Rokt helps marketers reach high-intent customers in the Transaction Moment™—when they’re actively completing a purchase online. Powered by AI and first-party data, Rokt Ads connects your brand with over 400 million global shoppers and delivers outcomes you can count on. Learn more at rokt.com/emarketer

to get started today.

Sara Lebow (00:00):

In today's economy, every ad dollar counts. That's why performance marketers are turning to Rokt Ads to reach over 400 million active shoppers in the transaction moment when they're completing a purchase online. You only pay when customers engage. Learn more at rokt.com/emarketer. That's R-O-K-T.com/emarketer.

(00:25):

Hello, listeners, today is Wednesday, May 14th. Welcome to Behind the Numbers: Reimagining Retail, an eMarketer show made possible by Rokt. This is the show where we talk about how retail collides with every part of our lives. I'm your host, Sara Lebow. Today's episode topic is generative AI in retail. Let's meet today's guests. Joining me for today's episode, we have one of our analysts from our retail desk, Carina Perkins.

(00:54):

Hey, Carina.

Carina Perkins (00:56):

Hi, Sara. Thanks for having me back.

Sara Lebow (00:58):

Thanks for being here. And also is an analyst from over on our tech desk, Yory Wurmser. Hey, Yuri.

Yory Wurmser (01:03):

Hey, Sara.

Sara Lebow (01:04):

Welcome back to the podcast.

Yory Wurmser (01:06):

Great to be here.

Sara Lebow (01:07):

Okay, let's talk about generative AI, everyone's favorite and least favorite topic. I'm going to start with setting the stage. Last week, Mark Zuckerberg said, "Marketing is dead." Well, at least he said it's dying. Mark Zuckerberg said, "In a way, Facebook's AI tool is kind of like the ultimate business results machine. I think Meta's AI marketing tools will be one of the most important valuable AI systems that gets built." He essentially said that Facebook's AI, Meta's AI is going to be able to do all of marketing without marketers having to do anything.

(01:40):

Around the same time, Amazon released Nova Act, an AI agent that can independently perform actions moving us further toward AI agents buying on consumer's behalf.

(01:51):

And also happening around the same time ChatGPT started surfacing product recommendations in its feed, so I'm going to start with a basic slash not basic question that I'm going to make both of you answer. True or false, generative AI in retail will change how people shop.

(02:08):

Carina, you're writing the report on this, so true or false?

Carina Perkins (02:11):

Yeah, I mean I think true in a lot of ways. I think it's already having some influence on how people shop. Will we see the extremes that some people are predicting or Zuckerberg would like to see is perhaps a different question. I think he's also just said that we're going to replace all of our friends with AI.

Sara Lebow (02:30):

He did.

Carina Perkins (02:31):

So that's another interesting point. But I think what we are really seeing here is that AI is causing a little bit of a shift in the kind of shopping landscape, and we are seeing tech companies really go for a kind of grab of commerce rather than just supporting it. They really want to become part of the shopping journey and own the commerce themselves. And I think that's probably where we're going to see the biggest shifts happening in the next couple of years.

Sara Lebow (02:56):

That makes sense. I mean tech companies have been trying to own since their inception. Yuri, true or false, generative AI will change how people shop.

Yory Wurmser (03:05):

True. Although I agree with Carina that Zuckerberg's comment is probably going to go down with his predictions about the Metaverse. I definitely think, speaking specifically about marketing, that all parts of the marketing stage are going to be transformed by AI and agents. Marketers will still want control, and the same thing is true of consumers. They're still going to want to be the final decision maker on a lot of purchases, most purchases, probably almost all purchases. There're going to be a lot of changes, but a lot is still going to stay the same.

Sara Lebow (03:37):

Okay, so let's start with current state.

(03:40):

Carina, you're writing about this right now. How is generative AI already being used along the shopper journey?

