2026.09.24 | GP | Internet | Agentic Commerce & AI Search

Associated:

BRBY:LN

URBN

MC:FP

RVLV

RL

BC:IM

WMT

TGT

AMZN

TPR

V

MA

WSM

Viewpoint:

C-Suite

Moderator: Raghav Nayar (RN)

| Sep 23, 2026 | 29 Min Read

Kareen Mallet (KM), Founder and CEO - Replika Software

SU MMARY

Overview

The discussion focused on how brands are leveraging AI search and platform commerce to enhance their market presence. The expert, a leader in AI-driven software solutions, shared insights on the transformation of consumer behavior through AI, the role of generative AI in shopping, and the emerging concept of Generative Engine Optimization (GEO). The conversation also touched upon the challenges and opportunities brands face in adapting to AI-driven commerce environments.

Key Insights

The expert emphasized the transformative impact of AI on consumer shopping behavior, highlighting that AI enables more personalized and context-rich recommendations compared to traditional search methods. The shift from SEO to GEO was discussed as a significant change, with GEO focusing on being chosen by AI for recommendations rather than just being found in search results. The expert noted that AI-driven commerce could lead to less traditional website traffic as more of the consumer journey occurs before reaching the brand's site. The importance of a strong data and commerce infrastructure was highlighted as essential for brands to effectively integrate with AI platforms.

Competitive Landscape

Company/Platform

Positive Aspects

Negative Aspects

Shopify

Tapestry

Preparing merchants for multiple AI platforms, not betting on one AI winner

None mentioned

Making products discoverable through AI platforms, potential for new customer acquisition

Early stage of rollout with some glitches

Walmart & Amazon

Building their own AI shopping environments

None mentioned

Sephora

Integrating product discovery into conversational AI

None mentioned

Quantitative Insights

Metric

Insight

Consumer Adoption

79% of surveyed luxury consumers have used generative AI for research

Average Order Value

Increased by 40% for AI-driven orders compared to non-AI orders

Engagement Growth

60% growth in weekly engagement for AI shopping agents

The discussion provided a comprehensive view of how AI is reshaping the retail landscape, with brands needing to adapt to new shopping behaviors and optimize their presence in AI-driven environments.

RN: Hi, good afternoon. My name is Raghav Nayar with the Consumer Insights Group at Guidepoint Global. This afternoon, we're discussing how brands are optimizing for AI search and platform commerce. We plan to address GEO, shopping in Gemini and AI models, if proprietary agents are lifting overall performance, and growth constraints for payment agents. As a reminder, if you have any questions for the advisor during the call, please email ask@guidepoint.com and we will try to get your question answered. I am very glad to welcome back Kareen Mallet, founder and CEO of Replika Software and previously a senior director of fashion at Neiman Marcus. Kareen, thank you for joining us this afternoon.

KM: No, of course. Thank you for having me.

RN: Of course. Can you please tell us about your work in AI, your background, and Replika?

KM: Yes, so before founding Replika Software, I spent 17 years at Neiman Marcus and Bergdorf Goodman as fashion director, and a big part of my job there was really about anticipating what the consumer would want next, like identifying the trend, understanding how consumer behavior was changing, predicting new trends, which one would matter, and then translating all this into merchandising and then the customer experience. So in some way that's very similar to what we are now asking AI to do, understanding an enormous amount of information, of data, recognizing patterns, anticipating consumer needs, and then making better recommendations. So today I'm the founder and CEO of Replika Software, and in one line Replika is the infrastructure that helps brands turn the people who influence their customers... like beauty advisors, estheticians, stylists, sales associates... into a measurable digital sales channel. So over the past 2 years, we really transformed the software around AI. On the brand side, we build an AI-powered dashboard where brands can really interact with all the data using natural language and ask to create reports. And for the user side, so the sellers, we're using AI to help them make better, more personalized recommendations. AI can understand the entire product catalog of the brand they work with, and then they can help them to make better recommendations to their customers. So that's really how I see the future. AI might know everything, but the human knows you. So AI brings the intelligence and skills and then the professionals, the seller, brings the relationship, the judgment, and the trust, and then at Replika we kind of blend those 2.

