Name: Recording for HUMAN Dialogue: Building Trust in the Era of Agentic Commerce
Uploaded: 2025-11-12T15:16:12.634Z
Duration: 44 min 10 s
Description: Recording for HUMAN Dialogue: Building Trust in the Era of Agentic Commerce

Transcript for "Recording for HUMAN Dialogue: Building Trust in the Era of Agentic Commerce": Great. Hi, everyone, and welcome. I'm John Searby, chief strategy officer here at Human. I'm joined today by, thrilled to be joined by, two two great guests, Sandy Carielli from Forrester, vice president and principal analyst covering primarily the application and product security market, and Tomer, our own Tomer Elias, who's principal product manager responsible for our agentic trust capabilities. You know, we're here today to talk about something that's shaking up not just cybersecurity, but, you know, all digital interactions as we know it. And that is the rise of agentic technology. I know another, webinar about agentic and agents and AI, but, you know, specifically here on this one, we're wanting to talk about, agent to commerce. So what is that how is that affecting the world and specifically that some of these focus areas for us. You know, AI, LMs, they all kind of taken the world by storm over the last couple years in in various flavors and fashion. They're disrupting so much of, our decisions around enterprise technology, our design principles, new product offerings, things we're doing every day. But what we're looking at here is, you know, how is that affecting the consumer first? What's changing at the the human layer, first and foremost? And then how is that disrupting really the fundamentals of the Internet, between advertising, ecommerce, search, all the things that we've taken for granted have been so fundamental in there for some time. You know, in 2022 when ChatGPT 3.5 was launched, there was this shift that happened. All of a sudden, this revelation by many of the the the human to machine layer just changed. I can now have a conversation with computer that's believable. It knows things. It it can answer questions. Right? And this is the fastest I think I've seen, an adoption of technology to go from, hey. Let's test this out to millions of users using it, new technologies out there, just a whole, adoption that's happened at rapid scale going from a chatbot to this now automated reasoner who can do things for us and answer questions and shop. You know, that has moved to this, you know, us from us going from the, you know, traditional world where we're used to operating with machines, you know, we all grew up with our laptop, our mouse keyboard, our kids have touch screens, we have voice age, you know, assistance over time, the mobile introduction has happened. Now we have robots in our pocket or our glasses that we can talk to that can answer questions and can do things for us. And that's fundamentally shifting the underlying economics, economics of the Internet in many ways from where we see it. The implicit agreement of, hey. I'll allow a search engine to scrape my site so that I can get a click and return where I can sell advertising space or gain attention of human eyes in some way, in order to sell them product. Right? The the the flow of money was very much built around human attention and that that has shifted fundamentally. I think that's where we see in, our customer base those, you know, supporting b to c across ad tech, those in ecommerce, the fintechs. Many folks who are trying to understand the transactional flow from that human interaction, and those those dinner digital interactions are seeing agentic in a in a specific line. Agentic commerce. How is it affecting us? How are we building technology and adapting to it in some way? You know, over the last decade, HUMANs been, protecting against automated threats. We are, focused on the bots that come in and automation was traditionally distorting things for away from that human interaction for many of our customers, whether that's from the the serving of an ad all the way to that transaction. And we've gone from not all automation being bad. It's not just is it a human or a bot, but is it a good bot? Is it a bad bot? Is it a good human with a bad bot? Is it a bad bot with a good human? All these variants now and complexity that's out there, in the world of agents has has fundamentally changed everything. So it's something that as an industry, we're all looking at. And Sandy, you know, very interested in bringing, you into the conversation here from your broader perspective, but I know it's it's now automation outpacing human interactions. We're seeing that, and we all need to be thinking about how are we not only from, you know, HUMANs perspective worried about how we serve our clients, but broadly as an industry, how are we adopting new concepts of identity, authenticity, and ultimately trust in some way to to really grow with that. So, enough of me, introducing or rambling on. Maybe we'll start, with a a few questions and tapping you all in. So, Sandy, I'd I'd love to maybe start with you just on, definitions. Let let's talk what do we mean when we say agentic? When we say I, and what's what how do you see, Forrester defining agentic