Behind the Numbers: How the Rise of the AI Shopping Agents Is Impacting the Customer Journey

On today’s podcast episode, we discuss how GenAI is changing the customer shopping journey the most, the impact of AI agents, and how to maintain brand messaging in a more conversational universe. Tune in to the episode with Senior Director of Podcasts and host Marcus Johnson, our Analyst Jacob Bourne, Vice President Suzy Davidkhanian, and Global Digital Commerce Senior Director, Strategy & Execution, Todd Hassenfelt. Listen everywhere and watch on YouTube and Spotify.

Subscribe to the “Behind the Numbers” podcast on Apple Podcasts, Spotify, Pandora, Stitcher, YouTube, Podbean or wherever you listen to podcasts. Follow us on Instagram.

Episode Transcript:

Marcus Johnson (00:00):

Closing the intelligence gap between data and insights is the key to transforming marketing from a cost center into an engine for growth, but where do you start? Great question. Find the answers in Zeta Global's latest playbook, Driving Growth in the AI Era. It's very free and you can download today, link in the show notes.

(00:26):

Hey gang, it's Friday, February 28th. Jacob, Suzy, Todd, and listeners, welcome to Behind the Numbers: an EMARKETER video podcast made possible by Zeta Global. I'm Marcus. Today we're discussing the combination of GenAI and shopping and what happens when those two things get together. For the conversation, I'm joined by three people, we have with us our Technology Analyst based on the West Coast, it's Jacob Bourne.

Jacob Bourne (00:50):

Hello Marcus, thanks for having me today.

Marcus Johnson (00:52):

Hey fella. Of course, of course. We also have with us our Vice President and Principal Analyst who heads up our retail desk living out life in New York City. It's Suzy Davidkhanian.

Suzy Davidkhanian (01:01):

Thanks for having me.

Marcus Johnson (01:03):

Of course. Finally, we have the Global Director of Commerce, Senior Director Strategy and Execution based in Chicagoland at Colgate Palmolive, it's Todd Hassenfelt.

Todd Hassenfelt (01:13):

Hello. Really appreciate to have the chance to be on the show.

Marcus Johnson (01:16):

Yes, welcome, welcome. Thank you for the time. We start with the speed intro to get to know our guests a little better. I informed Suzy and Jacob that they're not exactly going to be part this and they were very upset and nearly left the episode.

Suzy Davidkhanian (01:29):

Very upset.

Marcus Johnson (01:31):

You know those guys. Most of the questions are for Todd. They'll get to participate on the last one. 60 seconds on the clock. Let's do it. Todd, you are based in Chicagoland, but where are you from?

Todd Hassenfelt (01:41):

From Milwaukee.

Marcus Johnson (01:42):

Oh, very nice. Are you all the sports fan, fan of all?

Todd Hassenfelt (01:47):

I'm a little bit odd. I'm Packers for NFL and then I'm Chicago For everything else, I'm Cubs, Blackhawks, Bulls, Illini. Depending on what side of the family, my wife is all from Chicago.

Marcus Johnson (01:59):

Okay.

Todd Hassenfelt (02:00):

Depending on the season, I'm upsetting someone or bonding with them.

Marcus Johnson (02:05):

What do you do in a sentence?

Todd Hassenfelt (02:06):

I help find, share, and solicit solutions across the globe at Colgate Palmolive as part of our global digital organization. Really sharing best practice insights so we can outperform and accelerate our digital commerce growth.

Marcus Johnson (02:20):

Very good. What's your morning drink?

Todd Hassenfelt (02:23):

I'm not a coffee person. I've only had a half a sip of my life. I do Optimum Nutrition's Amino Energy one whenever they're ready to drink. Used to work there. It's a great product.

Marcus Johnson (02:35):

Not a drop?

Todd Hassenfelt (02:37):

No. Half sip. That's it. That was enough sample size for me.

Marcus Johnson (02:41):

Did you have it black? Is that what it was?

Todd Hassenfelt (02:45):

Yeah, it was that an Old Country Buffet, probably. I don't know if I put a whole bunch of cream and sugar, if it was black.

Marcus Johnson (02:50):

That's your problem. That's the problem. Last question. First record that you own, cd, cassette, 8-track maybe.

