Name: Agentic Commerce Is Here: Are Your Payments Systems Ready?
Uploaded: 2025-10-28T16:16:07.785Z
Duration: 44 min 2 s
Description: Agentic Commerce Is Here: Are Your Payments Systems Ready?

Transcript for "Agentic Commerce Is Here: Are Your Payments Systems Ready?": Hello, everyone, and welcome to our webinar in partnership with Cognizant and Microsoft. My name is Neil Perry, and I'm the broadcast director for FinTech Magazine. Now today's session is called AgenTic Commerce is here. Are your payment systems ready? I love a simple direct title, and that's exactly what that is. We're here to talk about something that's really shaking up the world of digital payments, and that is agentic commerce. Now what do we mean by that? Well, think of autonomous AI agents that can actually act on behalf of consumers, finding deals, negotiating, even completing transactions, all without a human lifting a finger. Well, it is happening now, and it is going to change how we think about the world of payments. As these AI driven agents become part of everyday consumer journeys, the big question is how do banks, how do fintechs, and how do payment providers really adapt to that new world? How do you make sure those transactions are seamless and, most importantly, secure? Well, that's exactly what we're diving into in this session. We'll look at what's happening in this space, the partnerships involving infrastructure, fresh authentication models, and most importantly, what it means for you. So this is what, we're looking to explore today. Here's a few checking bullet points for you. First of all, understanding agentic commerce, what it is, why it matters, and how it's evolving. Secondly, we're gonna be exploring the ecosystem shifts, key initiatives, strategic alliances, and technological advancements, supporting agent driven transactions. Next, we'll be assessing the impact on payments, how AgenTic Commerce is changing transaction flows, risk models, and customer engagement. We'll be identifying some challenges and opportunities from compliance and trust to monetization and new business models. And finally, strategic readiness, what payment leaders should be doing today to try and future proof their infrastructure, partnerships, and strategy. So by any measure, that's an awful lot for us to do and not a lot of time for us to do it over the next forty to forty five minutes. But thankfully, joining me today, I've got some real industry experience to take us through all of that. First of all, I am joined by Ashish Bughnaga, business unit head, banking and financial services at Cognizant UK and Ireland. Also with us, Colin Banfield, agentic commerce lead, Amir at Stripe, and Esther Duran, Chief Experience Product and Design Officer, EMEA at Cognizant Moment. A wonderful panel. Great to see you all. How is everyone today? Excellent. Thank you. Very good. Thank you so much for the Hi, EMEA. Thank you for having us. Hi. Very, very welcome. That was quite an introduction. That was a lot for me to get through. So, let's give me a break for a second and just allow you to all introduce yourself to the audience. Ashish, you were first on my list. So just thirty seconds on on you and your role first, please. Sure. Thanks, Neil. I I lead Cognizant's banking and financial services business in UK and Ireland. And within banking and financial services, payments is a area of deep specialization for Cognizant. We work across the end to end participants, by the same, and we help our clients design, engineer, and operate, secure and modern payment systems. Thank you very much. Colin, after you. Hello. Partner at AI solution architecture here at Stripe, and I help drive the AI agentic commerce and stablecoin side of our products within the EMEA region. And, obviously, enterprises are super keen on the agentic commerce space in particular. That's why we're all talking today. So very nice to meet everyone. Thank you very much. And last but not least, Esther. Thanks, Neil. Hello, everyone. As you say, I'm I'm the chief experience design and product officer at Cognizant Moment, and I've been, overseeing design and products in EMEA as well as The UK and Ireland for, I think, over two decades now, helping, helping organizations just to upscale and go through all the changes in technology and all these changes in design and and making sure that they are ready for the future. Thank you so much for having me today. You're very welcome indeed, and thank you everyone for doing those introductions. So I'm sure we've all been to plenty of events where the host has to sit there and read out long rambling biographies for their guests. I like letting you guys do it yourselves. It's a lot cleaner. Right. Let's get on to the questions. We've got a lot to cover off, and we we do have a few, seed questions from the audience, that we've had sent in as well that we'll be doing right at the end also. Right. Let's get into a bit of a a definition if we can. And, Colin, if you could kick us off on this and and how you define agentic commerce because I'm sure lots of people see it from slightly different directions. What what distinguishes it from traditional digital commerce models? What sort of macro and technological forces are driving the shift towards it? Let's go with your perspective. Colin, I'll let you lead the way. Thank you, Neil. I think from a Stripe perspective, right, and my own personal