Carina Perkins (03:47):

Yeah, I mean it's being used pretty much everywhere from discovery through to delivery, so we're seeing people using AI for AI-generated ad campaigns, product recommendation, AI search, customer service, review summaries, and even loyalty perks. But it's not without friction. So we're seeing implementation of those tools are being met with slightly different consumer responses. And also what we're seeing is that there's a lot of focus on the kind of marketing and the discovery element from brands when consumers are actually most interested in AI solving those post-purchase pain points.

Yory Wurmser (04:23):

I mean, Carina is exactly I think on point on both points. Discovery, I think, is going to, has been affected to some extent and definitely to consideration phase is already being affected by generative search.

Sara Lebow (04:36):

And generative AI is also obviously helping on the back end already. I think that's where it's most useful, merchandising, things like that where customers can't see it because this is something that you wrote, Yuri, in a report that technologists may be underestimating how much a lot of consumers don't like AI in their journey.

Yory Wurmser (04:56):

The pace of development of the technology, the underlying technology, is just mind-blowing. Every day, there's not one advanced or a couple advances. We're going to get to the point not that long from now where you'll be able to buy things automatically easily. I'm unsure how quickly consumers are going to trust that. That I think is going to be the biggest factor holding back automated buying.

Carina Perkins (05:21):

I think Yuri's got a really good point there, and it's not just how much they're going to trust it, but it's how much they want it. Because I think if we're talking about completely automating shopping, we are forgetting that some people enjoy the shopping process and they get endorphins from buying things. They enjoy the discovery. So I think that again, maybe that's slightly underestimated how much people actually still really want an active role in shopping.

Sara Lebow (05:47):

And what we're getting at here is agentic AI. Agentic AI is when AI moves from being copilot or a helper for the end user to being the actual entity making decisions and in this case making purchase decisions on behalf of the user. Did I get that definition right, Yuri?

Yory Wurmser (06:04):

Yeah, I mean it's basically, it's a task-oriented program that does what you want it to do.

Sara Lebow (06:10):

So how do retailers and technologists expect agentic AI will impact the shopper journey and how do you guys expect it will impact the shopper journey?

Yory Wurmser (06:19):

Technologists envision a future where these agents will anticipate what you want, will suggest things for you or even buy things for you if it's time sensitive or it's something that's somewhat routine and will help you pretty much in every stage of the shopper journey.

Sara Lebow (06:37):

And what am I doing in the meantime? Am I just working more? What am I up to during that time?

Yory Wurmser (06:43):

Yeah, it frees you. I think the thought is it frees you from the parts of shopping you don't want to do, and I actually think some people will really buy into it. The question is how many people. I have no doubt that there are going to be millions of early adopters that will be really into this stuff, but others I'm not so sure about.

Carina Perkins (07:00):

It will really depend on what you're buying. So it's great for auto replenishment, it's great for things that are boring to buy, diapers, dog food, things like that, but it's perhaps less exciting for things where the kind of discovery and the research is more emotional, so clothing and things like that. So I think it's really going to have a different impact depending on what kind of products we're talking about.

Sara Lebow (07:22):

Yeah, I see it as useful for maybe industrial shopping, buying supplies from an office, for example. Even for replenishment though, I mean not everyone uses subscribe and save on Amazon, and that is available to them. If people aren't using that, then why would Amazon believe those people might use a robot that then picks what they're subscribing to and saving on essentially?

Carina Perkins (07:46):

It does reduce customer control a little bit, and transparency is going to be really key here because there is a potential that these, it's going to reward whoever has the best AI integration rather than what is the best product for the consumer. And I think unless there's a real transparency around how it's coming to conclusions or coming up with decisions and products and suggestions, that's actually just really going to put consumers off because things like sponsored product listings are already creating a little bit of friction. And I think if consumers feel that they're just being served something which is just repetitive or not actually the best product for them, that's going to be a real problem.

Sara Lebow (08:25):

Yeah, I mean, let's talk more about problems. What is something in generative AI in retail that you each feel is being over-hyped right now?

Yory Wurmser (08:35):

Yeah, I mean, just to go back what I said earlier, I do think they're going to be some people who will be fine having automated buying, but I think in general that's over-hyped. I think especially older consumers are going to be pretty hesitant about that. So yeah, I think the area for me that is the most over-hyped.