Todd Hassenfelt (02:59):

Ooh, I was a cassette. Beastie Boys Licensed to Ill, the Fight For Your Right (To Party) and many other songs that are somewhat well known.

Marcus Johnson (03:11):

Good choices. Suzy?

Suzy Davidkhanian (03:14):

I mean of all the questions, this is not one that I have a really good answer for because I don't remember that far back. It's like 35, 40 years ago. I think it was Madonna. It might've been Cyndi Lauper, it might've been Tiffany. I don't know. One of these.

Marcus Johnson (03:29):

You're just naming artists who were popular back then.

Suzy Davidkhanian (03:33):

In the eighties, yeah.

Marcus Johnson (03:34):

Okay, good.

Suzy Davidkhanian (03:35):

I was a kid with obviously cassettes.

Marcus Johnson (03:37):

Okay. Jacob?

Jacob Bourne (03:40):

Yeah, I mean similar issue to Suzy, it was definitely a cassette, definitely some sort of rock album. I think it may be might've been Def Leppard, but I'm not a hundred percent sure just because of a long time ago.

Marcus Johnson (03:52):

Okay, very good. Well they are the three guests we have for you today. Welcome to the show folks. We start though before the actual content with the fact of the day.

(04:05):

Cows have regional accents. Okay, let's do it. Language specialists from the University of London studied cows and they found that cows from different herds have different moos.

Suzy Davidkhanian (04:25):

Is this true or false?

Marcus Johnson (04:27):

It's true. What do you mean? Is it true or false? You question the fact of that.

Jacob Bourne (04:28):

Well, it's true in birds. We know this about birds, so why wouldn't it be true in cows too? Yeah, different accents.

Suzy Davidkhanian (04:34):

The same type of bird?

Jacob Bourne (04:34):

Different accents, different dialects. Different languages.

Suzy Davidkhanian (04:37):

Wait, the same species or different types of birds?

Jacob Bourne (04:40):

I think different species. Different species depending on where they are.

Suzy Davidkhanian (04:42):

That makes sense though.

Jacob Bourne (04:43):

Yeah. A newcomer needs to learn the new local accent and dialect.

Todd Hassenfelt (04:50):

I mean, being from Wisconsin, I probably should be more well versed, but I'm not.

Marcus Johnson (04:56):

The other day we learned that cows have best friends according to the University of North Hampton, which is near my hometown. How are these university professors spending their time? You'd be like the Dean's like, "Oh, professor Williams, how's your research going this semester?" "Cows have best friends," so it worked out. This is insane. Cows. Here's a better fact for you. Cows have nearly 360 degree vision because their eyes are on the side of their head. Humans have about 180. Do you know where that fact came from? Cabot Creamery, who are apparently doing more serious research than English universities.

Todd Hassenfelt (05:35):

How does that impact the phenomenon of cow tipping then? How do they not get out of this?

Marcus Johnson (05:41):

You have to come up apparently from directly behind them.

Jacob Bourne (05:44):

Well, you do it at night. You do it at night.

Marcus Johnson (05:47):

Yeah. What do you mean you do it at night, Jacob? Something to say?

Todd Hassenfelt (05:50):

I have no experience. I've only heard about others.

Jacob Bourne (05:52):

Actually. Yeah, this is definitely second hand.

Marcus Johnson (05:54):

What you want to do. Anyway, today's real topic, how GenAI is changing the customer shopping journey. All right folks, let's talk about it. Todd, I'll start with you and we'll start broad. We'll start with the theme of the whole episode, which is how GenAI is influencing, is shaping, is changing the customer shopping journey the most. There's lots of ways, I'm sure it's impacting it, but Todd, what comes to mind when you think of how it's influencing this space the most?

Todd Hassenfelt (06:27):

Yeah, I think the timeline to summarize would be from a consumer impact, it's been reviews, research, and recommendations. You think about almost probably over a year ago now, Amazon and some others followed, summarized all the reviews. Now you didn't even have to look at the first page of six or seven or go scrolling through, it gave you a GenAI summary, which is more convenient. Then the recommendations, you don't have to ask anyone. You can ask whatever questions you want and depending on the accuracy, you got a lot of ideas that may have been different than a search term that you searched and you got some more information. Now the recommendations from that, from Rufus let's say, do people like them or not, based on the research now that they've done that everything has been easier. I think those are the three biggest ways. I think from a brand perspective, quickly, it's the content which does impact the consumer.