perspective and what I see in the industry currently, a definition of agentic commerce is like the use of agents to initiate and complete any action in in a customer's buying journey. Right? So that includes finding products, tailoring recommendations, and completing transactions. I think fundamentally, it marks it's a move from simply providing prediction engine or information, the the knowing part of AI to enabling AI to initiate and and execute actions, right, the the doing piece. And I think in terms of trends that we're seeing that are moving this shift, if we look at platforms like chat GBT, for example, they have enormous distribution, right, because they have 200,000,000 US users weekly, right, which is quite a lot. And Agenca Commerce provides merchants with that new channel, to access that customer base, right, and drive a new distribution channel. And I think there's also benefits then for long tail merchants. Right? So your local pizzerias that are stuck in legacy commerce with phone orders, right, or niche product sellers. Right? So you're like likes of you're looking for ceramic ornaments for your garden, for example. Right? It's usually lost in page fifteen, sixteen of a product catalog. Right? But they can now enter the commerce game by just providing an API endpoint. Right? So they don't need to have a custom Shopify setup, for example, or check out optimizations. And I think the final point on that would be around the industry focus and the investment on that side. Two weeks ago, Stripe announced instant checkout within chat GBT and the launch merchant for that was Etsy. And so with that announcement on that day, Etsy stock jumped 16% in a day, right, just with with that announcement alone. So I think enterprises are here to embrace agent ecommerce for sure. 16%. That is a nice little example to drop in there, Colin. That's, that's definitely something to to flag. Esther, Ashish, whoever wants to pick that up next, your your thoughts on that definition. Yeah. I completely agree with Colin. I think, just, a little sort of, thing would be that you're not really in the merchant domain. Right? You're sort of acting and interacting with an intelligent piece of software, which is acting on the goals and objectives that have been set by you. So, there's that sort of dividing line. You're not really in the merchant domain directly. And, I think, ChargeGPT, was a cultural turning point, right, in this whole journey. I think the public actually witnessed the power of Gen AI, you know, to engage in a human like dialogue. And since then, the interest has been, huge, in this area. So I'm not surprised that the jump has been 16% or probably higher. That's it. Yeah. I mean, I couldn't agree more that, Collin. It's SpotOn. I think, more and more businesses are gonna follow, exactly the pattern. So, we we we just need to reach out for it. I was having a conversation yesterday with a colleague, and we realized that the search engine as we know it is totally dead. People won't be searching on Google anymore. They would be going through all the channels. So, yeah, I think we need to reinvent the way that we search and the way that we purchase and the way that, you know, we digest information of it also. It's very exciting times. It most certainly is. And I think you're absolutely right. Particularly what you just said there, Ashish, about that cultural shift. It's not quite, sort of our generations. Where were you when when Kennedy was shot, Bills? Where were you when someone first showed you sort of AgenTek AI in in reality? That was I think everyone remembers that moment, someone running into their office carrying a laptop saying, oh my goodness. Look at this. Let's move on to user experience if we can. And, Esther, if I could get you to lead on this and how Agentic Commerce was gonna be reshaping user experiences and the implications this has on future payments. And I know you guys have got the, the new minds, new market paper that came out very recently, which we we will make sure we drop into the documents section of the this webinar. Your thoughts on that? Yeah. Actually, it's a really good question, Nel. So Identity Commerce is where AI agents start making all those purchases for us as Colin was mentioning. So not just helping us to shop, but actually acting on our behalf. So they will compare the prices. They will negotiate the deals. They will check the times, even tracking your habits. And and probably we're ordering what you need before you even notice that you are even running low. So I think they're gonna they're gonna become you know, we would be a little bit lazy in that respect if if you like. But so the user experience, I think, is going to shift, and it's already shifting from that decision making that we have to do, to probably delegating everything, to the agent. So instead of us endlessly scrolling and scrolling and scrolling, going to the bottom of that website, as Colin was just mentioning, just to find those garden furniture, It would be more about trusting your AI to act within, obviously, the parameters that that you said, but to act on your behalf. So, for example, designing my head, it won't be about menus and buttons to click anymore. It would be more about how can we design confidence or transparency or even giving users the the meaningful control of the the, the digital agents. And I know it sounds very abstract, but