RN: Great. Just starting with off-site AI recommendations, how do you define that and how frequently is the consumer engaging with AI over your GPT or other off-site platforms for items that they intend to purchase?

KM: An off-site AI recommendation is simply a recommendation that happens outside of an environment the brand controls, so the website. If I go to Coach.com, for example, and I'm searching for a handbag, Coach already has me. But if I go to ChatGPT or Gemini and say I'm looking for a sophisticated handbag for work under $500 where I can put my laptop, I haven't selected a brand yet. The AI is now helping me determine my consideration set. The same thing

happened in beauty. Instead of going to a beauty retailer and searching for a serum, I can say I'm 40, my skin is dry and sensitive, can you recommend something? That's much more powerful than traditional search because I'm giving the AI the context, what I'm searching for, and then the AI is making the recommendation. When I think of off-site AI, the important distinction is really where the decision begins. The brand may still ultimately fulfill the transaction, but it's ChatGPT or Gemini or Google AI that influence which brand I consider in the first place. And to me, that's the real disruption, and you ask how often the consumer engages with those. With the AI Overview, ChatGPT or all the other ones, I think it's growing extremely quickly, but I would be careful about giving one percentage because I think we're still early and the behavior varies by category. What is much more important is that consumers really are learning a completely new shopping behavior. Instead of searching—really, for 20 years they went on to Google—they trained us to reduce what we wanted into keywords, like "moisturizer," "black bag." But with generative AI, I can explain what I actually want in the same way I would explain to my salesperson in front of me, and that's fundamentally a better input, and we're already seeing meaningful consumer adoption. I was reading a Boston Consulting Group luxury research last week, and they said that 79% of surveyed luxury consumers had already used GenAI to research.

RN: Hi, Kareen. Are you there?

KM: Oh yeah, hey, I couldn't hear you anymore. It felt like I was disconnected.

RN: Okay, so you were saying BCG Group report 79% luxury consumers and then it fell off.

KM: Yeah, so they're starting to use that AI and I think it will vary by category, but I already know, if I know what I want, like I want a mascara, I don't need AI, but if I'm exploring, I think this is where the consumer will go.

RN: And how are AI Overview recommendations affecting brands? We follow home improvement and we had a large home renovation CEO say that off-site AI recommendations are actually driving up their costs for new leads.

KM: Oh my God, the phone keeps dropping.

RN: I'm here. Can you hear me?

KM: Hello, I can't hear you. Can you hear me?

RN: Yes, I can hear you.

KM: Yeah. So I think the biggest change is that more of the consumer journey is going to happen before the consumer ever reaches the brand. Like with traditional Google search, Google sent me somewhere, I searched for a product, I clicked the link, I landed on the brand website and the brand had the opportunity to educate me and capture my data and convert me. But with AI Overview, Google can answer much more of my questions directly.

RN: Hi, Kareen, are you there?

KM: Right, Raghav? My guess?

RN: Yes.

KM: Sorry, and I think it's easier if you call me. I don't know what's happening. It keeps pausing. I think you were asking me about the renovation CEO.

RN: Okay, we were talking about, you had said more of the consumer journey happens before the consumer reaches the brand and that's where we left.

KM: That's where it was reaching the brand. And then after with the AI Overview, Google can answer much of the question. I can learn about the category. So brands may start seeing less traditional website traffic, but the conversation is happening before they come. So this is actually good because there's conversation, but the brands might not know everything that happened before.

RN: And then what are brands doing to control the content of AI Overview and GPT today and what do you think they could do better?