Carina Perkins (08:53):

I'm going to be a bit controversial here, and it's not because I don't think it's a brilliant thing, I just think that it's not happening yet, and that's seamless personalization. So that is possible. That is the great sort of future potential of AI. But at the moment, according to consumer surveys, we're still seeing that AI is guessing wrong or it's overselling or it's kind of bombarding users with ads that they don't want. There was one server that found one in five consumers say that recommendations are still way off base. And so I think until that personalization is really honed and effective and accurate, I think that is a little bit hyped at the moment.

Sara Lebow (09:31):

I agree with both of you. I think it's an important time to remember that 80% of shopping still happens in store. So even what we're talking about right now for the time being at least is 20% of shopping, which is a huge amount, this is in the US, which is a huge amount, but it's a fraction of shopping and we're going to be talking about a fraction of that fraction.

Yory Wurmser (09:51):

But do you know what? Agentic AI is also going to have a big impact in store.

Sara Lebow (09:56):

Oh, say more.

Yory Wurmser (09:56):

And it can already do that. Just if let's say you take ChatGPT or take a picture of a shelf and a product on a shelf, it can give you reviews. It can tell you the price that you can get elsewhere. So even in-store shopping can be changed pretty fundamentally through agents.

Sara Lebow (10:12):

I'm going to put that in the bucket of one thing retailers and brands aren't paying enough attention to with generative AI, which is what I was going to ask you next.

Yory Wurmser (10:19):

Carina, you had something to say I think on that one too.

Carina Perkins (10:22):

Yeah, no, I was just going to say I totally agree with that. I think it's sort of potentially in store is really interesting and that kind of omnichannel piece. But again, I think it's convincing consumers to use the tool, so it's kind of changing that shopping habit, and that can be slow, but I think it will be quite transformative once it does start happening.

Sara Lebow (10:43):

I think something that is going to inhibit adoption is also fragmentation in this process. So yeah, if you have a chatbot through Amazon, then that can be integrated with Amazon products, that can be integrated across parts of the open web as well. But ultimately you're limited to Amazon. We already have all of these walled gardens that we struggle to work in between. There's not really any reason an AI agent would change that. In fact, I think a lot of companies are going to be even more cautious about keeping their walled garden walled in.

Carina Perkins (11:17):

Yeah. And I think that also speaks to a broader point about the need for better data infrastructure generally. Without really clean structured product and customer data, gen AI agents going to fall flat. So I think that data readiness really needs to accelerate.

Yory Wurmser (11:34):

I mean, one counter-argument to that is if a general kind of a universal agent takes off something like a super-powered Siri.

Sara Lebow (11:44):

Like a ChatGPT.

Yory Wurmser (11:45):

Or a ChatGPT, that's actually the main interface you have with your phone, that could break down some of those barriers into walled gardens because then it becomes the orchestrator. I mean that it will send you to the right walled garden for whatever that need is.

Sara Lebow (12:01):

Yeah, I mean a fair criticism of what I'm saying here that a technologist would probably say is that I am not thinking big enough. I'm thinking small, and I'm thinking current. On that note, what is one thing retailers and brands aren't paying enough attention to with regard to generative AI right now?

Carina Perkins (12:17):

So I would say the post-purchase experience. So as I said in my report, what I really found was there's been several studies that's showing a bit of a gap between where retailers are focusing their attention on AI, which is that top of the funnel, and where consumers really want that additional help and see the value in AI. So I think simplifying returns, offering smarter loyalty programs, tracking deliveries, and I really think the retailers that use AI to kind of fix some of those problems are going to be the winners.

Yory Wurmser (12:47):

I'm going to go with the impact of generative search and these agentic tools. That's going to have a big impact, first of all, on some of the more, the importance of things like reviews and price because they'll be ingested and synthesized by these generative search engines. So optimizing for search, paying attention to things like reviews and objective criteria, product specs I think are going to make a bigger difference than they currently do. And kind of an extension of that, I think price, there's going to be a lot more price.