that's exactly how the final experience is going to move forward. So on the payment side, I think it's more about the transitions will fade into the background. So everything will happen in the background. You won't be, you know, actively clicking anymore. So no more tap to pay or, you know, just validating and things like that. So it will become invisible, so trusted AI to AI settlements really. But the reputation and reliability, I think, is what is gonna matter the most. So more than the score the credit scores and all of that, it will be more about programmable money that could enforce the spending rules automatically. And, again, it sounds a little bit abstract, but it's exactly what it will be. So if you think about it, it is a bit like returning to the old corner shop when your mother was saying they'll go and buy some cheese or some, you know, bread. And the guy in the corner shop, they you know, he really knew you and your family because you were going there all the time. But think about, it's gonna be the difference in between the shopkeeper is now gonna be the dialing the corner shop anymore. It's gonna be an algorithm, and and it is trading on your behalf $24.07. I love that analogy. That is absolutely beautiful. And you were kind enough to share a few key takeaways from from the paper before we jumped on this webinar. And there's a couple of key things that jumped out to me there about, how the customer led movement can't be ignored. And I think there was one statistic about by 2030 AI friendly consumers being responsible for some was it 55 of consumer purchasing activity, which is quite extraordinary, but some 75%, going back to your point about control. Say they would be unlikely to allow AI to automatically reorder and pay high value items without their direct authorization. It's a fascinating area of research. So I I do urge people to take a look at that paper, when they do get a moment. As I said, that is in the document section just on the right hand side of your screen. Ashish, I'll come to you next on on on this. Your thoughts on, that user experience question. So I think, you know, again, I agree with, Esther's points, but I think AI will have to earn the trust, of of the user here. And, this whole user experience, I think, will evolve in phases. Right? So I think to begin with, you know, much of, AI, sits, agent ecommerce sits in the recommend space, right, where basically, AI agents are acting as smart advisers. You know, they're making purchase suggestions for you. They're making tailored recommendations. But they're not, fully autonomous. I mean, they're still, human in the loop, who's potentially, you know, sort of, going ahead with the rest of the process in terms of, checkout and the payment. Now I think, over a sort of evolution, journey, right, where AI proves itself, you know, to do well in low risk scenarios is is when, you know, it will start gaining more autonomy, you know, from a recommendation, to a transact, you know, to an orchestrating, complex task in a straight through manner, you know, end to end. I think that is a journey, and I think the user experience will evolve, in those phases. And, I think payments will basically, payments should just happen, right, so in a frictionless manner. So they will follow and they will enable, you know, this commerce. I keep calling. Yeah. I I totally agree with what she's said there. Right? In terms of, a user experience perspective, I think it's about creating that frictionless experience when we talk about the instant checkout and chat GPT. And the way I like to look at it is, you know, the typical checkout page is dead as we know it and intent is now the digital storefront. Right? So that's kind of how how we're approaching it at Stripe, and I think that does bring its challenges as well. You know, standard commerce, the seller controls the checkout. Right? So if we're going to have the agent come in and get involved in that, then I think it necessitates new protocols, new payment primitives as well, and that should be focused on standardization, data, and risk, altogether. And I think the agentic commerce protocol that Stripe and OpenAI have come out with is very much a lightweight way of getting that standardized set of APIs to bridge the gap between what's happening with the user and the agent without the seller having to change, their back end systems or a payment processor. Right? They don't have to change anything. But I think as well then, off the back of that, there's new risk and fraud mitigation that's needed. Right? So very much you don't have the same signals, I think, like device fingerprinting, for example, and when the transaction is going through an agent. So I think you need to leverage, like, a new primitive, like a shared payment token, which Stripe has just released. And I think that will help because it incorporates built in fraud signals. Right? It uses Stripe radar, and it has a risk engine that ingest data from our device graph, which is added to the shared payment token in real time. So that allows merchants to kind of assess the transaction as if it's come directly from a consumer's device. Right? And I think data was the the third one, mentioned here, and I think as well it's very much something that we need to consider, right, the implications and what the future looks like around that. I think, you know, the customer is completing