systems? Yeah. So let me talk about some of Forrester's definitions, then I'm gonna throw some numbers at you. So at a high level, Forrester defines, agentic AI as systems of foundation models, rules, architectures, tools that enable programs to plan and adapt to resolve goals by taking action in their environment. And this is critical part with increasing levels of autonomy. That autonomy is a pretty important part of AgenTic AI. And then when you go to AgenTic Commerce, it means taking all of that stuff, taking those models, those rules, those architectures, those tools. Those are the agents in question, and the agents are conducting or they're enabling commerce at some stage of the customer life cycle. Now there's a data point I pulled from some of our survey research that honestly kind of blew my mind, and it goes back to what you were saying, John, about just how quickly this is all happening. 9292% of b to c marketing leaders say that they are working on a strategy for agentic shopping. 92% when I would wager that eighteen months ago, none of us knew what agentic shopping was. And now you've got 92% of b to c marketing leaders working on this. So that speaks to just how quickly. this is happening. Right? Yeah. That's crazy. Yeah. Yeah. I mean, Tomer, from your perspective as well, I mean, you're obviously seeing this as you're focused on our products supporting many customers. What are you seeing as far as kinda how do we define this? Where are we seeing this shift at our customer base? What what is agent of commerce to, to our customer base as you've seen it? So getting back to the great definition Sandy just provided, the autonomy and the ability to take actions differentiates between any bot or crawler that we've seen up until now. Right? Bots, and crawlers, even l m crawlers are used in order to scrape data to index information. And now AI agents are here to conduct something that is a little bit more hybrid. They represent the actual intent of of a human or consumer. They interact with the Internet or the application as if it's a human interaction, clicking on buttons, going through flows, etcetera. So that delineation is is rather it's quite important, understanding the intent of the agent, understand what they are here to do, what they can do. That's something that, lots of current solutions or tools are blind to. Your analytics tools, cannot differentiate between an agentic session or a human session. Your fraud prevention solutions were built around a house household or a human engagement, and now they need to, allow agentic commerce to happen. And you see some protocols, that are being developed to support that, but agentic browsers will not follow protocols. So there are challenges and, already that we see in terms of, the adoption of AI agents and the use of that by consumers. Alright. Yeah. I mean, it's wild. I mean, the 92% adoption. And I know, Tom, you're you're speaking with all of our customers. And to your point, the the speed at which it's changing, new protocols coming out, new technologies, as well as the speed that that we're all playing catch up to it. I mean, it's just. unlike anything I've seen before. You know, I think you take something like the mobile phone and that was disruptive. Right? We all bought our iPhone and it changed the world materially. No doubt no doubt about that. But it took us a while to buy those things and adopt it and use it everywhere. I think I saw what was it, ChatJPT was the one that was a little faster than TikTok and Instagram to a million users, right, just immediately. And so just high speed adoption here and then obviously the the changing technology around it is just, is just shocking every day. And I'm excited to see how much the industry is changing. Right? Like you said, Tom, the new partnerships forming, the protocols being launched, all of us trying to adopt it, which again is is is a huge part of this. But, yeah, I think, you know, diving into that just further. So as we look at that more broadly, how is this changing kind of that that customer journey? We've all talked about the customer journey, the the digital experience. How how are we seeing this kind of fundamentally change from a consumer perspective and then how is that translating into the the enterprises supporting those consumers from what you were seeing? You know, maybe Sandy across your research like how what are these impacts across a brand and their loyalty and how they interact with the the consumer in some way? Yeah. There are a couple of elements here. And the first one, I actually wanna share a prediction from our digital commerce predictions that Emily Pfeiffer shared, which is that one third of retail marketplace projects will be deserted as answer engine steal traffic. And so basically saying that the GenAI search experience is gonna curtail the growth of a lot of online marketplaces. And that speaks to the fact that consumers are finding different ways to get their information. Sometimes they will start with a question, chat g p t, and it may bring them to the site, but it may also bring them information that sends them to another site. So there's less engagement with a particular ecommerce