KM: I wouldn't use the word control because brands can't really control what ChatGPT or Gemini say about them. I think the better question is how a brand can influence whether AI understands them correctly and considers them authoritative enough to recommend them. And that's where GEO, Generative Engine Optimization, becomes very important because the first layer is all the product information and brands need to really get after that; they really need to structure and detail all the product data. AI needs to understand much more about the product than just the product name. They need to know what it is, who is this product for, what problem does it solve, what are the ingredients or material that these bags or this cream, this moisturizer is being made of? What does it cost? Is it available? Is it competing with different products? What do other customers say about it? Who's recommending it? So all of this information that before was inside the e-commerce now needs to become part of the marketing so when they do their product catalog, they really need to think how this information is going to surface into generative AI. But the AI still has to decide whom they want to believe because that's what the brands say. But the AI will also pull information from other sites like reviews, editorial coverage, experts, Reddit, communities, what they say about the product. And the more those communities, those professional recommendations, those experts speak about the product, then they will surface it. Then it will signal to the AI something beyond what the brand says already about itself. And that's where I think this becomes very interesting for companies like Replika, because we sit in that recommendation layer. So if thousands of estheticians or users, dermatologists recommend a product from a brand saying, "This is the best serum," when somebody is going to ask, "What is the best serum?" they will surface this brand.

RN: We will go to GEO versus SEO. How is GEO different from SEO? And do brands and retailers have the right labor to move forward with a GEO agenda?

KM: So I think the simplest way to explain SEO versus GEO is: SEO is about being found, and GEO is about being chosen. So, for example, for the last 20 years, the question for brands was, "How do I get into page 1 of Google?" You optimize your entire website, your keywords, your content, your links, and ultimately you were trying to rank as high as possible so the consumer would click through to your site. With GEO, Generative Engine Optimization, the question becomes different: How do I become 1 of the products or brands that an AI actually wants to recommend? And I think that's a much more consequential shift than it initially sounded because if I search Google for "best moisturizer for dry, sensitive skin," I may get 10 organic links, or sponsored results, or shopping results, and pages of options. But if I ask the same AI, "I am this age, I have dry skin, everything," the AI will give me 3 products. So that's a completely different competitive environment. The AI isn't simply retrieving information for me anymore. It's helping me actually make the decision. That's why I think about GEO as competing. It's like the shelf space, when you were in retail, you're competing for your space. Here you're competing in AI. Physical retail has shelf space, and social media has the feed. Well, AI is going to have to have recommendation slots. So if there's only 3 slots, how you make sure you're going to surface in them. So I think GEO could ultimately become extremely valuable commercially. I mean, it's not the end of SEO, but definitely.

RN: And how are brands and retailers positioning to move forward with GEO? You think they move forward with doing it inside their organization?

KM: Yeah, I think they're not really... it's not a set discipline set in stone already. Most large brands already have an

SEO team, e-commerce team, content team, data team, product team. So the problem is that GEO kind of crosses over all these teams. So there is not somebody just dedicated to that position inside. But they are using all these teams and trying to make sure that the product will surface, because if you ask why did ChatGPT recommend my competitors instead of mine, there may not be one answer. Maybe the competitor had better structured product information, which is very important. Maybe they had better reviews or more editorial coverage. So, who owns it? Is it the SEO team? Is it e-commerce? Is it PR? It's all of them. So I think it will be a combination of all of these. That's why I don't think most organizations are fully structured around this yet. But I think we're going to see companies beginning to create cross-functional teams around AI visibility and AI recommendation and implementation in the same way that SEO became something normal inside companies.

RN: Got it. And I guess which brands or retailers are leading the shift to GEO?

KM: I think some of the most interesting companies are actually the platforms building the infrastructure around GEO. Shopify is one that I'm watching very closely because the strategy is essentially, we don't know whether ChatGPT, Gemini, Google AI, Copilot, or any other one or something else will become the dominant shopping interface. So then Shopify's merchants, product structure, they make that accessible to all of them. So for the brand, for the vendors, they offer those platforms. I think that's extremely smart what they're doing because they are not betting on one AI winner. They're just preparing all their merchants to sell wherever consumers choose to shop. Then on the brand and retailer side, we're seeing different approaches. Tapestry is making Coach and Kate Spade discoverable and purchasable through Gemini and Google AI. That's what we talked about earlier. Walmart and Amazon are building their own AI shopping environment, with Sparky and Rufus, and retailers like Sephora bringing the product discovery and recommendation into conversational AI inside ChatGPT so you can already use it; you can discover the Sephora product and then you go purchase it on Sephora. So I don't think there is one GEO strategy yet. Some companies are making sure the product can be recommended by external AI platforms while others are building their own AI recommendation experience onto their website. I think the strongest companies will ultimately do both.