Marcus Johnson (07:21):

Yep.

Todd Hassenfelt (07:21):

All the reports that you guys have done and others suggest that marketers are really using GenAI the most for content creation or optimization.

Suzy Davidkhanian (07:30):

It's interesting, I would've summarized one level higher, although I love the three R's, which is AI is just making the customer journey more efficient for consumers so that they get to the product that they think they want or that they were actually looking for faster. The summary of the three R's is perfect because that's the most tangible, the biggest wins, but I think it comes at a price in terms of losing the treasure hunt and losing the impulse purchases. I'm sure we'll talk about that more.

Jacob Bourne (07:59):

Yeah, I mean I'm looking at this from two angles. One is just more shopping activity. Actually Google, when they implemented their AI shopping features, they saw a 13% jump in digital shopping activity from that. Then there's also the coming AI agents where we're going to see the whole online shopping journey automated. I think those are the two areas, potentially more shopping activity and then the automation aspect.

Suzy Davidkhanian (08:25):

The Google one is interesting because I think we would answer the question differently. If you think about in-app AI tools, so like the Rufus versus the Google. Any GenAI that can tool, that can transcend to the apps and surf many different brands and many different retailers, will certainly have a different value proposition than Rufus.

Jacob Bourne (08:48):

Yeah, that's a good point.

Todd Hassenfelt (08:50):

Yeah, reducing the incredible amount of choice out there. I think even if going back to Amazon, a couple weeks ago, they kind of put down and shut down Amazon Inspire, which was supposed to be their kind of social commerce piece, but in their pseudo announcement it says, "Go to Rufus, use AI recommendations." They're really pushing that narrative and shut down Amazon Inspire.

Marcus Johnson (09:13):

In November, Chief Commerce Strategy Officer at Publicis and Forbes contributor Jason Goldberg wrote that, "Perplexity's launch of Shop Like a Pro," speaking of shopping agents, it's this new AI powered shopping assistant, "may be the tipping point," he said, "for a new era of AI agent commerce." He presents a world where you ask Alexa to find the perfect winter coat. Instead of just offering a list, it considers the current season's weather in your location, your past style preferences, your budget, customer reviews to present the ideal option or universe where you have Siri manage your grocery shopping by using your Instacart history, your calendar, your family's preferences to ensure that you're never going to run out of ingredients for school lunches, he says, responding to a simple command, "Hey Siri, keep my pantry stocks." Todd, it comes to you first. What do you think will be the impact of AI shopping agents?

Todd Hassenfelt (10:08):

Yeah, and I've tried out the Perplexity one. I think just quickly on that one, it's interesting even when you do a search of, "What's the best toothpaste for me," or, "What's a different product," I search a lot of different products, not all the results have that little tag that you can use the Perplexity Pro, it's only certain retailers. How should you think about that and is that a blocker or not? I think overall with all of these, there's still some hesitation from consumers to use these. It is that balance of trust and accuracy and then the convenience. If it's convenient but wrong or giving recommendations that aren't good, not great. How can Brian's kind of ensure that AI approach?

(10:46):

I mean ultimately it's going to come down to data quality in the back end too for brands. How well did you set up that new item setup form? How well are you really updating whether it's keywords that are addressing questions now, then even how do you see it through the consumer's eyes so that you can see what are the follow-up questions and does my content relate to this and am I getting recommended, let's say, at my market share. This may not be just an online thing, this could be an in-store thing when you think about smart cards or in-store navigation apps where maybe you take a picture of your shopping list or upload it and then it gets that for you, you just go look for some new stuff or the shop cart, the smart cart just kind of guides you to certain products. This isn't necessarily just an online thing, that's where everyone's looking now, this could be an in-store thing as well.

Jacob Bourne (11:37):

Yeah, I mean to Todd's point about the importance of data quality, I think with AI agents, we're really looking at brands having to now optimize for an AI audience in addition to the traditional human audience. I think it's going to be a learning curve and a balancing act, but I think AI is just really probably going to be the biggest shakeup in digital commerce since social media.

Marcus Johnson (12:01):

Wow.