provider or particular retailer. That changes things. The other thing that I think is really interesting in this and, you know, excites me as an analyst is going back to what you were saying about automation no longer being bad and no longer having good bot versus bad bot. Now we have this incredible set of blurred lines in terms of what's human, what's robot. Because now, frankly, if I'm asking a question in chat g p t, the fact that it's getting there through that, it's still tied to me as a human, and that still matters. So how do you understand what agentic traffic, is, who it's tied to, what the intent is, and how you should handle it in order to make sure that you are actually serving your human customers in the best way. But also to your point, making sure that malicious automation or automation that you don't want for some reason, maybe, you know, certain types of scraping isn't enabled. It becomes a much grayer area for folks to navigate, and that means that you're really having to dive more into the signals. You're going to have to dive more into understanding who's behind the traffic, and you're going to have to establish levels of trust with someone who could be going to coming to you through four or five different channels. Because I could be coming to you through chat g b t today. I could be coming through a browser tomorrow. I could be coming through directly. I could be coming through a particular agent. There are all sorts of different ways I may get to you depending on where I am, how I'm connecting. And you need a view that that's me. Yeah. That's difficult. Yeah. And mean, is that so much of the shift we're seeing? Right? Is it the I our concept of identity behind this, having to adapt accordingly? Right? The the end of there's a John Searby, we've all understood my login and my immutable kinda universal identifiers for some time, but how about what agents I use, how my paths to it? Again, my behaviors almost become the new form of identity in some way. And the problem actually is going to multiply. If you saw in July, for example, mostly consumer adopting Chejupt agent to try some some agent interactions. Later on in September, we saw a vast adoption of Comet browser even more than the Chejupt agent traffic that we saw before. And in October, October, we we saw it even tripled. So we see consumers actually starting to play and utilize different, channels in order to conduct, things online. As you mentioned, Sandy, how do I differentiate? How do I identify. that they represent the same consumer? That's a huge challenge. And the second thing around understand the visibility that you need as well as whether you can trust the agent, that's that's something that is is still, an issue. Right? Enterprises or ecommerce would like to adopt agentic sessions. They would like to adopt agentic traffic, but are they willing to open the gate with some level of trust? What's the level of trust that needed? So the question is not anymore whether it's a bot or not. The the question is whether you trust the traffic or not, whether it's an agent, whether it's a human or also a scraper or L and M crawler. And as you mentioned, everything is driven by an intent. The consumer intent can be a good intent. Right? They are here to conduct purchases, but the agents can easily be monopolized. So if you think about rogue agents, those have been given too many permissions and entitlements, they can do things that, applications and ecommerce are not yet there to support. And another thing, it's not just the security problem, it's also the business problem. When a human interact with the site, they have a way to see coupons or buy one, get one and upsell opportunities where the agent is driven by an intent of the consumer to buy shoes, for example, you cannot conduct that upsell opportunity in your inventory. So ecommerce may lose also that option of, building, let's say, more diverse, carts with agents. So in order to actually understand that, then that's happening, you need that level of visibility and trust of each session separately. That's actually what HUMAN security is doing with agentic trust. We categorize each session, let you know whether it's a human, bot, or agent, and provide you the clarity you need about the intent and whether, the session itself being compromised or not. Right. You know, Tomer, I think what's interesting about all of this and what's really probably very difficult for online retailers is they don't really have a say in how the customer is coming to them. You know, you can't you can't stop them from approaching you through certain channels. You can't say if it's automated, go away because then you lose the customer. And you can't throw out that friction that people used to do with various challenges and CAPTCHAs because you're going to lose people. We already had twenty one percent of adults that said they abandoned transactions because of challenges. But you can't throw up challenges to agents and or at least not not necessarily easily. And you you know, as as an online retailer, knowing that this is happening whether you like it or not, that you don't have that level of control that maybe you used to have in the