Jacob Goldstein
Jacob Goldstein spent more than a decade as co-host of the Planet Money podcast. He's also the author of the book Money: The True Story of a Made-Up Thing, which the New…
Join Pushkin+
Gain access to ad-free versions of 20+ podcasts from the Pushkin library along with exclusive bonus episodes and other member benefits.
Note: Today’s show is sponsored by PayPal.
Kartik Hosanagar is a professor at the Wharton School. The problems he studies include: What will it take for people to trust AI agents to buy stuff for them? And how should businesses prepare for a world where their customers include AI agents?
In this episode, Kartik explains:
- Why trust in agentic shopping will arrive category by category, not all at once
- How a single old Reddit post erased a thriving brand from ChatGPT’s recommendations
- Why someone needs to take on the liability risk when purchases go wrong
- How treating AI as just another sales channel is the mistake the movie studios made with Netflix
- How to get a language model to pay attention to your brand
Connect with us:
- Follow Jacob Goldstein on LinkedIn, X and Instagram
- Email us at problem@pushkin.fm
- Follow Pushkin on Instagram, LinkedIn or X
- Listen to Jacob’s other show, Business History
- To listen to the show ad free, sign up for Pushkin+
Pushkin.
00:11
Speaker 2
Hello, hello, Malcolm Gladwell here. Over the past several months across revisionist history and business history, we've been exploring the idea of trust, where it comes from, the moment it's earned, and where we decide to place it. This episode is sponsored by PayPal, trusted by consumers, built for all business. For over 25 years, they've been at the forefront of commerce, helping merchants of all sizes grow. So wherever commerce moves next, you can rest assured PayPal has your back. And now we want to leave you with one final discussion on this special episode of What's Your Problem? Today, Jacob Goldstein hosts a conversation about how consumers and companies are trying to figure out trust in the era of agentic AI.
00:59
Speaker 1
Let's listen.
01:11
Speaker 3
I'm Jacob Goldstein, and this is What's Your Problem? My guest today is Kartik Hosanagar. Kartik is a professor of operations and marketing at the Wharton School at the University of Pennsylvania. And I wanted to talk with him because we are at this moment when the way people use AI is shifting. We're starting to go beyond just having AI answer questions and toward having AI actually do stuff in the world for us, including having AI agents go out onto the internet and buy stuff. Karthik has been studying this shift. He's been trying to figure out how both customers and merchants can solve the problems that arise in the era of shopping with agentic AI. What will it take for customers to trust AI agents to make purchases on their behalf? And what do merchants need to do to prepare for a world where some of their customers are AI agents? In the first part of the conversation, Karthik and I talked about what this all means from the consumer's point of view. In the second part, we talked about what it means for merchants.
02:14
Speaker 1
You know, I will often begin my shopping research, having LLMs do the research for me. I'll send them on a deep research task and have them do all of that. I'll have them synthesize it for me. In some cases, I might be patient enough to read that long report. Sometimes I'll say, convert this into a table and just give me like, you know, what I need to know.
02:38
Speaker 3
This is pretty hardcore shopping. Like what kind of item are you having deep research reports written for? Tell me it's a t-shirt.
02:49
Speaker 1
Yeah, well, you know, sometimes it's like, you know, I'm looking to buy a car right now and trying to figure out what new electric vehicles are around. But honestly, I don't even know the research about whether electric vehicles are as environmentally friendly as originally touted to be because there's all this criticism that, hey, there's all this rare earth mining and the environmental damage of that is very high. So I'm not even saying which car should I buy. I start with, tell me The pros and cons of electric vehicles. In what sense is it good for the environment and what sense isn't? And in what sense is it convenient to own one and in what sense it is not? And then from there, it would be, hey, I've seen on the streets people driving Rivian. What do you think about Rivian? Is that any good? At times, it may not be deep research, but nonetheless, you know, I have a startup as well that I started a year ago. We're trying to make decisions for the startup like, hey, which.
03:52
Speaker 1
Software should I use for cap table management? Which software should I use for payroll? Which vendors? Who's right for a company our size? We are very small. We're just about 10, 12 people. Who's right for California? Who's right if we grow to 100 people and so on?
