to choose. The vendors are changing on a weekly narrative.

[00:08:25] Ricardo Belmar: But at same time, everything's moving so fast it's like you're losing more and more the more you wait

[00:08:29] Sabrina McPherson: yes, right. That's right. So you gotta something.

[00:08:31] Casey Golden: You gotta do something

[00:08:32] Sabrina McPherson: Yeah. Yes. Yeah Agree.

[00:08:34] You're Now Selling to Two Audiences

[00:08:34] Sabrina McPherson: do you feel like commerce experiences need to start supporting both the human shopper and,

[00:08:43] Casey Golden: aI agents?

[00:08:45] Sabrina McPherson: Oh, for sure. So right now it's absolutely two audiences, right? And I think that's also the big shift, and that, that takes work. So I think humans and AI obviously value different things, right? And you have to design commerce experiences for both audiences. [00:09:00]Yeah. and I'm literally gonna talk about AI like it's a human, because it is-

[00:09:03] it is like a human. Yeah. So but real humans, right, still value the brand, experience, how it makes them feel- Mm-hmm ... the imagery, the emotion that makes them want a product, right? But the AI, right, the machine, values really good, consistent data that it can trust.

[00:09:20] Casey Golden: It wants walls of text.

[00:09:22] Sabrina McPherson: Yeah. Yes. Yes

[00:09:25] Casey Golden: Where we close our laptop and put our phone in our pocket

[00:09:29] When we see a... when we see a wall of text.

[00:09:31] Sabrina McPherson: Right. Basically. but so, so you have, you have two audiences and they value different things. Yeah. And I think the opportunity right now is for companies to basically, you, you still need the human to want your brand, but you need to find a way to translate the product, the, the differentiation of your product and your brand into language- Mm-hmm. That the machine can read. Point blank. Yeah. Like, it sounds like i- obviously easier said than done, but that's that whole tricky thing of like, what is the product data that I need, right? Right. How do I talk about- Right ...

[00:09:59] Wedding Dress vs. Dress for Wedding

[00:10:02] Sabrina McPherson: what, what data do I put in about a wedding dress- that makes, the AI understand this wedding dress beyond just its price, and that it's a dress, and it's for weddings.

[00:10:07] And it's available here.

[00:10:08] Ricardo Belmar: Yeah.

[00:10:09] Casey Golden: And it's white or it's cream.

[00:10:10] Sabrina McPherson: Right.

[00:10:10] Ricardo Belmar: Exactly, yeah. All, all the different ways you could describe it,

[00:10:13] Sabrina McPherson:

[00:10:13] Ricardo Belmar: that the AI might be looking for based on what the original search prompt was, right? Right. It has to be able to show up in, in that list but

[00:10:20] Casey Golden: I think like even super random, but like wedding dress and dress for wedding, so not the same thing.

[00:10:26] Ricardo Belmar: Question, yeah

[00:10:28] Sabrina McPherson: Yes. And like the weather, right?

[00:10:30] Like- what- what- do wear for different seasons? Right. What kind of wedding is it? Right. I mean, all of that context, because it's intent-based search now

[00:10:37] Yeah ...

[00:10:37] it all plays in more. Yeah.

[00:10:39] Casey Golden: was never part of the product catalog.

[00:10:41] Sabrina McPherson: Yeah

[00:10:41] Ricardo Belmar: That's right. Well, and so I want to come back to, you, you, touched on a minute ago th- this trust element. Yeah.

[00:10:47] So in a sense, you're also looking for the machine, the AI, to trust- exactly ... your information. Right. But not just on your website, right?

[00:10:54] You, you... Don't you also really want and need, I guess I should say you need other people- Yes ... talking about your product, and [00:11:00] you, need that to also be in alignment with how you're describing your product.