purchases without visiting a merchant's domain, which Ashish had mentioned. And I think this leads to a reactivation problem. There's an attribution problem. Right? They don't have that traditional visibility of session data, browsing behavior, retargeting pixels. And so they're kind of at the mercy of what the AI AI agent is willing to share. And so I think we need to think about what work needs to be done and what that would look like as well. Well, I think that takes us on really nicely, Colin. Thank you to where I want to go next. And those are the foundational changes that payment infrastructures need to undergo to support these autonomous agents initiating transactions at all kinds of scale here. Ashish, if I could start with you, what are those changes for you in terms of those five, foundational infrastructures that need to be in place? Yeah. So I think if we look at the end to end, value chain, right, and, even before we sort of talk about the payment infrastructures, if we touch upon the merchant, for for a minute. I think, as Esther mentioned, I mean, the, you know, the merchants, the agents don't really care about things like UIs, buttons, and, you know, fonts which users of like us sort of seem to like. So merchants need to be ready for for agents and need to be welcoming of agents, you know. So in terms of being able to publish their inventories, you know, publish their offers, they need to refactor, you know, their current state and, design almost like a fork in architecture, which is welcoming not just of human beings, but also of agents. Right? So I think, that's, an important aspect. You know, first of all, I think on the security and trust element, right, I think, you're sort of moving from trying to prove that, you know, you're dealing with a good human to actually saying you're actually dealing with the right agent, authorized agent, which belongs to a good human. But I think that's a fundamental shift. Right? Today, all the identity and access management protocols, the traditional ones aren't geared up, you know, to sort of deal with agents. I I I I think the work that Colin mentioned around, the work that OpenAI and Stripe has done is quite, you know, partaking on the ACP and the MCP protocols. So that's that's fantastic. But they will need to sort of evolve, where they can actually work with agent identities and, you know, have more dynamic, you know, methods, as opposed to the static methods that traditional, you know, identity and access management, do right now. And it will have to have multiple layers of, you know, fraud and security controls. And it will sort of evolve, you know, from things like the tokenization we have today. You know, it would be just such an awful thing to have nontokenized data in the hands of agents. So tokenization is at the foundation, but you sort of take it one step further. You're talking about authenticated tokens, which are, you know, bound, to devices and apps. And then cryptograms, right, where actually, matching the intent, you know, what the agent is trying to do versus what its intended purpose and scope is and not letting it lose. I think that's a critical, you know, design consideration on the security and trust aspect of it. And, you know, finally, I think when, things do go wrong, I think the entire value chain, needs to have good, you know, liability frameworks, very clear dispute resolution rules, you know, audit trails of what these agents have been up to so you can actually, resolve, you know, some of these. So I think these would be some of the key considerations in terms of, you know, what needs to to change and what foundations need to be in place, to enable agent ecommerce. Well, Ashish, I I will class that as a thorough answer, and I I pity everyone following that because you've covered so much there in such a short space of time. Esther, I'll come to you next on that. Anything you'd want to add? Well, as you say, Nel, I think, actually, you just nail it. I totally agree with with what you say. Yeah. I think it's, accountability is gonna be really, really important going forward, especially, as you just mentioned, just, how are we gonna trust the agents? How are we gonna score them? How are we gonna do, as I mentioned earlier, for the experience? It's not about credit check anymore. It's gonna be about agent check. So how how can we do that? Again, it's fascinating because we are entering a completely different world that that we need to explore, and I'm sure we're gonna get it wrong at the beginning. But with this type of things, when it comes to banking, it's your money. I think accountability, again, is gonna be really important. Thank you. Colin? I think we've, we've definitely covered this one fairly well, but I just wanna maybe bring up, like, a point around we talked about the decoupling the checkout experience from the merchants domain. Right? And I think the shared payment token, which I mentioned earlier when we're talking about risk mitigation and authentication, I think from shared payment token perspective, I think, you know, it's it's a way that we can have that secure communication. Right? It's solving one bit of the puzzle, right, in terms of security pass and payment credentials to businesses, without exposing any raw card details, right, or any sensitive details at all. So straightaway, you're PCI compliant, right, between the agent and the