B (3:11)

Right. So this is quite a tricky one and in the short term I'm gonna hedge my bets a little bit and I'm gonna go with a 5.5 because.

A (3:19)

Unnecessarily, specifically, I'll give you my reasons.

B (3:21)

Why I think it's definitely gonna make a splash on the retail landscape. This is kind of now a gentic checkout. We've all been talking about it. It's here. It's gonna force brands and retailers to really think about AI as another commerce channel and it's gonna get competitors like Google to kind of hurry up and do their own thing. I'm less convinced it' an immediate impact on the channel mix and that is because changing shopping behavior just takes a really long time. The adoption curve is often quite different to the kind of tech curve. You just need to look at social commerce, voice commerce and I think social is a really interesting comparison because social is now a really big discovery channel. People using social for discovery for research for consideration but social commerce has been much slower to take off. So people actually making purchases on social media channels is not been as fast. And I think similarly we're seeing adoption of AI for sort of discovery research and consideration growing. There was an Omnisend survey in July which found over half of online shoppers in the US use Genai tools such as ChatGPT and Google Gemini for shopping related purposes at least once a month. However, there is kind of mixed messages coming out from the surveys. There was an August E Marketer and civic science survey which found just 2% of US consumers are starting their products product search using AI assistance or chatbots. So I think the kind of sweet spot for AI at the moment is actually the research and consideration phase, not necessarily even the discovery phase. People are know they want to buy something and they're using ChatGPT to research that purchase Instant checkout great in that it can, you can then purchase it without having to leave ChatGPT. But a that's a different shopping behavior and we know shopping behaviors take time to change and B that is going to require people to hand over their credit card details to to ChatGPT. We know that that's something that quite a lot of people are concerned about sort of data and Privacy, handing over personal information, a quite big concern still when people are dealing with AI tools. So I think there's. And also if we're talking about the channel mix in terms of E commerce versus, you know, in store retail, I certainly don't think there's going to be an immediate impact there because this isn't going to address the real concerns with E commerce, which are things like kind of delivery costs and speeds and the fact that people like shopping in store because they like to product, they like the social element of it.

A (5:48)

So really quickly, agentic checkout is that that's not just the thing buys the thing for you. Right? That's AI helping you along the shopping journey, like any kind of agent helping you to shop, not. Not necessarily the agent helping you to buy the thing by itself.

B (6:05)

Yeah, this is quite a tricky one. So technically agentic is something with a level of autonomy so it can complete a process on your behalf, whereas kind of normal, just gen AI requ continuous input and prompts. So up until now, ChatGPT has been able to help you with research and stuff because it can do prompts and things like that. But the kind of checkout part is more of an agentic capability.

A (6:33)

Okay, Zach, how much of a Splash out of 10 will instant checkout make?

C (6:36)

Yeah, I am in line with Karina's thinking. I 5.5. Well, no, I'm actually going to undercut it. I'll say a 4 4. But that's at the start. That's right now. Right now the functionality is not that great and the partnership with Etsy seems like a strange one to go out with. Etsy is a great platform if you need a needlepoint of your dog, but it's not a platform where you can very easily outsource like the everyday things that you want to buy, which I think is actually the utility of this sort of tool. It would be great if you could outsource the annoying everyday shopping tasks that you have to do. You need paper towels, you need toilet paper, you need Comet and you need milk and you want them as soon as possible, even if you have to pay a little bit extra for shipping. And to outsource that, have it do the work for you and figure out what's the best way of doing that. That's not what it's doing right now. Like you said, it's a single item that you can buy with this. That will change. I assume it will change rather quickly. The partnerships will expand beyond Etsy rather quickly. They have announced that they will broaden out to Shopify. So that is interesting. And once it does add that utility, I think it will be more useful. But for now, it just isn't that useful a tool beyond just the novelty of using this technology in a new sort of way. I, I mean, I also was thinking of social commerce and the role, the very slow rollout and growth of social commerce in relation to this rollout. And it took years for Facebook to figure out what social commerce should be, what it should look like, and what the utility of it is. And even today, our first social commerce forecast goes back to 2016. At that point, it was just 2.6% of overall retail e commerce sales in the U.S. it's now up to 6.6% this year. Now, on the one hand, you could say that's a small piece of e commerce sales, which are a small piece of overall retail sales. On the other hand, you could say that's also $85 billion. So, you know, I think this will have an impact, but it will take a little bit of time to figure out what it is and what makes sense for it to be.

A (9:42)

Yeah, it doesn't feel a. In terms of. Yeah, you said Etsy. So the things that you can buy from it aren't going to be mainstream. And also the behavior is going to take a while for it to become. Become mainstream. It's trying to. It's got this new Everyday Moments ad campaign which is kind of aiming to show real examples of how people use ChatGPT in their daily lives. They're cooking, they're doing homework, they're planning trips. And so this is supposed to be part of that, you know, shop, Shop with us as well. I said planning trips. They're also. We were talking before we hit record that they're also incorporating apps right? Into. They just announced that they're going to be incorporating apps into. Into. Into chatgpt. Tell us a bit about how that's going to work.