past. Because in the past, you could create the mobile app. You could put out APIs. You had some measure of control about how your customers re reached you. And now it's kind of the wild west, isn't it? Yeah. Yeah. I. mean, I think. that's the that's the thing. Right? Is it's this change again happening at the the individual level first as we're all playing catch up to it. And looking at that traditional journey, you know, you had kind of, roles, technologies, functions form across that journey. Hey. My job is to advertise. My job is to gain attention. Then my job is to to to present them products and actually make sure that they convert. And then the security team comes in, makes sure it's secure along the path, and then the fraud team is coming in after the fact and saying, okay, are we making sure that those transactions are not fraudulent in some way? We're not gonna run into charge backs or some other challenge. This world is us all adapting to a compressed journey. This could all happen with a single prompt. Right? And you compress that journey, which in many ways, you're seeing the need for us to to rethink some of the paradigms of kind of those roles and functions in an organization as well as the technology stacks that they're using. Tomer, I think you touched on it, like, to gain visibility. Like, your first step is, can we actually see this happening through the entirety of that journey so we can make decisions and policies accordingly? And and then subsequently, obviously, our our decisions around with technologies and how we maximize it. You know, it's it is interesting because it's not we we've also often thought of the threat, right, of automation. This is also the massive opportunity. We can't let this go. We can't miss this opportunity. And then, yeah, some of those those mechanisms we've used for upsell and conversion might not work the same way anymore. So it's a it's a brave new world for sure and it's affecting, many different functions and I I think, Tomer, from I I know from your and my perspective, it's very interesting working with our customers where that those conversations are now expanding. Right? From a maybe a team that we've traditionally worked with for, our product suite or set of teams, that's now we're now having much broader conversations and more holistic. So exactly that. More for example, more than 60% of of the agenda sessions are around a product. They are looking and searching for a product. And as Sandy mentioned, if you block all of that completely, you. leave money on the table. You frustrate your users. You you actually, in the future, the agent won't go to your site anymore. It will go to your competitor site to to conduct those actions. So brands today or merchants, if they block all alarm crawlers actions, their products will not be discoverable. If you they block all AI agents, actions, it means that they leave that money on the table. And if they allow all AI agents actions, you leave yourself valuable to risk, to fraud, to rob agents. So the idea of providing more granularity around the actions and controlling what are the things that AI agents can do versus not providing something that is more holistic, solution to adopt agentic commerce and agentic traffic. For sure. Absolutely. You know, and in that vein, just thinking through the the technologies today. So, you know, we I know where where we sit obviously as Human. We've traditionally been around bot mitigation. Ultimately, that's turned into what we consider cyber fraud. We protect the advertising ecosystem and looking for automation of, you know, invalid traffic and clicks and views in various ways. But this has become much broader. Right? I know we're having broader ecosystem conversations of who we integrate with and how we talk because of that. Right? That because of this greater need across functionally. You know, Sandy, from your perspective, what are you seeing across the stack of technologies that this is affecting and then kinda how is that changing? How do we have to work together or ultimately, in some way combine forces to solve this problem? Yeah. Great question. And, you know, I realized several months ago that thought management was no longer sufficient for what we had to do. This is application traffic that we're handling. So the application security vendors that and tools that I've worked with, they make sense to handle this traffic, but the way we have to look at it has evolved completely because now it's, as you said earlier, it's not bot or not. It's not good bot or bad bot. Now we have to analyze agent traffic and say, who is this? Who does it tie back to? Do we trust them? What do we allow for them? And how do we use that to better enable the business? From my perspective as an analyst, this is really exciting for me because the the the, soapbox I've been standing on for years about security to enable business is suddenly very, very relevant because security absolutely has to do this right or to the point you you guys just made, we lose customers. Right. We lose business. We don't show up in search results. And then the business stops working with a security well, and that's not what we want. So it becomes