Suzy Davidkhanian (12:02):

It's interesting because from my retailer perspective, I don't know, I think if you don't have an interface that's universal, so let's take Perplexity versus the Rufus. You already have to start with a universal one that you as the consumer trust, like Todd was saying, enough to give it access to all kinds of apps so that it could answer the question of buy a birthday present for my mom and make sure that she likes it in this price point and it reminds you, it makes her life better and is like, "Oh, your mom's birthday is in two weeks and she lives in Toronto." Ha, I answered where I'm from, so she can get her present on time.

(12:39):

That's all well and good, but does that not replace the idea of as Todd had, the recommendation, so the personalization? Are we just trading off personalization and then the dollars are going to start being spent in this sort of have a Perplexity or some other AI agents start servicing ads but in a different way so it feels personalized, but it's also not, I don't know. I think there's a little bit of potential of mistrust that could happen that if consumers are not a hundred percent comfortable with the recommendation being organic, they will have a bad experience and not come back.

Todd Hassenfelt (13:18):

A lot of this is going to be based on your shopper history, your glance views, what you've searched, and that's already baked in. I wonder if consumers would feel better if they had the opportunity, something they saw on TikTok, something they saw in a store wherever, and they could almost snap a picture of it and send it now instead of Google to look for it, you send it to your shopping assistant or your shopping agent and now it finds it so that way you got skin in the game and they're finding it and does it help or here's what I like. One, that's a little bit more trust on it, but I think also from the brand speak part of it, how do you get your innovation in front of consumers now?

(13:53):

If you're thinking on a retailer site and with these shopping agents, it's kind of like subscribe and save or auto-ship just accelerated now. What you want if you're the product on subscribe and save, but how do you get those new items, those new SKUs in front of them? You have to get creative with the content probably, and you have to really think about maybe paid Rufus recommendations. Maybe that's where that comes in.

Suzy Davidkhanian (14:16):

Even how do you get someone to even know that there's a shopcolgate.com, right? If you are only using the AI in your wholesale channels, then you're not going to get that direct to consumer sort of traffic. If you could have your own agent, then that would be the coolest, right, that understood you. I think Google's tried to do that a few times in different iterations.

Marcus Johnson (14:43):

Jacob, I want to ask you about that real quick. I mean, what we're talking about is so interesting. Mr. Goldberg from Forbes contributor was also saying the rise of AI agents are challenging traditional retail websites, what you guys are talking about, they've been the primary digital interface for consumer engagement, but for routine purchases, AI agents rendering them less relevant, shifting their roles to moments of inspiration and trust building, where those can be more and more paramount, besides also where you upsell and where you push impulse buys as well and collect data about your customers. It doesn't seem like something that retailers or brands are going to want to give up easily. Are we going to live in a world of multiple agents? Nicole Silberstein of Retail Touchpoints was noting that there are all kinds of companies who either have one or working on one saying Walmart, Amazon, target, Victoria Secret, IKEA, Instacart, et cetera. Are we going to just live in a world where we have 50 different agents working for us at any moment?

Jacob Bourne (15:45):

Yeah, I mean I think it's important to point out here too, that this technology is not at all mature. I mean it's still under development, but I think in the future this notion of one or more agents is going to be, the lines are going to be blurred. I mean, you're going to have an agent that really is probably multiple agents in one, or at least able to interact with other agents to complete some type of complex tasks. I think it's going to be different than thinking about an individual consumer. I think it's going to be more of a hive mind working in the background.

Marcus Johnson (16:21):

Mm-hmm.

Todd Hassenfelt (16:22):

From the consumer lens without the tech aspect of it, but do they look at agents and assistance differently? Depending on the product category, maybe even the length of research, agents maybe do more of what you would do is subscribe and save for more center store stuff where assistance, "Hey, I want to research it, I want to make the decision." Both are helpful, but in different ways for different categories.

Suzy Davidkhanian (16:43):

I think it depends on how you're defining Marcus, the 50 agents. If you shop at three different stores and they each have their own agent, then already you have three. Will we have multiple individual brand agnostic agents? I'm not sure that anyone can handle that many, right? You'll just go to your preferred platform. From my perspective, what I think is most important is if the agent can identify, so agent or assistant whichever, can identify the purpose of your trip before you even start going down the search rabbit hole, then they can help direct you to impulse fun versus get it done, "I need new toothpaste, so give me the whitening one, put it in my basket, let's move on with my life." I think because AI, GenAI particularly, as supposed to be smarter and more, they might be able to understand some of those nuances and move you in that right direction, which would be win-win for a retailer.