seller. It's essentially just a scoped grant. So they love the customer's payment method. Right? So you can have it can be limited by, you know, maximum amount, expired windows, and seller IDs. And I think that solves a part of the puzzle. Obviously, we need to build more on top of that. This all needs to be payment real agnostic, so we need to figure out, you know, how everything ties together very nicely. What I'd like to do next, if I if I can, is really focus on a on a particular part of this fraud question. And this is an area that absolutely fascinates me about how authentication and fraud prevention models are gonna evolve that can really distinguish between human and nonhuman actors, but also being able to do it in real time. From from my perspective, a step, you know, a step back, the level of technology and intelligence there must be to be able to be fluid between those two different types of interactions in real time is is quite astonishing. Colin, can I start with you on that? That feels like a big challenge for me. Yeah. It is a big challenge. It's definitely something that we've been put a lot of thought into and there's a lot of work done into. And I I have touched on that, I think, in the previous question around the Stripe Radar, solution and being able to append this, for fraud detection, to the shared payment token that I just mentioned. And it essentially is building a new risk layer, and it's got scoring and customer modeling, right, using all of Stripe's fast data. And it looks at the things from a agent specific perspective, you know, around fraudulent disputes, you know, is the credential using card testing, is it a stolen card, will the card issuer decline the transaction, is it being used by a good or a bad bot. So there's a lot of a lot of agent specific areas that we're looking at to include and expand the Stripe Radar product. And I think it's definitely a very strong starting point, but, obviously, more needs to be done as well. Absolutely. And Colin, I if I can say you are the king of the subtle, shameless plug. I'm I'm really enjoying the Yeah. I'm trying to hold back on those. Yeah. No. It's subtle. I like it. If I can come to you, next, Ashish on this, please. I think, it'll it'll be a case of, you know, fighting fire with fire when it comes to, you know, fraud management. It's, I think, the good AI versus the bad AI, if I can call it. But then, I mean, look, we don't know what fraud, vectors are gonna emerge here. Right? It's an evolving space. We'll have to sort of react to it as we look at it, but then, whatever we can do to reduce that surface area of fraud. Right? With the things that Colin mentioned and, you know, as those mentioned, things like, the tokenization and, you know, the fingerprinting, which is actually the signals aren't quite there as they would be with a human being. So I think it's about reducing that surface area of fraud as much as it's possible, is is what I would say. Thank you. Esther, your thoughts? Well, very good thoughts when it comes to fraud, Daniel. But, yeah, I think it's just fighting that, evolving, as Cassius just say. I think, agents will become pretty clever, and then they will create that fraud that we are talking about. It's always the good and the bad thing here. Right? But at the same time, we you know, people like Colin and and the great work that he's doing, I'm sure he he and his teammates is gonna be they're gonna be thinking about how to fight this in a in a greater scale because, obviously, consumers, we are kind of in the middle of it, but so is so are brands. Brands don't want to don't want to be notorious for fraud. So I'm sure they're gonna they're gonna be spending quite a lot of money on thought around that. Yes. I think that's a a slight, understatement when it comes to quite a lot of money. I think that's high on the list of priorities. Let's let's dig into the partnerships area of this now because what I'd love to know your thoughts on the the types of partnerships between banks, fintechs, AI platforms that are proving the most effective in enabling this agent driven commerce, reality, if you like. Esther, if I kick off with you, and there was one line that really leapt out of me from the paper we were talking about earlier, and that's how familiarity breeds trust. That that was absolutely screaming off the page to me. Can you talk me through your thoughts on that? Yeah. Absolutely. Yeah, Neil. I think in here, the magic, is going to happen when the banks and the vintage and the the AI platforms, obviously, pawned together the the right way. And and what what do I mean by the right way? I think it's it's it's got they're gonna have three different trades. So the first one is how are we going to build trust frameworks, and this is the first one talking about trust. So banks, will handle all the identity, the risk, and the compliance as they normally do, but the fintech probably will provide the, tokenized rails that Ashish was mentioning earlier just to make sure that they check that there is no fraud. But the AI platform will probably bring the intelligence and the user experience with it. So, basically, everybody knows the name. Nobody's crossing anything. So then, it's liability when something goes wrong. And we know in life, things will go wrong, by nature. So I think that's the first trade. The second one is probably composable APIs. So what what