really critical for us to have the tooling that's going to identify the intent of the agent traffic, identify the identity behind the traffic, still identify when it is hijacked or misused or malicious, and really be able to have that degree of continuous trust going on. So it for me, it is it is exciting to be in this situation where we really do now become a business enabler. The other part of this, I will say, and, Tomer, you alluded to it a little bit, that you're dealing with a broader set of, team members than you ever have. You know, for years, when we looked at bot management, there was always the the approach and the thought that you needed marketing and ecommerce and the digital team and security and fraud involved. And the actual adherence to that varied. Let's let's put it that way. Now it you can't. You have to have all of those people at the table because this directly impacts business. And it's just it it's so fundamental that we now really do have to practice what we're preaching and bring all these teams together because it is a holistic problem. And, For sure. again, as an analyst, this is the fun part for me because it's making everyone come to the table. Yeah. Absolutely. Yeah. It's it's funny. I mean, we're looking at it just again back to the kinda integration ecosystem and, you know, Tomer hop in, but it's we're we're looking at how are we working with those customer identity access providers. We're working with payment fraud providers. We're working with, those who are tokenizing wallets and trying to apply that to agents and a whole collection identity verification providers going into kind of more fintech and now your customer type use cases. There's there's a whole ecosystem now that is all looking at this saying, how do we work together to support this use case? And do we have commonality? I think that's key. Right? Is we can all be rolling out our version of agentic views into pieces of this puzzle, but I think a highly integrated solution where we do give complete visibility is the first step is is key. You know, the the commerce team, the digital team, the marketing team, understanding what agents are coming to site is the same collection detection and, visibility that the security team has and and subsequent teams as well. So I think for us is is, equally as exciting because of the complexity of those relationships, not just the the overall market, but, you know, at customer deployment. So I know we're asked being asked, Tomer, to to integrate with new tools and our visibility that we we haven't in the past, which is really exciting. Yeah. If you think about it, most of the protocols are there to allow the transaction to happen, but they have no visibility into what happened before the transaction itself. Whether you, can trust the agent, conducting that action or transaction or not. If you think about it, if you provide a an individual a credit card, are you always willing to allow them to conduct purchases? What about the other types of of risk that would happen even before the purchase took place? So that's something that, here's come to come to mind, the needed visibility and intent, understanding even before the transaction happened. And those protocols are great. They allow some level of, control, you could say, and it really depends on whether the agentic solutions are willing to adopt them. You can see as an example the bot authentication web bot authentication, protocol. If you think about it, some commercial AI agent solutions adopted it. So, cryptographically, you can verify the traffic. You can understand it that it's coming from a specific vendor, let's say, OpenAI or AWS. That's something that we did as well, but some didn't. Right? So opening that up without the ability to really understand each session separately and its goal, it's it puts yourself still in a risky situation. Right? You need that kind of a holistic, you could call it firewall, or less line of defense even for those that adopted the protocol, and and especially for the those that didn't. Right. I agree with that, Tomer. I think that the protocols get you part of the way there. They certainly address pieces of it, but they alone don't help you establish ongoing those trust relationships and help you validate that the intent of the traffic remains what you think it is because it doesn't account for things like intent hijacking. So there there is that ongoing need to also still collect the signals, conduct the analysis on top of the protocols. You need the protocols. Don't get me wrong, but it's not enough. So it, yay, we get to solve hard problems. You know, and and is this something just to to to take it up a notch? And maybe, Sandy, from your perspective where you're looking more broadly, I mean, we're talking across our customer base, but broadly, is this something that, the market is understanding today? Is this an, a trend that goes beyond us all seeing it coming, but are people actively working towards this problem? And do you see it really having a kinda sustained impact? I know we all we're all saying it's there, but is this is is this something again that the that the market is saying, yeah, we absolutely need