RN: Got it. And you said that SEO is still going to be important, but how important will be GEO?

KM: How important will be GEO? It will be very important. It is because AI is becoming so important, I think because customers have embraced it and now they are on these platforms asking questions, making their own research. It is very important that brands take it seriously and do everything to optimize to resurface in it. It will be more organic because here it's just not, they can put all the right keywords, but then they will not surface because there is no press or there is no content around it.

RN: Got it. We'll start with Tapestry, the announcement that you mentioned. So what are your thoughts on the announcement coming from Tapestry and also the timing in the fall of 26? And earlier we had talked about this not working for either of us.

KM: Yeah. I think it's a very important announcement because we are moving beyond AI simply recommending products. We are now moving towards the conversation itself becoming the stores. Before, it was just like I'm asking all those platforms about a product they recommended. But now, when you can buy it directly into the conversation, so into those platforms, it's different. Until now, Gemini or ChatGPT might have helped me discover a Coach handbag, but ultimately I would leave that environment of my discussion, go to Coach.com, and complete the transaction. So what Tapestry and Google are doing to change that... if I'm using Gemini or Google AI mode and discover an eligible Coach or Kate Spade product... is a buy button can appear directly in that experience. If it worked, I can check out with Google Pay without being redirected. So we're really collapsing the discovery to the recommendation to the transaction, and it's all done in the same environment. That's a significant change in commerce. When we say you tried it, I tried it, it wasn't working, but the announcement was, when was it, like on the 15th or the 16th? So it's live in the US.

RN: Thank you for joining us today.

KM: Maybe for some products, but the way it's supposed to work, if the product appears when I'm doing my search, I would select the product, use Google Pay, explicitly approve the transaction, and then complete the purchase without being redirected to checkout. So they say it's officially live, but clearly it's still in the very early stage of rollout. Google itself described the UCP checkout as being available to select merchants and eligible products. So maybe we didn't choose the right product. So I wouldn't expect every query or every product or every consumer to necessarily see that experience yet. But we are talking about something being deployed right now, not a mature commerce. They are the innovators, so there will be some glitches at the beginning.

RN: Got it. And what does the brand need to have to make Gemini or AI mode shopping possible?

KM: What do they need to?

RN: What do they need to have or what do they need to do, whether it's from data or interface with UCP?

KM: I think the first thing is they need to have a very strong commerce and data foundation. What we spoke about before: you need excellent product data, accurate pricing, real-time or very accurate inventory because that's where AI is going to pull from, shipping information, returns, and all the payment infrastructure, fulfillment. And then you need the technical integration that allows the AI environment to communicate with the merchant system. So, for example, for Google, that's where UCP, the Universal Commerce Protocol, comes in. I think of UCP almost as the common language between the AI and the merchant, so Gemini needs to be able to say: is this product available? What does it cost? Can it ship to these customers? Then it creates the RN: Thank you for joining us.

KM: The checkout, you complete the transaction, and the merchant systems need to be able to answer all of that. So this is not simply turning Gemini shopping on. There needs to be a real infrastructure underneath it. I think that's maybe another reason Tapestry was able to move quickly. In its announcement, Tapestry specifically pointed to investments they had made in technology infrastructure and data and consumer insights. I think that's really important for other brands who want to do this, who want to implement this, because agentic commerce is going to expose every weak commerce infrastructure you have very quickly. If your product data is messy or your inventory isn't accurate, your APIs aren't reliable, or your order management, you're going to have problems participating in this environment.

RN: And what are constraints to geographic expansion? The Tapestry corporate press has stated it's supposed to be live in the US with plans to expand across geographies in the future.