do we mean by that? The vintage layer is kind of the bridge, and then it's wrapping the payments, applying the spend limits because, obviously, this can go you know, we can spend all the money in one go. We don't want that. So and, obviously, as well, feeding that insight back to the bank. So it's the data, exchange that Colin mentioned as well. It is kind of the plumbing that lets AI transact, safely without actually touching the core banking system because, obviously, that's where the fraud can happen. And the third one, I think, is the control autonomy. So agents acting within the the parameters that, obviously, we consumers put, but acting with the scope tokens, human in the loop for approval as, Ashish, I think, was mentioning earlier. I think it's early stage not to have a human to actually say yes or no. And the audit trails as well that I think he mentioned earlier. So it is autonomy, but with accountability, which is, I think, the most important part. Now are we actually doing this in a good way right now? I'm not too sure. I think it would be a 100% defined in 2026. So, you know, we are starting to do this. So, for example, Visa's intelligent commerce program certifying AI agents is doing it well. Then the Mastercard's agentic token is giving the AI's, permission spending power, is doing it well. And then Google's AP two protocol is standardizing the agent to payment communication. It's also doing well. But are we doing it well together? Probably no no to that extent that I was talking about. But I think the sweet spot right now will be the bank plus the AP, API vintage plus the AI platform, trio. So, basically, as we normally do is to pilot a small to prove they're safe. So, you know, it's not fraud. It's it's no, you know, all those things that we don't want to happen and then, obviously, to scale it. So I think the winners in in the years to come are gonna be those who are blending, I suppose, as, you know, normally it is the case, the innovation with control, but also to speed with trust. And I think that is kind of the key the key message for for me. What is the speed with trust? How can we do that in a way that is is controllable, basically? Because because what fascinates me there is is how the consumer is part of that partnership in in a more of sort of a wider broader perspective. Because one thing that jumped out at me from, from the paper, was how consumers were becoming more comfortable, you know, ship more sensitive and relatively risky activities, I suppose, you know, on their devices. For example, I think the stats I've got here, 52% are responsible, happy to use it for banking services, 44% booking travel. It's it's fascinating how they're part of that partnership. Although, personally, I have to admit the one thing I'm still never happy doing on my smart device is is booking a holiday. That has to be done on the big computer. That is a big computer job. I don't know about you guys, but that's just me. That's just me. But you see, this is totally different demographics. I can guarantee you that if you have kids, I have kids. So in the years to come, the the consumer behavior that they have versus the one that we have right now is totally is gonna be totally, totally different. So we need to keep that in mind as well. Absolutely. Absolutely. It's amazing how much, my children teach me about technology, which scares me a little bit because I'm meant to know a bit about it. Ashish, if I can come to you next, your your thoughts on that partnerships question? Yeah. I think there's a lot of interesting partnerships that are developing. You know, I love what Shopify is doing with its MCP, you know, catalog servers, and, that's like a merchant of merchant Shopify. So that's got a profound effect, on on the take up of agent ecommerce. Quite excited by what Nike has done, you know, particularly with the launch in, China around agent ecommerce with WeChat, partnering with WeChat. Of course, I must mention, you know, Stripe, Visa, OpenAI, that's a phenomenal, you know, partnership to enable agent ecommerce again. So there's there's number of, you know, things. I mean, Amazon, Alexa with Walmart, Sparky. So it's an interesting space. I think, look, alliances will have to ensure that a brand's products are discoverable and and purchasable through popular AI agents. Right? So, you know, you you would not get surfaced as a brand if you're not, partnering in the setup. Right? So I think that's that's an interesting area and lots of partnerships taking place here. Thank you. And, Colin, Nasheesh has, teed you up very nicely there, so I'll let you carry on. Yeah. Nasheesh has a very interesting point you raised around, merchant discoverability. Right? This is an area that I keep getting asked the question and what that looks like. Right? So right now, you have to sign up and go through, the KYC with OpenAI, for example, and to become a merchant that's discoverable within ChatGPT. Right? And that can take some time. So I think in terms of partnerships, so what we're talking about here, there's a huge opportunity for, companies to figure out how to do that much better. Right? Feedonomics is a platform that does it relatively well. I think there's probably a few others out there, but it's definitely an area that that can improve and and and become a lot faster to make this a little bit more frictionless. Right? And the other areas, Visa