Jacob Bourne (17:42):

I think that's kind of the dream of having an AI agent and also can kind of act like an assistant, in that it understands the consumer it's working on behalf of. It knows the person's interests, it knows its kind of consumer profile and can kind of act with that in mind.

Marcus Johnson (17:59):

Yeah.

Todd Hassenfelt (18:00):

Can you get increased conversion rates on the same traffic and reduce your returns, which are constant issue all over the place, right? Will it help with that? Potentially. I think there's an option, but also retailers and brands, how do you negotiate this in a JVP, joint business plan, where could it be leveraged where the retailer could turn off your SKUs as a punitive thing for a time being for a shopping agent or an assistant results? Maybe.

Jacob Bourne (18:29):

Yeah.

Todd Hassenfelt (18:31):

I think it's going to be interesting just how you negotiate for data, how you work with this, and then who helps the consumer ask the right questions of their shopping agent?

Jacob Bourne (18:40):

Right.

Todd Hassenfelt (18:40):

We haven't done a good job as consumers with Alexa, we ask it to play music, set a timer and an alarm, how do we know what are the right questions and do I want a lot of reviews or do I want a little reviews because that's a newer item. What is the price? What are the kind of the questions I could ask of my agent? How does that get disseminated throughout the consumer community.

Jacob Bourne (18:59):

Yeah.

Marcus Johnson (19:00):

Jacob, how close are we to having agents that are ready to perform at a level that they need to? We already mentioned this about the data piece of it being paramount, but Maxwell Zeff of TechCrunch was saying real-time information is key, shopping agents, they're scraping retailers websites and giving you information about their products, but there's a disconnect between what Perplexity tells you and what stores actually have in stock of a time lag. How close are we to agents that actually act as though humans shopping for themselves?

Jacob Bourne (19:30):

Yeah, I mean, I think the technology is close. Part of the issue is that there's a lot of risk. There's heightened risk with these agents in terms of digital privacy and cybersecurity.

Marcus Johnson (19:42):

Mm-hmm.

Jacob Bourne (19:43):

That's for instance, why OpenAI really delayed the launch of its operator agent. I think that part of getting the technology to where it's mature, that we can say we really trust it to do all this online automation, is really trusting that this is not going to result in more security breaches and privacy lapses.

Todd Hassenfelt (20:09):

Yeah.

Jacob Bourne (20:10):

I think that's a big thorny issue for this technology.

Marcus Johnson (20:13):

Todd, let me start with you for this final question. Going back to that Forbes piece, Mr. Goldberg wrote, "Today, brands spend billions on search ads and retail media to capture shopper intent at the point of purchase." He says, "What happens when AI assistants eliminate that step altogether? If a consumer asks an AI to reorder peanut butter, where does the brand place its ad?" Todd, how do you maintain brand messaging in a more conversational universe?

Todd Hassenfelt (20:45):

Yeah, I agree with the point. It can happen to a degree, but it's not going to go away completely.

Marcus Johnson (20:50):

Yeah.

Todd Hassenfelt (20:50):

I just stood in back of someone writing a paper check the other day at the grocery store, so this isn't going to really disappear that quick. As I think about, well then it's your more, if you still believe in the funnel, it's more top of funnel, you get it out there to get your brand name out, the remembrance, get your name out, whether it's social to connect to TV, et cetera, but then also I think this is probably building a case for the in-store retail media piece, right? Let's say there is a big loss of spend for onsite retail media, well retailers and maybe brands are going to say, "Well, where do we reallocate this money?" In store where they're still shopping. The point of purchase is there, depending on where the smart cart's doing for them, but I think this is a case for that, and it'll be interesting how that shift happens.

Marcus Johnson (21:35):

Mm-hmm. Yeah. Suzy, any thoughts here?

Suzy Davidkhanian (21:38):

I mean, for me, the ads start to feel a little bit more like fake personalization, like bot personalization in a, I don't mean it in a ugly way, I just mean retailers and brands start to share information. They know so much more about the people that are there and the ads are there to help make my life as the consumer easier. I, as the consumer, fully trust this process now. That's why there's a lot of data that shows that consumers don't actually mind the ads. They find when you think about this retail media world, that they actually find it helpful to get to what they need faster.