04:10
Speaker 3
I do feel like this distinction between using AI, using a language model to do research for shopping and and actually handing off the purchase decision to the agent is really important right now. Like I basically trust the AI to give me information about, you know, product reviews and that sort of thing. And I'm curious, like when you see people who are using AI on the research side, not, you know, we'll get back to agentic in a second, but on the research side, like how are they using it differently than traditional search?
04:43
Speaker 1
Well, I think the biggest difference with traditional search is that the search engines are merely curating the information and then you are taking the trouble to synthesize it. And you get exposed to a lot of information about various brands, even if you decide not to eventually buy from them, or even if it's not in your shortlist, you still get exposed to a lot. Yeah. In contrast, when it comes to AI for shopping, I'm outsourcing a lot of the synthesizing. And then depending on the type of buyer, the effort level varies where some might say, synthesize it for me. And then there might be a lot of back and forth, but what about this? What about that? And in other instances, you might outsource the decision itself in terms of, bring it down to the short list. And so you're starting off with a short list of three brands and a comparison table, and that's where the human work begins. A lot of it is based for me in terms of the risk of making a wrong decision. And if the risk is high and I'm spending a lot of money, then I'll actually go back and forth with the LLMs. But if the risk is low, then I'm pretty much, for the most part, outsourcing everything except the final approval decision to the LLM.
06:09
Speaker 3
And that seems pretty common, right? It seems like, by and large, a lot of people are comfortable using AI for doing the research, but they're a lot more hesitant to let the agent take that last step. They're wary of letting AI actually buy the thing. And, you know, it reminds me a little bit of the early days of the web, basically the 90s, which I am old enough to remember. And, you know, you would go and companies had websites and you'd look at stuff. But the idea of putting your credit card into the web was like, no, there is no way I'm going to put my credit card into some random website. And of course, you know, PayPal is in fact an example of one of the early pioneers here. They were navigating these shifts in real time. But now, whatever, 25, 30 years later, we have solved that. Everybody buys stuff online all the time. And, you know, I would say at this moment, I am not ready to just have agentic AI go off and buy stuff for me without my input. I could imagine getting there. I just don't trust it yet. And so I'm curious, you know, we've solved the trust problem before. How do we solve it this time? Like what has to happen to get to the point where consumers trust agents to buy stuff on their behalf, on our behalf?
07:27
Speaker 1
Yeah, I think there's a few things that, need to happen, some of which is probably the tech infrastructure needed to support something like this. The second is a behavior change. People doubted in 1998, 1999, whether it would ever happen that we would put our credit cards and buy. And there were others who said, no, no, no, it'll happen next year. And there were folks who said, no, it'll happen in five years. And so that's the speculative piece, when it happens and for what it'll happen and so on. I remember back in the day, I had my own students who, you know, I teach at Wharton. So my students at Wharton who started the company Warby Parker, where you buy eyeglasses from, when they first came and actually their term paper in my course was a business plan for Warby Parker. And at the end of the semester, I chatted with them. And to me, it felt like at that point when I chatted with them, it was like, I can see some people buying their eyeglasses online. But I think when you're buying eyeglasses, you're wearing one, I'm wearing one. I feel like I want to try it on. I want to see how it feels. And that is extremely important. So I kind of my reaction was, it makes sense, but I still don't see how it's mainstream. But they changed it.
08:48
Speaker 3
Well, they sent you the glasses. The thing of sending you the tester pair, was that because of your feedback? Did you get equity?
08:56
Speaker 1
No, I didn't. I'm sure others brought it up to them. I won't take credit for that, but I will give them credit for kind of saying, we hear you on the problem, but we disagree that there is no solution to the problem. And so that's the part that I was saying is speculative in terms of, when it will happen and what needs to happen for that full behavior change to happen. I think someone will crack it. And I think it'll happen in phases just like e-commerce did. We first felt comfortable buying books and we said, hey, a book is a book. I don't need to worry about how it feels in my hand and so on. And then it became electronics. Hey, I bought books for $ 20. Maybe I can buy a $ 200 or $ 500 electronics online and it's going to be okay. And then it became clothing. And then it just kept expanding. And so I think it's going to be similar with agentic where there are some categories which will be easier than others. And there'll be some customer segments where it'll be easier than the others. And that's where things will start. And then how that percolates or spreads out and becomes more mainstream across all categories, that's the subjective piece. And that's what I was kind of saying. It's hard to say when that will happen.
10:13
Speaker 3
You mentioned books. I mean, famously, Amazon started out selling books.