KM: Yeah, I actually don't think the protocol itself is necessarily the biggest constraint. I think it's more everything around the transaction, like every country has different payment preferences, different currency, I don't know, tax requirements, privacy... like in Europe, consumer protection law... all these, even return policy. So that can get really complicated. Google's current UCP documentation also made clear this is still like a rollout when I was looking at it, the merchant center integration is rolling out just to the US at the beginning. And the same other markets will follow, but each market is different; when you implement different e-commerce things, it's different whether you do it in the US or in Europe. So I would summarize the geographic issue this way: the protocol can't travel much faster than the commerce infrastructure around it. So that's why I think Tapestry's comment about being protocol-agnostic is important, but they're not really building for one Google feature. They're really building a commerce layer for a future in which shopping can happen through many different AI interfaces. So I'm sure there's a much bigger strategy story around it.

RN: And is there something unique about Tapestry's Coach and Kate Spade assortments that make Gemini and AI models appropriate paths to purchase? So the question is, is this a good option for certain types of businesses or can all brands get on this?

KM: I think all brands can own this. I mean, brands like Tapestry or brands like skincare, where I think when a fashion or an accessory or skincare, where it's not just one product, but where the AI can actually either build an outfit with your handbag... maybe you need the matching wallet, with your skirt maybe you need the blouse, with your serum maybe you need the moisturizer... and how to use it or how to sell it, that's where AI has become really super interesting.

RN: Great. We'll go to AI model target customer. Who is Tapestry really targeting? Is this a highly engaged customer or a first-time shopper?

KM: I think there's really 2 types of customers. One is the existing Coach or Kate Spade customer for Tapestry using Gemini as just another way to shop. They already know the brand and they're using this just as another way they would shop. But I think the most interesting customer for Tapestry is someone who hasn't decided that they would buy Coach or Kate Spade yet. Imagine, if I go to Gemini and I'm just looking for a sophisticated leather bag under $500 that fits my laptop. I haven't asked for Coach. I haven't necessarily even thought about Coach. But if Gemini recommends a Coach bag and I purchase it, Tapestry has potentially acquired a new customer at the moment of intent before I even selected a brand. So I think that's where the real opportunity is. So yes, your own customer will use it, but you will get new customers that didn't know about you or didn't think about you.

RN: Excellent. And if Gemini or AI mode is actually working, and I know we're beating up on Tapestry, but just any brand, how incremental could this be to total digital sales? Do you expect it to be incremental or is it taking from their existing business in the short term?

KM: I think it's far too early to know. I would be careful even about predicting a large number today. But I do think there's a real incremental opportunity. I think the reason is that Gemini can capture the consumer before she has even decided where to shop. That's very different than if I go to Coach.com, because if I arrive at Coach.com, Coach has already won a huge part of the customer acquisition battle, right? I'm already there. But if I start in Gemini with a need rather than a brand... I have a need, that's why I'm there, I'm asking what I want... Coach now has access to intent that otherwise would have gone to another retailer or another brand. That's the incremental opportunity. So I would look beyond simply saying, oh, Gemini generated 10 million in sales. I would want to know how much of that 10 million would Coach have gotten anyway. I think that's more of the important question.

RN: Excellent. This might be a premature question, and if we're not able to answer it right now, it doesn't matter. But are there shortfalls in the AI mode customer experience, and what would you like to see happen in AI mode and maybe ChatGPT, whatever it is that the customer is using off-site?

KM: How? What are the shortfalls?

RN: If there are shortfalls and if there are opportunities to improve the customer experience?

KM: Yeah, I think it's extremely early. You and I both tried to coach and Kate Spade and couldn't complete the transaction this morning. That alone tells you we are at the beginning of this. But even when it works, there are still limitations compared with a sophisticated e-commerce experience. Right now, for example, the experience focuses on a single item checkout, but that's not necessarily how people shop fashion. That's what I was saying. If I'm buying your handbag, maybe I want a matching wallet. Maybe I want to see the bag in 3 colors. Maybe I want to compare 2 sizes. Maybe I want to see how it looks on someone. So a great shopping experience isn't simply recommending

product by product, particularly in fashion, in luxury, as you know, discovery and consideration are part of the experience. So AI has to become much richer before it replaces everything a good brand website can do. So I think the other place is more for the conversation and the discovery, but the brand website is still there.