C (10:26)

Yeah, I think this is really interesting because this shifts ChatGPT into a platform where individual companies can increase the functionality within that platform. And the companies that they're going to market with are really interesting because two of them that caught my attention were Expedia and booking.com and if you talk to just about anybody who has used ChatGPT, one of the things that they say is that this is a great tool to plan a trip. That's how I have planned every trip since I began using ChatGPT because it is great. And planning trips and booking the individual elements within the trip is annoying. And if this can remove friction from that, that has a tremendous amount of utility. And I could see people using this because it does save time.

B (11:22)

I could see this integration being really good for grocery as well. I now plan all of my meals using ChatGPT. So I get it to give me recipe ideas or I'll give it a recipe and then I'll say, okay, make me a shopping list for all of this stuff and it automatically generates you the shopping list. You can add whatever you want to it. If there was then an option to say, okay, now order all of that for Click and collect from asda, I would probably do it like that's just another stage of the process like Zach said, that you would kind of automate and it just takes a bit of a pain point away.

A (11:54)

Yeah, yeah, they've got a lot of different partners. You can use it to Spotify playlists, to collate different music. There's. Yeah, Zillow is one of the apps that being incorporated. But there's, it seems like there's a big potential shopping audience here. I mean, if you just look at the people who use ChatGPT, we estimate it's 90 million Americans using it this year alone. And there's some data suggesting that those folks are probably quite lucky to buy stuff. Michelle Fradin, OpenAI's product lead for commerce and ChatGPT, says that more than 10% of folks who use ChatGPT have some intent or interest in making a purchase from the data that they've seen there. 10% is not a small number. If you look at the total number of folks they have using the platform. And then, Zach, you had some numbers as well. Nearly one in five consumers using an AI assistant or large language model for product searches in the past year, according to Salesforce. So it seems like people do want to buy stuff. What kinds of things, Karina, are people most likely to use this instant checkout feature to buy?

B (13:03)

Well, like Zach pointed out, at the moment this is limited to Etsy and perhaps Shopify merchants and it's single purchases, although they have said that they're going to expand that to kind of multi purchases. Should they manage to expand that to multi purchases and get some partnerships with grocery retailers? Like I could see that being a strong potential market. However, at the moment I'd say it's going to be like low cost, low consideration goods, kind of small fashion and apparel purchases, phone, accessories, gifts, novelty items, that kind of thing. In the immediate term, maybe kind of beauty, cosmetics, that kind of thing in the short term, I think I would.

A (13:42)

Imagine gifts jumped out to me actually because there was some data supporting that Adobe analytics data from March showing 1 in 3 gen AI shoppers use the technology to get present ideas. And so being able to complete the purchase there I think can be quite powerful for buying gifts. Zach, what do you think this can be used for the most?

C (14:01)

I think GIFS is interesting, but I think the real bread and butter is replenishment and routine items. The things that you buy on a regular basis that you know you need, you don't need to. Like Korina said, they're low consideration purchases. When it's something that is a high consideration purchase, that's where, I don't know, I think it's a, it's a bigger hurdle to clear for this to become the means of, of shopping.

B (14:31)

I think it goes back down to what do consumers actually want, what do they need? Like are we solving a need here? And we know that consumer journeys are actually super complex. Like people will spend a bit of time on social media, they'll go on a marketplace, they'll compare prices, they'll kind of bounce back and forth and then maybe go to a brand website to complete the purchase. This kind of instant checkout via AI is trying to collapse that purchase funnel, which, yes, is going to reduce some friction. But do consumers actually want to collapse that purchase funnel is the other kind of question. And like how comfortable are consumers going to be with collapsing that purchase funnel? Because I think that actually people like using different channels to get kind of different inputs from different channels and people enjoy shopping. Right. So there is always going to be a kind of limit on how much people are going to use it because they do sometimes want to do their own shopping.

C (15:24)

Yeah, I, I think that's right. I think like people like it and people are accustomed to it.

B (15:29)

Yeah.

C (15:30)

And those two pieces are the critical elements that make this difficult to, to come to fruition is habits are hard to break.

B (15:39)

Just look at, look at the, you know, in store versus E commerce split. Right. Like we saw a big acceleration during COVID but then actually especially in the uk we've seen it like slip backwards a little bit. There's a ceiling on it because people enjoy shopping in store and it takes a long time for people to change their sort of behaviors. So I don't think this is going to be any different really. Yeah, yeah.