10:16
Speaker 1
Yeah.
10:18
Speaker 3
Do you have a bet for what's going to be the books of agentic? What's going to be the first category where people are comfortable having an agent go make the purchase decision for them?
10:29
Speaker 1
Yeah, I actually feel like groceries are a really good candidate for me. You know, and this is something that I actually ask my students on a regular basis. I've been doing this for about two, three years where anytime I'm teaching my MBA, undergrad, or executive students, every session I have on agentic shopping, AI shopping, I ask them, would you use an AI agent to shop for groceries? Just your usual grocery list. And perhaps you could argue that if I'm using Instacart or if I'm using even the Whole Foods or the Amazon app to order my groceries where the list is there and I'm saying go buy it, at some level, I'm kind of saying, hey, here's my usual list. You buy it. You make the decisions on substitutions and do it on my behalf. So I've already outsourced that. That's not a big deal. Tomorrow, if someone were to say, here's an AI agent that can do end-to-end, just upload your grocery list once, and then you don't have to ever ask again, that's one I'd be willing to do. The Instacart, the human making that decision, for me, is not that different than an AI agent saying, hey, I'll substitute this particular brand of milk with that particular brand of milk and so on. I know what this customer wants. And I'll know in two orders that this is working for me or not. Am I willing to try two, three times? Yeah.
12:03
Speaker 3
And it's relatively low stakes, right? It's not like buying a car. If you do it and it kind of sucks, you're like, well, I didn't get the right milk this week. Not that big of a deal.
12:13
Speaker 1
Exactly. So I'll happily try that a few times. So I feel that's a good one. I also think the other one is where individuals are doing a lot of research and the research is a big part of it. Insurance is a good example. If I were buying car insurance... Mm-hmm. I would be happy to have Muse or somebody else go to all the different websites, fill in my information, get me a bunch of quotes, and then pick one. Maybe with car insurance, I might still want the final decision, like an approval gate. To me, that's still agentic if it's doing everything.
12:50
Speaker 3
I mean, that's a big purchase, and it's a lot of money. And if it gets it a little bit wrong, you're paying... you know, or buys the wrong policy. Like that one feels way scarier to me than milk.
13:03
Speaker 1
Yeah, but give me the chance to approve it. And that's all I need. You give me all the data. You've done all the work. In effect, from the brand's perspective, AI is the customer. It's doing all of the work. And I'm just kind of saying, okay, everything looks good. Let's proceed. That's the one right I want with insurance. But with groceries, I don't even need to approve it. I'd say just do it. If it breaks two out of three times or three out of three times, I'm not repeating it. But if it works all three times, I'll do it forever.
13:29
Speaker 3
So let's talk about what happens when something goes wrong. If the agent does something I don't want it to do, it makes a mistake. If it goes off the rails, we know that businesses are split over who should be responsible in this kind of case when there's a dispute between the AI platform and the customer. And it's clear that most businesses think that the merchant who fulfilled the order should not be the one who's financially responsible when things go wrong. And so the question is, who should be on the hook if the AI, say, makes an unauthorized purchase?
14:07
Speaker 1
That's a great question. I think anyone who's pushing agents to make them mainstream for shopping, any company that's doing that will have to assure customers, we are on the hook, we'll take care of it. So that's when it goes mainstream. I think that's when you get past the trust hurdle. So if Muse wants to do that for me, or Instinct, another app or startup that wants to do this for me, if they want me to go full agentic, they've got to take on the liability risk and say that if our agent does something that deviates from this intent that you had, then we will reimburse you.
14:48
Speaker 3
You mentioned Muse. I mean, that's the meta... agent that is new and buzzy and instinct, which is sort of a startup, right, that wants to do something similar. Invite only, I believe. I haven't been invited yet.
15:03
Speaker 1
I can invite you.
15:04
Speaker 3
Okay.
15:04
Speaker 1
I'll do that.
15:04
Speaker 3
I mean, I'm curious, what do you think it's going to take? You know, there is this leap that a lot of consumers haven't yet made. That leap is letting an AI agent buy something for them. And I'm curious, what is it going to take for that to happen? And And do you think agents like Muse are going to be part of the key to this transition?