to rethink this all. We're getting there. I think one of the things you said earlier, we talked about how quickly this market is moving and how much everyone is struggling to keep up. So I do think that there are some folks there that are thinking about this, but there's going to be a really long tail here. And part of our job as people that are looking at this is to help bring folks along and help understand some of the nuances and challenges of this, that this isn't just bot management. This is botan agent trust management, that this isn't just about good bot, bad bot, but this is really about understanding the individuals and the intent behind agents, that there's a lot more to it than there used to be. So I think that there are some customers that are starting to get it, but I think there's a lot of work that we're going to have to do as industry professionals to this will sound corny, but spread the gospel. Yeah. Yeah. For sure. Yeah. I I mean, Tomer, I know we're seeing the same thing. Right? I think it's, everyone knows it's a thing. We're all seeing thought pieces out. We're all seeing the research. We're seeing protocols. We're seeing partnerships announced. Then when we, you know, we hit the ground and we have a conversation with someone who's living and breathing and protecting their resources or trying to to sell more, I think the unknown is it it the the consumers outpacing in, the the enterprise is supporting them in such a material way. Right? And they're saying, I I'm not even sure whose profit is in my organization. So, like, how I know how we've been kinda coaching. Hey. Let's bring them in the room. Let's talk about our solution. We're trying to do it. But, you know, between, from both your perspective, how how do we continue to push that? How do we make sure that people see it in a way that they see the problem and they can start to build update their policies and review their technology stack and and how they're supporting? Like, how do we push the, the gospel, if you will, Sandy? I mean, in a material way beyond the the we've gotta show the data. We've gotta show how we are analyzing traffic. We have to show the signals. We have to show what things mean. the conceptual? We have to show the examples of good and bad. We have to get people need to see the data. I mean, I had this conversation a few weeks ago even with during a panel with someone on our b to c marketing team and talking to her about, you know, how should security be approaching marketing about some of this stuff. And what she said to me during that conversation was, you know, we'll we'll accept the feedback, but you've gotta come on with the data. Yeah. So we if we're going to convince anyone that this is happening now and that this is something they need to continue to prepare for and adapt to, we gotta bring the data. Yeah. And that's exactly what what we see. As I alluded before, we see AI agents already searching, for information. We see them also interacting behind login and also with the account information itself, those that are changing password information, credit card, trying to do that, as well as lots of registration attempts. We saw at the beginning of, at the September, saw AI agents trying to complete check checkout and payments, but that was actually before, some, protocols were there as well as before agentic browsers that are built without guardrails, also were were announced. So at the October, we already saw AI agents, conducting more check checkout and payments together with some malicious activities such as fake account creation, etcetera. Our trending intelligence team also found that some of the traffic is already spoofed. So we saw malicious, players already, getting into sites as if they are, driven by OpenAI traffic, but it's actually not, trying to scrape the data, steal information, and conduct unwanted activities, clocking themselves. So, we already see that statistics and, sharing that widely, hopefully, for more, companies to realize that the change is already here. It it's happening. Yeah. I think, I mean, Sandy, you touched the data and visibility. Right? I mean, Tomer and I, I think, speak, you know, multiple times a day about this as you can imagine. Here at HUMAN is, let's get this in the customer hands. We can see it today. Right? We are seeing this traffic. We've categorized it. As Tomer's point, we've seen certain trends starting to form and and new threat vectors as well. But the goal for us is, let's let's actually get it turned on and and roll it out to as many of our customers as possible so that we can start to make those decisions. We can iterate. It's it's changing every day, but so many folks are just flying blind here, in this space. And and I think to your point, that's that's fundamental to the change, and that starts to build that cross functional bridge as well, across teams at an organization. Yeah. You know, I Sandy, from your perspective is that, you know, we we've seen a a natural place as as those who played around bot mitigation and and similar space for some time like this. We were already there. We were seeing this traffic, so it made sense for us to to really dive deep into this. Do you