A (16:04)

I do wonder, Zach, to push back a little bit. I do wonder if this is the place for those immediate kind of household good type of products because people, this does seem like they're Using this for research. I think that same Adobe survey had 55% of people saying that they can use Gen AI to conduct research on things specifically. That seems to me more like the type of products you might buy, which are gift ideas. You put in a bunch of information about a loved one, about a friend, about whoever, and then you kind of have a conversation with it back and forth to get an idea. As opposed to buy me paper towels. Do you. Could you see that behavior changing?

C (16:44)

Oh, I mean, I see that being a real use case here. I just think the bulk of shopping are those everyday purchases, you know, so, yeah, sure. But like if this is going to become a massive play, you need to capture those everyday purchases and it's probably easier to do that because people don't. That is not. That is not the fun part of shopping. The fun part of shopping is going to a store, getting inspired, saying, oh, this would be great for my niece. She loves this or that. Going to Target to buy toilet paper is not fun.

B (17:29)

And that's kind of why I think again, the grocery angle could work. You know, it fits into an existing workflow that ChatGPT people are already using ChatGPT for. You know me, a survey of one, I use it for recipe planning. I know other people use it for recipe planning. You can generate an ingredients list that saves you a load of time if this then enables a kind of quicker purchase of those products that would fit quite nicely into that work, that kind of flow. Similarly with travel, right? People are already using it to research travel. So again, like Zach said, that could take away a pain point there. I think with grocery though, the kind of food thing is that we know that the reason people still buy food in store is because again, they like to go and see the food, they like to touch it, they like to see the freshness of the vegetables and things like that. So there's a lot of kind of interesting barriers that it's going to be interesting to see how they play out, I think.

A (18:20)

Yeah, Fascinated me as well that folks are already pretty trusting of AI shopping assistants. According to YouGov data from this summer, only around a quarter of folks think AI shopping assistants would try to sell them things they don't need. I bet that share is a lot higher for people talking about human sellers trying to push them things that they don't need and upsell them. Or a really low share of people believing that it's going to give them inaccurate recommendations. As well as Accuad some data, 86% of users now trusting AI product recommendations up from 46% a few months ago. That's from Salesforce.

B (18:58)

I think though, the trust in the recommendations is very high, but the trust in kind of data privacy and security is more of an issue. So I think people can trust that the AI is giving them a good recommendation. But are they going to feel comfortable with sharing a load of kind of personal financial information with the tools? That's going to just take a little bit longer, I think, and it will come like the more that these people get used to handing over their credit card details to these kind of platforms. But I think that is in the short term going to be a bit of a barrier.

C (19:30)

I think that's true. And I also think like before this function functionality existed, it was easier to see this just as like a trusted source of information. But once you see this as a tool that is trying to sell you something or get you to purchase something, I wonder if that shifts a little bit.

B (19:51)

Yeah, you're right. That's a really interesting thought actually, isn't it? Because it's not just giving you recommendations, it's now saying you can buy it through us.

A (19:58)

Yeah, yeah. Speaking of sharing data, Emily Pfeiffer, a principal commerce analyst at Forrester, was saying the merchants get the sale, that's the benefit. But they lose the customer loyalty potentially. How much of a concern is this? How much will this potentially hold? Instant checkout back. Merchants being reluctant to give up that piece of the equation.

B (20:22)

I mean, I think they lose more than just the loyalty. They lose the possibility of kind of upselling and they lose. So they lose that kind of discovery possibility. But I think even more significantly, they could stand to lose retail media revenue. Right. Like if people are no longer visiting their websites because they're just doing everything and searching in and out AI and making the purchase in AI, then that could be quite a big impact for the sector. And we've already seen quite a big impact on search. Again, I'm not 100% convinced, convinced in the immediate term that this is going to have a kind of huge impact on the amount of people going to retail websites to purchase themselves. But I think it is something that brands and retailers are going to be quite concerned about.

A (21:08)

Yeah. Bell Lin of the Journal pointing out that the chatbot does also send customer payment information through the payments processor stripe, which forwards the data to the merchants. And OpenAI says that allows the merchants to maintain control over the customer relationships. But we'll see to what extent. Zach, did you have something to say?

C (21:27)

Yeah, I mean, I. I just think It's a similar calculus that that brands have when they go on marketplaces. And a lot of them, yeah, make many of them decide to do so. You know, they trade some brand visibility but they gain incremental sales and that's worth it if it exposes you to a whole big customer base that you aren't otherwise able to reach. I mean, this is why Amazon has a marketplace that, that is where like 60 something percent of sales on the platform stem from. It's, it's a powerful tool.

B (22:03)

But interestingly, if we go back to kind of social commerce, I think that's one of the things that's held back social commerce is actually brands and retailers being willing to sell on social commerce platforms directly because they're not guaranteeing that kind of same volume as a marketplace can. And they're kind of, there's concerns around things like fulfillment and yeah, who's owning the customer data, who's owning the relationship. So yeah, I think, I think you're right, Zach, in that it will be a something that they're going to weigh.