15:26
Speaker 1
Yeah, I think agents like Muse are key to the transition. So what Muse and other agents can do is you can articulate your preference, just like you do in a prompt on an LLM. They can go to different websites, gather the information, reason through that, fill forms if need be, make a decision, present it to us. So all of that, those systems do. And like I said, I use it up to that point. They can also go and transact and that's where they need to get past the trust hurdle. And I think for that to happen, first, they need to insure me about things going wrong, as you were just alluding to earlier. So they have to tell me that there is some protection. Just like, by the way, a credit card company, when it's giving us a credit card, they're saying if there's some fraudulent transaction, You're not liable. You can appeal and you can say, this was not me. Something's gone wrong. And they will make sure that no harm is done. And they're providing that guarantee. That's what needs to happen in this world. So someone needs to be the guarantor or the trust builder here. So that's the piece that is needed beyond just the technology needs to work.
16:41
Speaker 3
And it seems like that will happen, right? It seems like the nature of the history of finance is... lowering friction is the way things go. There's less and less friction to buy things. A credit card is lower friction than cash. Shopping online is lower friction than going to a store. And it seems like ultimately, agentic AI will be lower friction still, and someone will figure out how to solve the trust problem, and people will use agents to buy things on their behalf as often as we buy stuff online now.
17:14
Speaker 1
Yeah. I mean, look, I'm not a betting man, but if you were to say, hey, What's your prediction here? Pick a side. For me, it would be, you know, there's no doubt the history of shopping shows that we're always willing to adopt technologies that add convenience while guaranteeing trust. Now, is it going to happen for every kind of shopping? Maybe not. Is it going to happen for every person on earth? Maybe not. But is it going to be mainstreamed? For large pockets of shopping, my bet is that it will happen.
17:52
Speaker 3
We'll be back in just a minute. So we've been talking essentially about the consumer point of view, about what it's going to take for people to let agents buy stuff on their behalf. I want to talk now about the merchant point of view, about the retailer point of view. This is a thing you have written about. You wrote an article about it recently in the Harvard Business Review. And, you know, it does seem like most business owners largely trust AI platforms to represent their brand to customers. But it is clear that there is, you know, to some extent, a trust issue here, that the AI won't accurately represent the brand. What are you seeing here? What are you hearing from the merchants you talk to?
18:48
Speaker 1
Yeah, I think merchants have to worry about this too. And there's many layers of the merchant concerns. One is, are the agents representing the brands correctly? And so I'll give you an example. I was talking to this seller of vintage clothes in London. And so this person owns this brand. They sell in stores. They also sell online. And they sell vintage women's clothing. He actually bought the brand off of the previous owner. And he was kind of saying, it's shocking that we are the number one seller of vintage clothes in the UK. And you go to ChatGPT, you go to Gemini, and you ask this question. I'm in London. Where can I buy vintage clothes with 70s look, 60s look, and things like that? or Vintage Inspired rather is their brand, and his brand would never show up. And he was like, what's going on? And so I actually went through this a little bit. We found what was happening was that there was a post on Reddit where someone had mentioned that this brand is shutting down. And the history there is that the previous owner had put it up for sale And probably the word got around that the previous owner was struggling to make it work. And then, of course, new buyers took over. But there was this post online. So ChatGPT had probably decided that it shut down or something like that and was not showing the brand. And so that's an example of the brand being misrepresented by an LLM.
20:31
Speaker 3
What do you do about that? Like, this is a problem, and presumably it's not in the LLM's interest, right? The maker of the model doesn't want it to be wrong. The merchant doesn't want it to be wrong. Is this a solvable problem?
20:45
Speaker 1
It's a tough one to solve because the maker of the LLM is not going to say, every time you're not happy with an answer, appeal to us, and then we'll have a human look through it and correct it. That doesn't work for them. And so while it is in their interest, there's just too much information on the web, too many vested interests. And then how do you also address people who might complain about subjective things and say that, how come I'm not listed as the top three vintage brands? And in fact, our survey shows that we're among the top three. And so they're not going to provide resources for that. So unfortunately, what it ends up becoming is This is what I tell a lot of brands, that this world is a lot like your customers are sometimes misinformed. Your customers also need to be persuaded. And just like you've spent decades thinking about, how do I ensure I inform my customers about my brand? How do I make sure I persuade my customers about your brand? You have to do the same for AI agents now. And you have to start thinking about the AI agents as another customer, And the same questions you had to have for human customers, now you need to answer those for AI agent customers. And it's a new problem for marketers and brands and so on. So, right.