and Mastercard that Esther touched on in terms of, you know, solving for trust and consumer protection. I think you've got Shopify, Etsy, OpenAI for your ecom doing that discoverability and and for products discoverability and the selection piece. And then the final hurdle is, like, the Stripe OpenAI piece, right, where you're actually buying things. So I think all of these need to play nicely together. Some of them are right now, but there's definitely work to evolve. I love the way you phrased that. Yes. They do need to play nicely together. One last key question for me before I get on some of the, pre submitted questions from the audience, which I'm I am keen to cover off if I can. Where I wanna finish with is is the risks of delay. Because if organizations aren't adapting to iGentic Commerce now and adopting now, what are the biggest risks they're facing and what say one big action they should be taking today to stay ahead. Ashish, I'll I'll kick off with you if I can. Sure. So so I think look. I mean, agentic commerce in my view would not completely replace everything else. Right? I mean, the traditional, methods of commerce will continue to exist. I mean, ecommerce, mcommerce, social commerce, look at these will all continue to exist, including physical commerce. Right? I mean, that's not gone away. So, I think in terms of biggest risk, I think, look, agent ecommerce will take away a certain x percentage of of the market, but it will continue to coexist with other means of commerce. And, I think the biggest risk here is, you know, organized tent, would potentially risk losing that x percent. But it's not as if, you know, agent ecommerce is completely gonna replace everything else. I think it will coexist, in a certain form. So I think the one action merchants, sort of go back to merchants here, that they can take is to start working on that, fork in architecture, as I mentioned, where they are welcoming of agents. They embrace them, you know, and work alongside humans to to enable that, you know, flow of revenue from from the stream is what I would say is important. I think so. And I I wish I could remember who said it to me on stage at one of our fintech events recently. And that was if your, your organization or your shop front wasn't visible through Google. People would think you'd absolutely lost the plus. You know, we're we're getting into that same territory again. It's it's almost amusing to think that, yeah, what would they say the best place to bury a body is the is the second page of Google, isn't it? So, yeah, that was an interesting, perspective on that. If I can come to you next, Colin, on that, those risks and the action people should be taking to stay ahead. I think in for an for an organization, they they should be exploring what this new set of protocols is first and then how it integrates with their existing stack. Right? And I think the key question they should be asking their payment processor, for example, is what's their agentic commerce strategy. Right? Because these protocols are coming out. They're designed to be lightweight. They're designed to allow merchants to keep their existing technology stack, but they need to also have assurance from their existing technology stack that, you know, that they are here to support this new open standard. So I think that's the question that needs to be asked now. Thank you. Esther. I think from my part now, it's more about experience. Technology, obviously, is the conductor, as usual, but I think it's more about how can we implement the multichannels entrance when it comes to the journey. So as Ashish was mentioning, I don't think tomorrow everything is gonna collapse, and we're gonna have only one way to purchase things. I think it's more about the transition culturally that we need from generation to generation to change those habits. But I think at the moment, it's more about how can we create a multitenal experience. So whatever and however you start your journey when it comes to purchase, your agent will know where did you start and, obviously, what is the endpoint that that you want to go to. So, yeah, I think, a a lot of things to explore in there. Thank you very much. Right. We've got a few questions that I've got here on a list that were presubmitted by our audience. And I I do warn you all, some of them are pretty blunt and pretty to the point. So if we can get some quick answers on some of these, I'd be I'd be grateful so we can cover off as much as possible. Now I don't know who sent this one in, but this is brutal. Will agentic commerce completely replace traditional commerce? Now that's a that's a mic drop question if there ever was one. Colin, I'll start with you. Good luck. Your thoughts? Yeah. It's a good question. I I think I think the answer is no. I believe it's a new channel of distribution. Right? It's not a replacement. Right? I think, you know, 200,000,000 US users in chat GBT Weekly, I mentioned at the start. Right? And I think we're already seeing stats around, you know, agenda commerce and the conversion clicks into sales is two to six times higher than in Google, is what we're already seeing. So I think it needs to be exposure to this new channel, but absolutely not a replacement. Thank you. Prashish, anything you'd add? No. I think I've I've spoken about it. I think the x percent I mentioned, but the risk is the x percent will continue to