So hi everyone. We are very excited to have you here. Already while we are recording, there are hundreds of people that signed up. This is a very hot topic and I'm very excited about the guests we have today. It's my first time trying to host a podcast, so I hope I'll do a good job. If not, feel free everyone here to direct me. The idea here is that we're talking a lot about how AI is changing the world, and today it's about how AI is going to change commerce from shoppers to brands and what does it mean. We brought the builders from Shopify that actually have a front row seat to how Shopify thinks about it. We also have our VP of Strategy, who spends a lot of his time researching and thinking about AI and commerce. And myself, one of the two co-founders and CEO of Yotpo, also spending a lot of my time specifically on discoverability with AI, but AI and commerce also on a broader spectrum. So I'll let our guests introduce themselves better than me.
Great, I can start. I'm Ellen Dunn. I'm a product director at Shopify. I've been at Shopify for almost a decade, and I have been working a lot in conversational commerce from years ago, which was focused on merchants actually talking to their buyers in real life and trying to sell through those conversations. And so I'm really excited about this era that we have entered and this future that has arrived with AI really bringing conversational commerce and fully automated conversational commerce to life. So excited to be here to talk about this and how we can ready everyone for this future.
Hi everyone, my name is Mucha Kambati. I'm sort of new to Shopify, but I've been in the industry for well over two decades and I've spent the last decade doing work on AI. I've seen all the failures and all of the successes lately with AI, the extreme interest with agentic capabilities that AI offers and how consumers are now rushing to do more things with AI, particularly with shopping and discovery. We'll get into some of that, but it's been quite fun diving into shopping with AI as sort of the focal point over the last several months here at Shopify. And I'm excited to talk to you all about some of the stuff we're doing and seeing as we work on this together.
So I'll follow it. I'm VP Strategy at Yotpo. Strategy at Yotpo basically has two sides. So I deal with product strategy, our own research and thinking about how AI is going to change commerce and what that means for us and where we should invest. The other side is actually strategic partnerships. So one of the areas I work on is our Shopify partnership. So this topic is close to my heart in this forum. So yeah, very excited to be here today.
Thank you. So to be honest, when we were preparing for this show, we did a prep talk and it was amazing. We had four different topics that we outlined that we think are interesting from our experience with brands. The first one is discovery in the age of AI agents. That's going to be the first topic we're going to cover. After that, in that kind of a world, what's even a PDP and if a product page has a new kind of a role. We're going to talk about authenticity, user-generated content, and trust in general in a world of AI, which is becoming more and more important. And of course, some practical strategies that brands can actually take today to drive their strategy in an agentic world. There's a lot of statistics around how AI is impacting discovery, product search, shopping. We did our own survey and we saw 48% of consumers are now searching on LLMs. I know Adobe had their own survey that the Shopify team shared that 39% of US consumers are already using generative AI in discovery. So definitely it's massive. Even if conservatively we say in three years, 10, 15, 20 percent of commerce is going to move to agentic, this is probably the biggest shift. It's like saying retail is now opening on the moon and there's a billion people on the moon. So this is something that everyone needs to be prepared for. It's something that we picked up a few quarters ago, probably like a year and a half ago, that we started to double into that. And I'll shamefully just mention my own newsletter about it, Commerce GPT. The entire purpose of the newsletter is exactly the hours that we spent with our own research, with our own data science team, to share what we see working in discoverability in LLMs. So Commerce GPT, check it out. And leave that aside. So now let's talk a little bit about discovery at the age of AI and what's called agentic commerce, right? So where shoppers are not going to go to the storefront, they're actually going to do the research and they're actually shopping inside the LLM. We saw Shopify announce some really exciting stuff from the MCP capabilities to the checkout inside the ChatGPT that was announced a while ago. So I'm just curious for the builders here that are actually helping everyone from an ecosystem standpoint moving forward, how do you see and how do you view discovery in the age of AI in general before we go to some tactics?