22:08
Speaker 3
So, I mean, the kind of central theme of the merchant point of view section of your article is companies, merchants, marketers need to think about the agent as a customer. And sort of you dig into that and you lay out these three common misconceptions that you have heard, you know, in your work with retailers and marketers. And so I actually want to walk through those and have you kind of talk us through them. So I'm going to read them to you and then I'm going to ask you what they mean. So the first one, misconception number one, is there will always be a human at the other end. Tell me about that. Like, what are you hearing? Why is that a misconception?
22:49
Speaker 1
Yeah, so I think a common one that a lot of marketers and brands think is, okay, AI is probably another channel like search, where in search engines, we have to think through how we show up in Google's results pages, but there's always a human reading it and making the decision. And so if you think about it that way, then you would say, hey, maybe the human is doing the research through chat GPT, but then the human's making the final decision. But I would argue that the direction in which we are headed is where people will say there are certain decisions where, you know, I don't need the headache of going through all of this stuff. It's not a lot of joy for me to sit and comb through the information and make sense of it. And I'd rather outsource it.
23:38
Speaker 3
Outsource it to an AI agent.
23:41
Speaker 1
To an AI agent. Yeah. Right. So groceries, we already talked about as one example of that. Yeah. There could be many others. I gave the example in the article about like tickets to a music concert. So I know I'm interested in, let's say, Coldplay concert. I want that ticket or in my daughter's case, you know, Taylor Swift, and we want that ticket. So I just tell an AI agent, hey, listen, just keep searching on the web. If it's under a certain price, just go buy it. If it's anywhere in my vicinity, go buy it. Because if you're going to find it and then flag it for me, but I'm busy in a meeting and I don't respond fast enough. And then the tickets are gone. I'd just rather just go search. If you get one, then go just buy it. So those kinds of things, I kind of feel like, you know, they're going to train people more and more to use agents. So this idea that a human being is making the decision is not quite there in those instances. So You know, it's the agent that is getting informed, the agent that is getting persuaded to make choices for the most part, the AI agent.
24:49
Speaker 3
Did you see Coldplay?
24:51
Speaker 1
I saw both of those concerts, actually.
24:55
Speaker 3
Okay, back to your article, back to that list of misconceptions. We are at merchant misconception number two, and that misconception is we just need AI crawlers to read our content. And in fact, most businesses say that they have started structuring their catalogs for AI, although for the most part, they're not all the way there yet. They're not fully ready. But why is this a misconception? Why is it a misconception that merchants think we just need AI crawlers to read our content? What's going on here?
25:25
Speaker 1
Well, I would say it's necessary but not sufficient to succeed in a world where AI agents might be shopping on behalf of customers. It's necessary because, yes, you want to make sure AI crawlers can read your content and understand how you show up. But that's not enough because their answers to customer questions or the way they reason through all of the information is not just informed by reading through your page and making sense of it. They're also reading a lot of other stuff. For example, they're reading third-party reviews. Some may be biased. Some may be objective. They're reading through those, making sense of it. they're increasingly starting to process videos. So I increasingly see that Google AI Answers and Gemini Answers are shaped by YouTube videos increasingly. They're shaped by Reddit discussions. So there's a lot of other things that goes into them. So the fact that they read your content does not mean that they are going to make decisions that are favorable to the brand. So just like every marketers had to think through how do you persuade humans? They have to think about how do you persuade AI? But with humans, we understand a few things. We understand $ 9. 99 is not treated the same as $ 10. It doesn't make any rational sense, but it is true that people sort of react to these prices differently. We also know that social proof matters and saying that your friends have bought this matters or you know that scarcity matters for people. Only three items in stock can influence somebody to buy it.
27:05
Speaker 3
Buy now.
27:06
Speaker 1
Buy now, right?
27:07
Speaker 3
Exactly.
27:08
Speaker 1
So we know those things matter for humans. But if I were to ask any marketer, what are these biases that AI agents have? Or how do you persuade an AI agent? The answer would be a blank because none of us truly understand it. So that's why it's important to understand that misconception and really start thinking hard about how we correct that.
27:28
Speaker 3
Does anybody know anything about how to persuade an AI agent or a language model making recommendations?