RN: And then we'll go to who owns the customer. So we have Gemini and AI mode. The Tapestry press release expressly said that the customer could bypass Coach.com or KateSpade.com. If that's the case, what happens to direct website traffic? And who should be controlling the data right now? Is it Google capturing the customer visitation and transaction data and sending it back to Coach or Kate Spade or what is happening there?

KM: I think that's 1 of the biggest strategy questions in AI commerce. Technically, Coach and Kate Spade, Tapestry still owns the transaction and the customer relationship, but Google increasingly owns the discovery experience. So I'm not sure what's going to happen. Some traffic will move away from the brand website if consumers can discover and purchase products directly through AI. But I don't think brands' websites disappear; their role changes. The website becomes more important for the deeper brand experience, storytelling, loyalty service, while AI will become another front door tool to those new customers. So that's the difference.

RN: Okay. We'll go to monetization. This is really a question around monetizing traffic and transactions that are off- site. What are the possible contractual scenarios, Tapestry versus Google or whoever, whatever has taken place so far?

KM: I think for Tapestry, the monetization is kind of straightforward. You sell more Coach and Kate Spade products and potentially acquire new customers. So for them, the real value is if Gemini brings Tapestry a customer who wasn't already planning to shop at Coach or Kate Spade, that's the value. If I asked Gemini to recommend or give me a bag, that creates that incremental revenue for new customers.

RN: Does Tapestry pay Google a fee per visit or a fee per transaction?

KM: I don't know. I don't believe the specific commercial terms between Tapestry and Google have been publicly announced. So I wouldn't assume what you're paying per visit or per transaction. When you think of the different scenarios over time, I think we can imagine managing several models between Google and a brand: that can be a transaction or referral fees or cost per acquisition, revenue share, even paid or sponsored product placement. So Google will definitely monetize this. Now we'll have to see what it is for the brand, so I don't think the brand, though, will own the data if the purchase is made in those platforms.

RN: Got it. And we'll go to agents. We have a number of agents that have come to market. What makes an effective proprietary shopping agent? And which features do you think really drive usage, conversion, and sales?

KM: Okay. So for me, we work with a lot of agents, actually. So for me, there's 3 things that are super important: context, personalization, and trust. A good shopping agent shouldn't just find you products. It should really understand what you as a customer are trying to accomplish, know your preferences, and really make the recommendation that is not, "they recommend this for you." That's really genuine, that feels relevant to you. So the features that are super important are personalization, memory... remembering you bought this product so you might not want this... product comparison, review summaries, real-time inventory, and especially basket building. Amazon is really a great example because they have an enormous assortment, obviously, and an enormous amount of customer data. So their AI shopping experience can help you navigate millions of products, summarize all the reviews, compare the products, and really answer really specific questions. So instead of searching through hundreds of coffee machines, for example, I can just explain exactly which espresso machine I want, and their AI is much better than any other traditional search and filters where you can never find it. So Amazon has another huge advantage: they already know you as a customer, but now that you have purchased these brands and you buy and you reorder, they get all this

data, and the agent is becoming really smarter at making better recommendations. So for Amazon, the agent can potentially move from simply helping me find a product to understanding my entire shopping mission.

RN: Got it. We'll go to engagement. So we have engagement metrics in Walmart on a relative basis. How are the shopping agents which meet your standards for context, personalization, and trust, how are they tracking versus your expectations?

KM: I think the early metrics are very encouraging and probably stronger than I would have expected. But for me, I think the most important metric isn't how many people try the agent, on that Sparky that they've been saying. It's whether they come back and use it again, because the first interaction is always curiosity. But if you use it, so that would be interesting to have the data on repeat usage, being the consumer actually found value the first time and they're really using it. But they do say that their average order value is up 40% compared to non-Sparky orders, which is really interesting. And we are seeing this with our agent, actually our own agent inside Replika. For our brand, the sellers using it, they have the average order value is plus 51%.