Yeah, maybe I'll kick off and Ellen you can follow in on some thoughts. What we're seeing, of course, is an early wave. We have from BFCM, the Black Friday, Cyber Monday that happened in the past, already some very high numbers where users were converting more when they went to stores that had agents on them. We were also seeing that a large number of purchases were being made, and this is the quotes that Tomer was mentioning at the beginning, Adobe numbers, a large number of purchases that were being made through some AI interactions. And we are seeing this, it's quite common in our day-to-day that we have incorporated AI into our day-to-day lives and so have consumers. So when it comes time for particularly considered purchases, there is a lot of actions that users are doing inside say ChatGPT or Perplexity to weigh options and get some research. So you're not as a consumer having to read all these different reports and websites and compare, and you get a little bit of intelligence back. We're seeing that start and more users are doing it for more actions. The thing though is if you're a merchant, you got to be asking yourself, the conversation may be happening outside of your storefront and it may be important for you to show up in that conversation in an accurate way. When users are asking what are the best shoes to go buy, is your brand showing up amongst what is considered by that AI as best? If the user is asking I want to buy a luxury watch, is your brand considered a luxury watch? Because that is an ambiguous term that somebody like an AI is disambiguating. But if you follow through on that discovery part, users will then make decisions based on a lot of different aspects like return policies, shipping to your destination, very specific information. And we know that AI needs to rely on accurate information to be able to answer those questions. So I want to just pause here to motivate the fact that this is happening. We're seeing this happen. We have evidence even from the last BFCM. We should probably see more of it at the coming BFCM. But I want Ellen to be able to jump in because there's a bit more towards making your brand participate accurately in this discovery journey.
Yeah. I think there's two things that are interesting. One is just to put a finer point on what Mucha just shared. I think a really easy way to think about it is in the old world someone's typing into Google best running shoes. In the new world they're saying I'm training for a marathon, I need good shoes for flat feet and running on pavement or something like that. So does your product information deliver on those types of details? Or if a buyer is asking a similar question around best brands for marathon style running shoes, do you have brand information that would indicate or structured information that would indicate that your shop specializes in that kind of thing? Right? So that discovery, that understanding of a brand is really easy to do when you land on a storefront because these merchants, these sellers have spent a lot of investment making their storefront bring their brand alive. But when these conversations are happening on other surfaces, how do you translate that brand story into a ChatGPT or Perplexity? So that's what I think is really interesting and those are some of the tools we're trying to build to help merchants show up in that way. Even when it's not just product information, which I think we'll get into a lot in this conversation, but if it's your brand and your brand story and who are you and who founded the brand and what do you really care about and are passionate about, that kind of structured data, we're also at Shopify helping merchants to put that into a place that is accessible for LLMs to pull out and be able to tell that story on your behalf.
Yeah, it's almost like the...
It's almost like the store needs to get richer or more structured.
Both, right? And more nuanced. Also, something you both are talking about is the difference between old search, right? You get a lot of results and you click in and you do your research. And here, so much more of the research is being done for you and is taking you so much further into this research journey. Being specific, answering the use cases, the needs, how you differentiate yourself is so critical. And it's one of the places where Shopify is so unique to be able to power so many different DTC brands with so many specific brands and specific products. It really is critical to add that kind of information and really explain to the agents what your brand and your product is all about to be able to be surfaced. Yes.
And I would say the other thing that we see, which is pretty cool, is that buyers who are coming out of these conversations where they've actually been able to get all this information, when they eventually do land on your storefront to make the purchase or to make their final decision between two products, they have way higher intent and they're way more likely to convert because they've already done all that research upfront inside of an AI agent. And so they land on your store and it's like boom, they get what they need and they make the purchase. That's the great news. Storefront traffic that comes from these agents, really high intent, really likely to convert.
Yeah. So I have a question. One of the things we did in our customer advisory board, we showed our top people from the merchant side a few capabilities of how we can take their reviews and reviews attributes into the LLMs. And one of the first things that came out is control, right? Because in the storefront they have 100% control and in LLMs they have zero control. So what do you recommend brands, like practically, to think about the control inside LLMs?
That's an excellent point. I think when we see storefronts from the pre-AI, let's say two, three years ago, they were being optimized for human eyes, right? Like on your mobile screen, on a large desktop screen, make sure that it's visual, the information is summarized because people are busy. But now you've got AI and it's got this voracious appetite to read and ingest and it can reason over large volumes of information. So you almost have to have the human eye version and the AI detail version so that the AI that is sort of summarizing, if it's Perplexity for instance, it's getting the right information and not missing things. You can put more for that AI. And so we've been thinking through how merchant storefronts can evolve. Sometimes you can put stuff on your storefront and AI will reason over it because it has the web and it can search, etc. But sometimes it's just like you don't want to clutter up your storefront. And so we've created an app that we announced in May at Editions called the Knowledge Base app that gives merchants the ability to view what we already know about your store. You install the app and there's auto-extracted FAQs that show up and you can say, 'Yep, that sounds good. I have nothing more to add.' Or you can look at that and say, 'I do have more to add. I think it's missing a brand story.' Or, 'I think I would like to say something about how I handle tariffs. Maybe that was sensitive and I don't want to be able to talk about that on my storefront.' So you can now use the Knowledge Base app to sort of create this extensive version if you want to that AI will have access to in order to answer questions about your store. Remember AI is just going to try its very best to answer users' questions and the more information you provide it, the more it can then be factual and accurate about your store.