27:36
Speaker 1
We're starting to understand some basics. And so, for example, I just mentioned that with Google, I have enough data to suggest that YouTube videos matter. And so in your marketing strategy, you should be thinking about influencers and you know, the relevant, who are the relevant influencers? Which influencers are shaping, say, Gemini or AI answers in my business vertical? And who should I go talk to? Or understanding, you know, sometimes even page layouts and how they matter. For example, and this could change, by the way, what I'm about to say could change in six months because new models can come in and they may reason through information differently. But I've already seen, for example, that presenting information in the form of tables somehow has a greater impact than having the information in paragraphs, especially when you're comparing things. So why, who knows? I mean, but for some reason, if that same comparison is presented as a table versus three paragraphs, it somehow seems to persuade AI differently. So there's a lot of small things and big things.
28:52
Speaker 3
So at least for now, language models love tables.
28:56
Speaker 1
At least for now, they love tables. Don't ask me why, but that's what we find.
29:02
Speaker 3
Okay, so we're going back to your list of common misconceptions that marketers, retailers, merchants have about AI. The last one, misconception three, is AI is just another channel. Why is that a misconception?
29:18
Speaker 1
Yeah. I mean, I think the most common way in which brands perceive the impact of AI is to think of it as another channel. Your customers can come to you directly. They can come to you via social. They can come to you via search. And now they can come to you via AI search or via AI agents. And that perspective just says that, okay, we've got a bunch of channels and we need to show up in this other channel. And what I worry about that is that you are mistaking something that I view as a system level shift in how buying is happening And you're mistaking that for another distribution channel that emerged. The analogy I would give is that when Netflix got started, all the movie studios said, okay, here's another channel through which I can distribute my content. I can get my movies out via the movie theater. I can get it out via pay-per-view on TV. I can also give it to various TV channels or I can give it to Netflix through streaming. And this is yet another channel And what they realized later was that, whoa, it's not just another channel because what's happening is it's changing my customer behavior and training them to work in entirely different ways. And it's changing who has the power and so on. And ultimately Netflix took over that industry because Netflix said, hey, maybe I don't need your content anymore. I can just make my own content and deliver it and so on. And so I think there are similar analogies here, which is to say that If you just think about it as another channel, you're missing two things. You're missing one, the fact that your customer behavior is changing in big ways. And as a result, you could lose market share because if all you're thinking is it's another channel and I'm on that channel, your competitor that is thinking smartly about how to inform and persuade this new type of customer AI agents will be ahead of you and will steal market share from you. The second thing is you're also missing how the power dynamics might change. If people are doing more shopping through, I'll pick an LLM like say ChatGPD, then does it shift the power balance and those LLMs become more powerful? By the way, Amazon has acknowledged that this could be a huge impact on them, which is why Amazon is reluctant to provide its full product feed to the ChatGPDs of the world. And instead, Amazon is building its own AI agentic shopping experience, which is called Rufus, and trying to persuade people, hey, have AI do your shopping for you, but do it within our Amazon environment. That's because Amazon recognizes that the power balance will shift if the agent doing the shopping is not their agent, but a third party, because then they're at the mercy of the third party deciding whether it'll finally buy it from Walmart or whoever else. And so they recognize that. And that's an example of saying AI is not another channel.
32:27
Speaker 3
Well, I mean, that last example is compelling, right? Amazon is one of the few retailers with the scale and the money and the technical expertise to be able to say, oh, no, we're not going to let these outside AI agents shop with us. We're going to build our own. Like, clearly, businesses are worried that If they don't adopt agentic commerce, then they risk losing customers or sales to competitors. So if you're a small company, what should you do?
32:56
Speaker 1
Yeah, I think you need a holistic strategy. So by the way, when we talked about e-commerce, it's not like stores don't exist today. They are there. And a good strategy, like we talked about Warby Parker, figures out how do we mix stores for humans who want that experience walking into a store and trying on glasses with an online strategy for humans who want the convenience of going to a website and with four clicks getting their eyeglasses. And so it may be a strategy where you say, how do we have a strategy where we provide an experience for humans who want to shop themselves? And also an experience for agents who have already taken over a big chunk of the shopping work on behalf of humans. And so how do you support both?
33:41
Speaker 3
Although now you're walking back to the It's Another Channel, right? Like that is the It's Another Channel story. But the Netflix analogy suggests that you might end up getting eaten in the end.