RN: Wow. How long?

KM: I think here they say 60% of growth in weekly engagement is impressive. I think this is great and people are excited to try new things, but I'd love to see what's next in retention. Are the same customers continuing to use Sparky month after month? And if they are, that's a much stronger indication that AI shopping is becoming really a habit rather than just, oh, I'm just trying it.

RN: You had talked about your own agents driving AOV growth at 51% with sellers versus orders that don't have the agents. I'm just wondering if you think this lift is because of newness, because of the loyal base of shoppers, or is it something that you think is a long-term driver for sales?

KM: I think it's newness, but it's actually cool because when you start interacting and it's really personalizing and it's really telling you, for example, for us, it's in skincare. And when they recommend products, they don't just recommend products, but they recommend how to use it. And if you buy 3 products, they tell you first to use this one and in which order... like serum and then moisturizer and then SPF... and it's giving the consumer, it's not just "here your product and go," it's "here your product and this is how you can use it." And so there's really that personalization that you were getting at the counter, and I think this is why consumers have been embracing it and have been sharing the detail why they want this product, on Sparky, on Amazon, on our own agents. So I think it's a little bit of all; there's probably some early adopters and novelty effect, but I think the mainstream customers really are going to embrace it.

RN: Excellent. We'll just end with the payments agents and KM: Okay.

RN: payment companies. So Visa Intelligent Commerce and its Trusted Agent Protocol are in development and MasterCard has agent pay. What are Visa and MasterCard building? How are these different from Google Pay?

KM: Yeah, so they're essentially building the payment and trust infrastructure for AI agents so those agents can transact on behalf of customers. Like if I tell my AI, "Find me the best flight to Paris under 1,000 and book it," the merchants need to know that it is really me, my authorized AI agent. I am giving the permission to spend my money. So Visa and MasterCard are building this infrastructure to make that possible and that it is secure. So it's a little different from Google Pay. I think about it very simply: Google Pay answers, "How do I pay?" And Visa and MasterCard, Agentic systems also need the answer, but it's not just the payment; it's identity, authorization, and then trust. So it goes a little bit further than just Google Pay.

RN: Good. And what are constraints right now to agent-to-agent payments commerce? We know that these are sort of in pilot mode or test mode at Visa and MasterCard.

KM: I think the biggest constraint for me is trust. If my AI is going to communicate with a retailer's AI and actually spend my money for me, we need to know, is this really my authorized agent? How much can it spend? Who is responsible if the agent buys the wrong thing? And what happens with fraud, returns, disputes? I mean, all this. The technology is moving very quickly, but I think the consumer will take longer. I think this part will be a little, you know, searching and researching is one thing, but allowing your agent to go spend your money is another.

RN: And do you think having Visa with their Intelligent Commerce agent, if this could accelerate agentic commerce sales or do we really need to see something more from the industry side?

KM: I don't think consumers are waiting for an Agentic payment. They are waiting to feel comfortable really to say, "I trust my AI enough to let it buy this for me." I mean, Visa and MasterCard's role is to make the security and authorization behind the experience almost invisible. So I think it's more about a big psychological hurdle, you know, saying, "Choose the hotel and spend $2,000 for me." I think we will get there, but for right now, it's just at the early stage.

RN: Got it. Kareen, thank you for joining us this afternoon. I'm sorry we had some difficulty at the start of the call, but this was an excellent call and I learned a lot.

KM: All right. Thank you very much.

RN: Thank you. We'll be in touch. Bye-bye.

KM: All right, bye-bye.

DISCLAIMER: The statements or opinions expressed today are those of the Advisor and not Guidepoint, who disclaims all liability for the content provided. The Advisor may not disclose material nonpublic or confidential information or any information that would cause the Advisor to breach any duty or obligations. Guidepoint is not a registered investment advisor and the information provided is not intended to constitute investment advice.