(22:16):

When I think about GenAI and where the ads will sit, I think privacy is an issue. Risk is an issue. At the end of the day, whether it's in store or online, from a consumer perspective, if you can make my day go by faster, take out some of the tasks, makes them easier with less friction, and to Todd's point, not everybody understands how to use it, help me understand how to use it and I see value in it, then I'm more likely to adopt it. I think maybe five, 10 years ago people hated the ad sponsored stuff, it just felt like clutter, but retailers have understood how to take that away and make it part of the authentic messaging so that it feels more familiar and easier to navigate.

Todd Hassenfelt (23:01):

Will it evolve to a point of we have sponsored brands and sponsored products now in traditional onsite retail media, will you get to a point of, I don't know sponsored bots or sponsored choice, AI choice that somehow the shopping agent is seeing something there. If they do well, should that be a different CPC, like a cost per click than a human looking at it? Now we have good bots versus bad bots because the invalid traffic concerns, we always say, "No, we don't want it," now we'd be encouraging it. You have a new layer and acronyms, but of measurement and potentially, potentially, costs from CPCs or CPM perspective.

Jacob Bourne (23:39):

Mm-hmm. Yeah. I mean, to kind of piggyback on what Suzy was saying, I mean, yeah, so consumers don't necessarily mind the ads. They don't necessarily mind the personalization and actually expect it as long as a privacy is maintained. I think in the AI era, the essential thing is that you have all this automation. I think the mistake is to think of it as just a productivity tool versus something that enhances the human element in crafting the content. If you are able to really keep the human voice within the content and able to still connect with the human consumer in an authentic way, I think that AI is something that just adds value versus detracts from the experience.

Marcus Johnson (24:30):

There's something in here, there was a survey from AI powered SCA platform, Chadix, 70% of consumers feel emotionally manipulated by AI shopping assistants. It didn't go into why, the reasons for that, but just the fact that so many feel emotionally manipulated, maybe not realizing or understanding where these suggestions are coming from, are they paid for, are they not, et cetera.

Jacob Bourne (24:55):

Yeah.

Marcus Johnson (24:56):

It's going to be something that industry's going to have to figure out.

Suzy Davidkhanian (24:58):

You know what? Once people understand that, "Oh, you're only showing me ads that there's inventory for because you have that computational capability of doing it so quickly that you're not showing me an ad for something that isn't available right now for me to purchase in my location," and they start to see how it makes their life that much, reducing the mundane tasks and making it faster, then I think you'll see that number go down significantly.

Jacob Bourne (25:22):

Yeah.

Suzy Davidkhanian (25:22):

I think people just don't understand yet the power of this helper.

Jacob Bourne (25:28):

Yeah, I agree with that. I think also just because people feel manipulated by something, also doesn't mean that they won't use it. They might continue to use it even while they feel negatively about it.

Marcus Johnson (25:28):

Yeah. Todd, final thought.

Todd Hassenfelt (25:40):

This is similar to like 25 whatever, two, three decades ago where we didn't trust online banking. We thought all our money was going to get stolen. Now people rarely go inside of a bank or online dating. It was terrible to meet someone online. Now it's the norm. I think, yeah, it's a progression here.

Marcus Johnson (25:54):

Yeah. Yeah. All right, gang, that's what we've got time for, unfortunately for this episode. Thank you so, so much for hanging out for the conversation. Thank you first to Todd.

Todd Hassenfelt (26:03):

Thank you so much. Really enjoyed the conversation.

Marcus Johnson (26:05):

Yes, sir. Thank you to Suzy.

Suzy Davidkhanian (26:07):

Thanks so much for having me.

Marcus Johnson (26:08):

Yes, indeed. Thank you of course to Jacob.

Jacob Bourne (26:10):

Thank you so much. It's been a pleasure.

Marcus Johnson (26:11):

Yes, sir. Thank you to the whole editing crew, of course, victoria, John, Lance, Danny, Stewart runs the team, and Sophie does our social media. Thanks to everyone for listening in. We hope to see you on Monday for Behind the Numbers: an EMARKETER video podcast made possible by Zeta Global. Happiest of weekends.