see that when you look at kind of the vendor space that this is the natural progression of. bot management and it's a. place where you should see this coming from? Absolutely. And that's why I made the decision a few months back to really change the name and focus of the market. It's not bot management anymore. It's bot and agent trust management because you're dealing with traditional bot traffic. Sure. That's still there, but there's agent traffic as well. And it's not just about good agent, bad agent. There's that trust enablement, trust establishment as well. So I felt it was really important to include that in there. And I think as I look at how this market is evolving, yes, we're see it's such a natural place for the formerly bot management market to go because you all were collecting the signals anyway. It's application traffic. It's understanding how the application traffic is flowing and the signals around intent and all of that stuff. And so now what we're looking at is more of that, harder, more integrations to your point, more different sets of signals to collect, but this market was already on its way to being able to do that. So, really, what I'm expecting is an extension of capabilities that really help customers navigate this world. Yep. And and, actually, in order to help our customers adopt Adenty Commerce in the coming shopping sprouts here in, globally, we decided to, provide our enterprise customers, the ability to, use, agentic trust, in order to understand the the needed level of intent, visibility, control, and trust for agentic traffic coming to their site to conduct purchases, so that, they could really adopt identity commerce while protecting themselves against the malicious activity that might happen, coming December. Yeah. That's the, For example. the the concern for us. Right? As we, you know, we are we are trusted in so many ways, of course, to, to protect them at the business time of year and making sure a bot doesn't take down your, you know, Black Friday sale in the traditional sense. You know, what, gave us, concern was for our customers was are they gonna miss the agent shopping, right, when when they could be in the, you know, the newest, shoes drop or the great, you know, whatever the the new toy is, whatever the Tickle Me Elmo of, today is. Right? And, and everyone points their agent to go buy it. Are we gonna block that? Is that lost revenue? Right? And so getting this on immediately was key for us, and I know we're we're aggressively making that available where we can. Well, great. I'm looking at time. I I think we're, you know, probably coming towards a a natural point here. Any additional thoughts, Sandy, Tomer, kind of we've covered a lot. Obviously, what's changed the consumer level, what's changing, in the overall market, who's adopting it, who it affects, and then obviously where we see, you know, the folks in in our space about mitigation alike, needing to evolve to support that. But any additional thoughts from I'll throw one thing out there, Yeah. Aaron? which is that, ecommerce providers and online retailers need to keep a pulse on consumer preferences in terms of what they trust and how they use agents. We've looked at there was we did a Consumer Pulse survey in May, and at that point, 27% of US adults said they trusted tools more than people to choose a movie or a TV show to watch, but only 19% trusted tools more than people to pick out clothing to purchase. So I think one thing that's gonna be really important for folks to keep their eye on is how do consumers look at use of agents and automation to perform different shopping tasks. You know, just the thing they're buying you know, what's what's the level of trust they have in an agent versus doing it themselves, and that's going to impact how they engage with the site, which if you're a site owner, then staying up to date on that is going to impact how you choose to engage with some of those folks and what you learn to watch for in terms of changes in agent traffic. Yeah. I would add to that that the future is also around identity. Right? We we mentioned here, the consumer identity. Right? Who's doing the whose intent was that? But, also, soon we'll see the agent identity. Right? An agent represent, multiple consumers. It has its own, maybe, identity or ability to identify so that you can trust it better. You can, allow it to, go on and conduct the purchase versus not. It's more fundamental for maybe the future of Swarm, where multiple agentic systems will communicate, with each other in order to complete an actual, transaction or happen. So they, let's say the intent of the consumer gets, lost sometimes, and also the identity of the consumer when it's like a chain of agents doing things on be on behalf of that consumer. So I think that that's gonna be an interesting future for registry of agents and, more of the a n s situation if you compare it to the d n s, world. And that's something that we also are already part of, like, defining what would be the future with multiple organizations like, OS and agency, etcetera, to be able to adopt multiple, protocols as well as, ways to identify the agent. Yeah. Yeah. I think, you know, I