Sara Lebow (13:24):

Price competition.

Yory Wurmser (13:26):

Price competition.

Sara Lebow (13:27):

Yeah. Yeah. I mean if you can pick out the cheapest high quality item, then that'll push the prices down theoretically.

Carina Perkins (13:35):

Yeah. I think further to Yuri's point as well, brands need to be thinking about how they're optimizing their websites and their data so that it is optimized for these gen AI tools. You want to make sure that you are going to be discoverable in ChatGPTs, shopping, however you do that.

Sara Lebow (13:53):

I sound like such a Luddite here, but if we have agents making shopping decisions on behalf of us on websites where agents are making merchandising decisions, what's the point? What are we up to? What am I doing?

Yory Wurmser (14:11):

Yeah. So I think if you focus on the parts of shopping you don't like, and that's I think what will disappear most quickly because there's an obvious need. But if you're shopping for a car, you're right, you do want to spend the time to actually go through it, try different things, research deeper and deeper. So you'll still be doing stuff, but the stuff that you most want to do, and theoretically you can get rid of this stuff that's tedious.

Sara Lebow (14:37):

Yeah, even retail, I think a big part of shopping that people don't like is looking at their body in a mirror over and over again in different clothing. Can an AI agent help that? I don't know.

Carina Perkins (14:52):

Not yet, but yeah, probably one day, right? I mean, we have AR, there's all those capabilities when they come together have the potential to completely transform the shopping experience.

Yory Wurmser (15:03):

These virtual try-on tools are forms of generative AI.

Sara Lebow (15:06):

That's a fair point. Yeah.

Yory Wurmser (15:08):

Yeah. I mean, it could and as they get better and better, it definitely could reduce some of that.

Sara Lebow (15:13):

Yeah. Of course, you'll have to have virtual try-on tools that can report, this is like what you said with specs, that can report up accurately on size. We've talked before on this podcast about how it seems like sizing is a mess right now. Measurements are a mess. That has to do with production quality. Yeah. And theoretically, generative AI can fix that too, but we're talking about a massive system where everyone needs to be bought in, and we're definitely not there yet.

Carina Perkins (15:38):

No, and I think, again, this comes back to a broader point of it's very easy to get obsessed with the new shiny tool. And yes, we will have these AI agents that can automatically purchase something for somebody, but are consumers actually more concerned about the little frustrations in their day-to-day life with the shopping? Yes. So I would recommend that people use technology first to address those frictions and make it as seamless as possible because that's what consumers face in their day-to-day life, and that's going to be the bigger result in the end, I think.

Sara Lebow (16:15):

Okay. So Carina's advice to retailers is focus on day-to-day frictions, make experience seamless for consumers?

Carina Perkins (16:24):

Yeah. Focus on the basics, get the basics right, and once you've got the basics right, then you can start looking at wonderful, transformative new shopping experiences.

Sara Lebow (16:32):

That was Carina's one piece of advice. I think my advice right now as these kinds of search engines become more popular in shopping is to pay really close attention to how people are talking about your brands across the internet, in reviews on social media, and honestly and particularly on Reddit, which is a place that a lot of AI is scraping from and is a place where a lot of people talk about brands.

Carina Perkins (17:03):

That's a great one.

Yory Wurmser (17:03):

Yeah, no, that's excellent. You guys took the two best ones. I'm going to say just focus on first-party data. I mean, this is just a kind of an evergreen quote, but first-party data is going to first of all, help you design the products that you want, it's going to help you market better. It's probably going to help you, without the first-party data, your own internal agent tools aren't going to work well and personalization is going to be much harder.

Sara Lebow (17:31):

Great. All right, well that is all we have time for today. Thank you to our listeners and to our team that edits the podcast. They're always generating a great podcast. We'll be back next Wednesday with another episode of Reimagining Retail: An eMarketer Podcast. And on Friday, join Marcus for another episode of The Behind the Numbers Show, the eMarketer podcast made possible by Rokt.