C (22:38)

Up really and it'll take some time for OpenAI to build that trust that brands feel like it's worthwhile for them to get on there.

A (22:52)

One of the questions that jumps out to me as soon as I read this news final question here is as long as it's quick enough to react, won't Google just copycat this with its AI mode Gemini feature, cutting ChatGPT off before it gains traction? And Karina, you reminded me that Google is working on its own offering. You sent me an ask call. It was may ask called TechCrunch's Ivan Mehta, writing that Google was planning to add multiple AI powered features for online shoppers, including new Visual panel in Google searches, AI Mode, personalized pricing tracking notifications with Agentic checkout and virtual try ons. Karina, won't Google's version just blow ChatGPT's instant checkout feature out of the water when it hits the market?

B (23:37)

I mean, not necessarily like I think ChatGPT obviously has a much smaller user base, but its users are very engaged. So I think there was July data from Sense tower which found ChatGPT mobile users open ChatGPT more than the top browsers or search engines now. So it is kind of having quite a loyal user base. Obviously Google has much wider reach. I think another big advantage is that it already has people's payment details through kind of Google Wallet. It's also already partnering with big payment providers such as MasterCard and American Express. And PayPal on a protocol for AI driven payments. And I think given its search dominance and its existing relationship with brands and retailers, it might just be able to scale a little bit quicker. ChatGPT instant checkout is partnered with Stripe. It's open sourcing the tech that powers Instant Checkout so more merchants and developers can build their own integrations. It's quite simple for merchants already processing payments with Stripe to use it. It's a little bit more complex but still possible for those with other payment processes to use it. For a bit of context there in 2024 I found some data saying that Stripe process process payments worth 1.4 trillion and that's versus MasterCard's 9.8 trillion. So if MasterCard is on board with Google, that is going to give it a little bit more scale, I think. But I think really it's going to come down to the experience and we don't know yet because we haven't seen this in, in AI mode, but it's going to come down to the experience that each gives the consumer.

A (25:12)

Yeah, Zach, agree. Disagree.

C (25:14)

Yeah, I agree. I mean I think Google has some massive built in advantages. I think it's massive trove of product data is one that is just hard for OpenAI to compete with. I also think Amazon has a lot of advantages. And is this going to steal customers from Amazon? I don't know. I don't think so.

B (25:35)

Amazon has, Amazon has also launched Buy for Me, which is where people can buy other buy products from other brands from within the Amazon platform.

C (25:44)

Yeah. And it also shares some of the functionality that this tool is offering just with like Rufus and all of that. And also it's Alexa devices where a lot of the utility that we were talking about that perhaps this instant checkout tool could offer in time. Amazon's offering already. I mean you can have an Alexa device in your kitchen and you can say hey, I want to bake that banana bread that I made last week, add all of the ingredients I need for that to my cart as well as ice cream and chocolate syrup and it'll just pop up and you can say okay, that's all looks good and check out. And that you know, is super easy and doesn't require you to, to shift your, your buying habits away from Amazon which you're very accustomed to buying with. And so we've said it before, like habits are hard to break. And so I, I think this will take some time to, to actually shake out.

B (26:53)

I think it's really interesting that you talk about voice commerce there because. And things like Alexa. Because actually that's one of the examples of the kind of tech curve being much faster than the adoption curve because voice commerce is still a very tiny percentage of E commerce sales. Like it has not caught on like people thought it would. But interestingly, I think Genai could be the thing that really moves the needle on voice commerce, because one of the things that was holding it back was just it was a kind of super clunky experience. And I think Genai, Amazon's already announced Alexa. It's already kind of enhancing the capabilities of its voice commerce. And I think that that is going to move the needle. We know a lot of people use these voice assistants, but the amount of people actually using them for shopping them is quite low at the moment. But I think kind of AI could perhaps accelerate adoption of that.

C (27:41)

Right. It doesn't make my life easier. That's the question that it needs to solve. Exactly. Always to show the consumer that it does. And that's not easy.

A (27:53)

No. We shall see. That's all we have time for, unfortunately. For today's conversation, thank you so much to my guests for hanging out with me today. Thank you to Karina.

B (28:00)

Thanks very much.

A (28:01)

And to Zach.

C (28:02)

Yeah, thanks for having me.

A (28:05)

To Frank.

C (28:06)

Thank you for Frank.

A (28:07)

Let's get that official, please. And to the whole editing crew and to everyone for listening in to reimagining retail email podcast made possible by Fetch Rewards. I hope to see you on Friday for the behind the Numbers show, where I'll be speaking about the major ways the digital world has changed this year.