[00:11:03] Two Layers of Trust AI Must Earn

[00:11:03] Sabrina McPherson: Yeah,

[00:11:03] No, absolutely. So it's, it's two layers of trust, right? We need the humans to trust the AI- Mm-hmm ... and, and really believe that the AI is not just pushing a product, right? So, like, if you're going to Amazon and you're talking to chat, right?

[00:11:14] Mm-hmm.

[00:11:15] Ricardo Belmar: it's not gonna just push

[00:11:15] turns out.

[00:11:16] Sabrina McPherson: right? Right. Right. Whatever someone's paid for. Right. number one. And then the second thing is you need to trust-- the AI needs to trust. Again, AI is like a human. The AI needs to trust the data 'cause it's not gonna recommend your product if it feels that it's not consistent, if it's getting different signals across different channels- Mm-hmm

[00:11:31] if it doesn't think the inventory is accurate, right? Yeah. And so you-- this whole having clean data and consistent data, it's, it's huge. It's not even just about what data to have. It just needs to say the same thing across many different areas- Yeah ... that are being, know Yeah ... scraped by the aI.

[00:11:45] Ricardo Belmar: Yeah. And I coming back, to trust part, in the last, call it 10 days or so, right, there, there's new personal agent, obsession I guess I'll say out there when Meta released Muse, and they're not the only one, right? But these personal agents that it [00:12:00] seems like a lot of people are experimenting with that if we go by the download numbers in the app stores. But at, at what level do you think consumers are really going to trust these agents to go out buy for them?

[00:12:10] Sabrina McPherson: Yeah. So I think it's, it-- we have a long to go, right?

[00:12:13] Would You Let an AI Agent Buy Toilet Paper?

[00:12:13] Casey Golden: What is so hard that we need an agent to go buy it for us. I think, I think the use case first, we need to, like, narrow in on-

[00:12:20] Ricardo Belmar: Really understand use case

[00:12:21] Casey Golden: I know

[00:12:22] Ricardo Belmar: Exactly.

[00:12:23] Sabrina McPherson: There's also that

[00:12:23] Casey Golden: there's like function, utility, fun.

[00:12:27] What is this product category? I can say like, I need a list for certain complex product categories, but, like, I think we have to really nail down what is the merchandise category that the agents will have the autonomy to shop for.

[00:12:42] Sabrina McPherson: Yeah, I'd like some commoditized category ...

[00:12:44] Casey Golden: that's not... that's hard to shop- and not fun..

[00:12:47] Sabrina McPherson: Yeah.

[00:12:47] Ricardo Belmar: right

[00:12:47] Sabrina McPherson: like, or like toilet paper. Right.

[00:12:49] Casey Golden: like, but that's like easy

[00:12:50] Ricardo Belmar: yeah, that's an easy one.

[00:12:51] yeah, get that off, get that off my plate

[00:12:52] Casey Golden: button, right? So it's like, it's something hard that's not fun. And I think i'd like to... I don't know, I think I wanna make a whole list of, what is that product category? [00:13:00]

[00:13:00] Sabrina McPherson: You should ask, ask chat what it thinks.

[00:13:02] but I, but realistically, right, there's also everyone's like, "Oh, the trust spectrum from, like, we used to trust, like, AI to, to give us some responses in chat or in, in search to we're gonna have AI purchase." Like, there's a huge spectrum there. Yeah. And y- I mean, we all know that trust doesn't go from here to there, right?

[00:13:16] There's no jumping five levels ahead on no Trust gained as you go. And Right people are experimenting with, with all the new personal assistants, but we all know things are gonna go wrong.

[00:13:27] So it's gonna go this way. People are gonna fiddle, and that's a very small part of the population as well. And then they're gonna go back to being like- Yeah ... "Eh, I don't trust it. It's hallucinating. It's, like, telling me that this thing is available, trying to buy something for me, and then I got a back, like a pushback," We have a long way to go.

[00:13:40] Ricardo Belmar: Yeah.