think back to your point, Sandy, is, you know, when we look at recommendations for, organizations and how we stay ahead of it, let's let's actually make sure we're ahead of what's being adopted, what are users using, let's get the visibility. I know I personally, I'm increasingly trusting in my own personal LLMs and agents to do things on behalf. They've gotten so good and frankly are. terribly convenient. But, you know, where where does where does my consumer level trust stop? I'll have a book of reservation, but maybe not buy my clothes yet. Right? As an example. And so understanding that, I think, is key. And then having visibility on the enterprise end as well as the consumer end, is key and I think, Tomer, bringing that to what you were describing, as we continue to evolve the underlying infrastructure of this and we're signing agents, we have the concept of the identity behind it, we verify that identity in that wallet for the financial transaction, Where do we make sure that if we continue to maintain understanding of the intent through the process? Where do we, keep that visibility up? Because it's gonna continue changing as organizations if we start to direct it purely between, you know, direct the agent to an agent to to conduct the transaction, do we lose something there? And so I think, you know, coming back to that visibility piece is key, and the identity piece. You know, I I see a world where identity is, again, is no longer John Searby and some of my basic information, it's John Serby and he uses GPT to do this and he uses perplexity for this and here's his behaviors around it. So building this much more robust knowledge graph around what we can trust and what we expect from, the identity behind this these agencies, I think, is is key. But in until we, until we gain the visibility into it and we really see it across the board, we're we're gonna be flying blind. So that's super important. Awesome. Well, I I really appreciate it. I know we could keep talking. I mean, this is, we've talked a number of times and we could keep going here and, where it might go, some of the latest tech. There's some really amazing stuff being built every day, I know, and we are continuing, to focus on it. But, you know, maybe we'll have a follow on session as, in a few months when this actually changes all again. But, thank you so much for the time. This is this is moving so rapidly. I think it's it's critical for, you know, Human, it's critical for our customers and Sandy, you're you're, preaching the the the good news here, if you will, on on these changes coming which is hugely informative with all of us. You know, I think just in so many ways, the old foundations and frameworks on which the Internet was built are changing. Search is changing. Ads are changing. Where we place things, how we get attention, and then obviously the transactions we talked about. And if we're not having these conversations, it's gonna either become some Frankenstein we can't control or we're gonna lose visibility and not seize on the opportunity as well. Sandy, I know we I think I failed to mention we have a number of documents here posted in, in the webinar. We do have live chat as well for any q and a. Sandy, anything you want to highlight for the audience here on on on where they can look for more information? Yeah. I mean, from my perspective, I think, we've provided either in the chat or somewhere else, links to a couple of blogs and resources that I've, contributed, including my blog about about an agent trust management market and the the thought process behind that renaming. I also I think we have the link there to the rest of my blogs if folks are interested in taking a look at any of my research around application security more broadly. You know, obviously happy to share that that information out. So yeah. Hold up. Thank you. Yeah. And again, from a HUMAN's perspective, thank you all for joining. I know we've been, we've been protecting, many of you for a long time from the threat of automation and, this is evolving. And, you know, it's no longer we've we've always evolved ahead of the threat actor. It's now how are we evolving with the consumer as well and seizing the opportunity in addition to particular threat. So, you know, Tomer's focused on it every day. We're releasing new features, as well and I think, again, a lot of content here that you can, you can see from the webinar, but also feel free to ask us any questions or or reach out. We'd love to, to hear your perspective from the audience and what it's changing for you and how we can help. So, Tomer, anything to add? No. Actually, we covered a lot of ground here, and I'm excited to see what would be the future of AI and agenda commerce. Hopefully, we'll have a follow-up session with, some great, insights. I'm here if you wanna ask any questions on the chat. I would love to to hear your thoughts. Perfect. Sandy Tomer, thank you for the time. Appreciate the conversation as always whenever we chat. So thanks again, and I, you know, look forward to continuing to explore what the world of agentic commerce brings for us and ultimately the idea of agentic trust as well. Thanks Thank you. you. again.