33:55
Speaker 1
Good pushback. But by the way, notice that I did not say, how did you, how do you support humans who want to shop on their own with humans who want to shop through AI. What I said was, how do you support humans who want to shop and AI agents who are shopping? And so I'm really saying it's another type of customer. It's another audience and you need a real strategy for that audience. And so for that, it might mean, I'll give you again an example. You might say that Either the way I respond to it, I have to make shopping so enjoyable and so luxurious and so high-end that it's worth it for the human to show up. And a very high-end luxury brand would have to ask, how do I make it so amazing that the person says, I'm not asking any AI agent to do anything. I want to show up at the car dealership and I want to sit in this car and experience it. And that is what shopping is going to be. So you either make that call or you've got to make the other call, which is to say our primary customers will be AI agents. And so what I'm going to do now is have a team that is solely focused on winning with AI agents. And that is to ask, how do we inform these new customers? How do we persuade these new customers? And unleash this new team to go chase after that business And make that work. I think you've got to first make that strategic choice. And once you've made that, your playbook is going to be extremely different for these two strategies. But that second strategy is, again, I'm still saying it's not another channel. It's here's your new customers. How do you win those customers? That now means you hire new researchers who are figuring out how are these AI agents reasoning? I've actually just created a new startup called Bodium Labs, and what we are doing is doing that on behalf of brands. We are basically building systems that brings in interpretability or explanations for how AI systems reason. So it's no longer this channel. It's almost just like you have a data science team that's thinking about how do you inform and persuade humans. You have an AI research team that's thinking about how do you inform and persuade AI.
36:17
Speaker 3
We're going to take an ad break now and we will be back in just a minute. Your HBR article lands with this idea that AI is a new kind of customer, right? That the, at this moment, we're in this hybrid moment where people are using AI, but people are still clearly the customer. Um, we will probably soon be in a place where the AI is actually making decisions, actually buying things. And so like, ultimately, what does that mean? What does that world look like? How do you think about it?
36:54
Speaker 1
I'll share one thought I have about how it plays out. By the way, in software settings and in B2B settings, we already see it. You asked me for early examples where this is already happening. Many people are building their online storefronts through tools like Cloud Code. And they're not hiring developers to go build it. They'll just say, hey, build a storefront for me. I want to sell XYZ and Cloud Code or Lovable or these other no-code tools that will have AI write the code, no humans writing the code, will build the storefront. When they're doing that, they're making decisions. They're making, for example, how are you going to accept payments? They, the agents. Yeah.
37:43
Speaker 3
So the agents are making decisions about things like payments, these fundamental business trust things.
37:49
Speaker 1
Yeah. Yeah. And if you now take that to the consumer world, I think increasingly we will have a shopping assistant. I don't know who wins that. Is that meta with Muse or is that going to be a next generation thing that Apple launches with Siri and Maybe in the future, or Google does something, but it might be on the phone itself. It might come built into the phone. But a system where you might say, hey, I want to buy this kitchen timer for when I'm cooking in the kitchen, or I want to buy a new phone, or I want to buy an alarm clock because I don't want to set it on my phone, whatever the reason. But you might just speak that into the AI. And you might either say, do all the work, make a decision, confirm and get my approval before buying, or you might say, proceed to buy it and you outsource that. And it's going to be that seamless for a customer. But in that world, we then have to think through what does that mean for businesses? You know, how do you stay relevant as a business when you're branding? What does brand even mean now in a world where AI agents are deciding that? What does persuasion mean when AI agents are making the choices? What does advertising look like in that kind of environment? How do you advertise to AI is unclear, but all of those will have to evolve. And I think there will be a whole AI economy coming up. When AI is buying and selling, you might have a whole new layer of the economy open up with services targeting as either helping brands sell to AI or helping AI make better decisions when buying on behalf of some people. And so I think that's an interesting new era that's potentially coming up in the next several years.
39:44
Speaker 3
It was great to talk with you. Thank you so much for your time.
39:47
Speaker 1
Thank you for having me. It was my pleasure.
39:56
Speaker 3
Karthik Hosanagar is a professor of operations and marketing at the Wharton School at the University of Pennsylvania. Today's show was produced by Gabriel Hunter Chang, engineered by Sarah Bruguier, and our showrunner and editor is Jake Harper. I'm Jacob Goldstein.
The Host