grow. Right? And, I I think, a lot of people are turning towards, AI to at least, get some good recommendations, you know, smart, tailored recommendations. So I think that is something that we would keep growing. So, you know, to your point, Neil, where, you know, earlier it was considered you need to be on Google and you need to optimize for search engines. Now you need to optimize for, generative agentic, you know, engines and be discoverable there. So I think it's it's gonna be very important. Yeah. Just when everyone thought they had their SEO game on point, here comes the curveball. Let's do it all over again. Esther, your your thoughts and from you about that? Again, it's just multichannel. I agree with Colin. I don't think it's gonna happen anytime soon. Okay. Just in case of being ready. Right. Let me pick another question from this list. And whoever sent this one, I wish they wrote some of my questions. This is direct to the point. How do you see fraud levels in agentic commerce? Will it be lower slash higher than traditional commerce? Now that we could fill an entire webinar with that particular question. Ashish, I'll start with you on this one. I I I don't think I know the answer to this. Maybe you should ask me in in a couple of years' time, but it's it's a it's a thing that I think about a lot now. I think the one good safeguard which, we should use is I think agents come with an identity. Right? I mean, similar, to human beings. Right? There is there is an agentic identity. And within that identity, is stored the scope and purpose of, what that agent is meant to do. So I think matching, transactions against, you know, that stated scope and purpose, to see whether these agents are actually are acting outside the boundaries that are meant to could be a particularly important, you know, mitigation mechanism, I I feel, apart from everything else that, you know, we will do on the security interest element. But, it remains to receive nothing. We know it's too early days. We don't know. I mean, I I don't know. Yes. Well, if I could borrow the, Cognizant, time machine, I'd be grateful just so we could clarify that, that answer. If you can lend that, I'd be very grateful. Esther, your thoughts on that, the fraud question, just briefly? Well, again, no. I think fraud is something that's been happening since the beginning of time. It's just that it changes and morphs depending on the technology and and the the customer behavior. So is it gonna be fraud? Absolutely. I can guarantee you it will. But, is people gonna spend money trying to combat you know, to fight the fraud and and make sure that everything works well? Absolutely as well. Thank you. Colin, your thoughts briefly? Yeah. I think, you know, the actual answer is nobody knows, but I think we need to have AI driven fraud mitigation layers. Right? And I think that will go a long way to to helping solve the problem. Lovely stuff. And one final question. I'm just gonna squeeze in on the end. So if we can keep these answers pretty brief, I'd be grateful or else I'll start getting, grumpy messages from my producer, which you never want because when we finish the, finish the session, then they come chasing after me. Final question, and that are the signals and metrics organizations should be tracking, to measure the success of agentic commerce initiatives beyond just transaction volume. So what are those signals and metrics organizations should be tracking? Estelle, I'll start with you, please. Well, that's one that I really don't have the answer, but I think KPIs, as we know them today, they would have to be redefined. I don't think the KPIs that we apply to, you know, how many people has clicked on this website, how many minutes they could be staying here. I think, yeah, they they would have to be redefined. So, again, the answer nearly is we don't know, but we would have to be working towards that as well just to measure that something works or something doesn't. Yep. To be confirmed. Colin, your thoughts? Yeah. Absolutely agree with Esther. But I think, you know, there's maybe an evolution of of some of the existing metrics like conversion rate, customer acquisition costs. What does that look like in the agentic world? Average order value, is it bigger on the agentic channel? You know, there's there's definitely trade offs, but, yeah, they'll just be new there introduced, I think. Yeah. Thank you. And, Ashish, I'll give you the final word. No. I agree. I think art abandonment or the lack of it, you know, acceptance rates and, you know, displacement against other traditional commerce, I think that is an important metric, I think, to track you. I Think so. Well, we have fitted an extraordinary amount into the last, sort of forty five minutes or so. I'm afraid we are gonna have to wrap up there. Just a reminder, the New Minds new market paper, that's over in the docs section of the platform here. That's just over to the right hand side of screen. A little bit of bedtime reading for you if you so wish. As I said, I'm afraid that's all the time we have for now. My thanks to all of you, for joining me. We'll speak soon, no doubt. The recording of this webinar will be available shortly, so you can come back, watch again if there's any bits you wanted to, review, or you can share it, of course, to anyone else you think may find it useful. So for now, thanks again for watching, and bye for now.