[00:13:41] Casey Golden: What do you think this means for, like, retail and brand agents, versus the personal assistant? A lot of sites are starting to block other bots, and right?

[00:13:51] Sabrina McPherson: And that is a huge trust issue, right? So I think for a retailer or brand to get over the hurdle of a human not thinking or thinking or [00:14:00] not trusting that they're gonna actually bring back a selection of something that meets their needs versus pushing something that someone paid for- Yeah.

[00:14:06] That's gonna take a long time, right? Like- Yeah ... th- there's already so much distrust in like, every- I mean, I, I know. I'm sure- Yeah ... you guys feel it. Every time you go on, you're " like, oh, who paid for that to- surface- in my selection . Right.

[00:14:17] So if you have your own brand,

[00:14:18] Casey Golden: Well, so So much, I mean, I have to go double check to make, see if this is a dupe of something else, because because never heard the brand. Is this fake website? Right. Because, like, there's so many. I'm like, i'm trying to find the real one.

[00:14:31] Ricardo Belmar: Yeah Right.

[00:14:32] Sabrina McPherson: No, exactly!. I like it,

[00:14:33] Casey Golden: but like is this the real one?

[00:14:35] Sabrina McPherson: Yeah. Or with, within a brand, right? Even if it's in the same brand, I'm gonna always just assume that they're trying to show me something that they're trying to offload because they have too much inventory of it.

[00:14:42] Right. Not because it's the right fit,

[00:14:44] Casey Golden: we're too savvy of shoppers.

[00:14:46] Sabrina McPherson: But like, I'm gonna be like, "Well, okay, this is like the day old eggs. Like, you're trying to sell me day old eggs!

[00:14:52] Ricardo Belmar: Exactly. Yeah.

[00:14:52] Sabrina McPherson: because I want the best egg for my need

[00:14:54] Ricardo Belmar: Yeah.

[00:14:55] Sabrina McPherson: So..

[00:14:55] The Fake Product ChatGPT Recommended

[00:14:55] Ricardo Belmar: Well, and there's definitely one, like, I think it was maybe couple months ago that, someone, an influencer [00:15:00] staged a direct-to-consumer, beauty product on website that was completely made up.

[00:15:05] But they created the entire site, they gave it all sorts of product descriptions, and within two weeks it started showing up in results on chatGPT.

[00:15:11] Sabrina McPherson: Oh, I mean...

[00:15:13] Ricardo Belmar: But it was a fake product. It didn't exist and it was getting recommended.

[00:15:15] Sabrina McPherson: Exactly.

[00:15:16] Ricardo Belmar: So yeah, there's, there's that.

[00:15:17] ...

[00:15:17] Ricardo Belmar: well, Sabrina, thank you so much for, for joining us. This has been great conversation. We'll have to have you come back in the future to really dig into, as this is constantly changing.

[00:15:25] Casey Golden: Exactly. Don't forget your sunscreen!

[00:15:28] Ricardo Belmar: That's right

[00:15:28] Sabrina McPherson: Sounds good. Thank You!

[00:15:29] Ricardo Belmar: Right. Thanks everybody

[00:15:31] ​

[00:15:31] Show Close

[00:15:37] Casey Golden: I know you loved this episode, so drop us a five-star rating and review on Apple Podcasts, Spotify, or Goodpods. And if you're watching us on YouTube, like and subscribe before you go. I'm Casey Golden

[00:15:48] Ricardo Belmar: Follow us on LinkedIn, Bluesky Threads, and Instagram, and subscribe to our Substack for highlights and bonus content. For transcripts and guest info, visit retailrazor.com. I'm Ricardo Belmar

[00:15:58] Casey Golden: Thanks for joining us on [00:16:00] Retail Transformers, part of the Retail Razor Podcast Network.

[00:16:03] Ricardo Belmar: Until next time, stay sharp, be bold, and keep transforming retail.

[00:16:07] This is the Retail Razor Retail Transformers!