I would like a funny, as a former designer and having been in this world for a while, everyone remembers like above the fold and below the fold when you're designing, right? So as a visual, you have to have all the important information above the fold. Like LLMs in this new world of agentic commerce gives us the opportunity to put so much below the fold that you would never have put back in the old days. But all of that additional content, all those rich details can now live in a way that LLMs can pull them out, summarize, search through them, find the pieces of that information that are relevant to the exact user who is doing the search. But you don't have to have a product page necessarily that's cluttered with all that information. Like your above the fold is still looking great for those who are actually landing and interacting with your storefront. So I like that analogy because it's such an old term that goes way back to designing.
So basically you're saying like PDPs and storefronts in general are no longer for just humans.
They're also for agents. And therefore when designing, Shopify is building capabilities that can help. Yotpo is building capabilities that can help. But more in general, less tactically, what does it mean that I need to now think about my storefront, some of it for humans, some of it for agents. So above the fold, below the fold is a great example. Any more thoughts on ideas that brands need to think about when they think about designing their storefront also for agents?
Yeah, I'll offer one more. Remember when users come to storefronts, we sort of define the thing that they go see. We give them maybe a popup that says you can get a 10% discount if you give us your email. We might give them some choice spots. Enter this door if you want men's stuff. Enter this door if you want women's stuff. And we sort of leave it to the user to navigate. And we try our best to make the visual appeal of our products show up very well. Users aren't often now operating in the same way as they were operating two, three years ago. And we know this from how you see users change their search habits. The big question is what's happening to Google? Is ChatGPT taking some of the queries away? Well, the honest truth is behaviors have changed. Users are wanting more conversational ways to look things up. So you've got to maybe embrace some aspects of your storefront to be more conversational. You can have your visual elements as you had before, but having the ability to have natural language questions where users can search or maybe ask detailed questions. I'm on a PDP and I now want to know what your...
Returning our product roadmap. We're going to announce a few stuff very soon. You're seeing it.
We'll talk more, but I do think storefronts get evolved in conversational ways too.
Yeah. And rethinking your search bar on your storefront even, what does that return if you ask a long-winded question? What is the user going to get results that make sense? Like let's hope so, right? Because that is the behavior that we're seeing.
And it's such a power, right? Because now you're able to actually get from people the questions that they want answered. It's not just clicks and monitoring where they're going and the funnel to go to conversion and what products they looked at. They're actually asking you and you're learning from that data. It's gold. And then to Tomer's point, our roadmap is how do you then take that and turn it into a loop of collecting the right data from the users that would end up getting surfaced and sorting and showing the right user the right data to those customers and the agents back on the PDP is such a different kind of mentality to how e-commerce was happening on the DTC website until now, right? The guesswork is kind of getting done.
A great flywheel, right, is what you're describing. And that could even change based on the season or the trends and what are people asking about? How are they asking it? You being able to as a seller see that in real time and then adjust the descriptions that are on your product page, etc. What do people really care about? They care about where it's manufactured or they care about its sustainability or what kind of materials. Those are all things you can dynamically be updating to make your products again have a better chance of being discovered for buyers where those are the things they've expressed to the agent that they really care about. And that's just a really cool thing to be able to do in such a quick iteration.
Yeah. I have to say one of the things that maybe it's my entrepreneurial spirit that gets me excited. It's a reset, right, in user behavior like you said. So now I can compete with Nike, I can compete with the big ones. When you think about entrepreneurship in general, so many small brands now, they need to figure out a way and they can have an early move advantage versus the big ones that in SEO, go beat Nike in SEO. Good luck with that. That's probably a war that no one is worth fighting. I don't know any brand by the way from our customers that is really trying to win versus the massive ones. And on the other hand, we have some massive brands that I see how they're enjoying the organic traffic coming from SEO. And now with generative engine optimization, this is why I think when I at least talk to our customers I try to instill urgency because if you do move quickly, and I think it's phenomenal that Shopify is helping its merchants, because then you can get a head start and you can optimize and then you can do it by season, you can get better. That doesn't really matter as long as you're in the game. And now many brands are not in the game, right?
