Retailgentic, presented by the ReFiBuy Media Network and hosted by Scot Wingo, is where Agentic Commerce meets retail innovation. We help retailers and brands prepare for the future of agent-driven shopping with actionable news, expert insights, dxis, forward-looking predictions, and practical guidance on Agentic Commerce Optimization (ACO).
Yeah. What's gonna happen is there's gonna be... You know, out of 25,100,000 shops online, there's gonna be 75,000,000, AI agents trying to, deal with that. So how are people going to... People still have the problem of how do I find what I want to buy.
Speaker 2:Welcome to the Retailgentic podcast, where we explore the intersection of retail, ecommerce, and AI agents, which we call Retailgentic. Our mission is simple. We are shining a bright light on the new Agentic AI shopping megatrend which will disrupt the trillion dollar digital retail market. The stakes are high. Google, OpenAI, Microsoft, Perplexity, Amazon, Apple, Meta, and Anthropic are all fighting to be the Consumer AI Shopping Agent winner.
Speaker 2:There are gonna be some tough decisions for online brands and retailers to make. What happens to checkout? How should payments work? Could you continue to invest in retail media networks? What does the world look like when over half of your traffic is not from humans but from AI agents?
Speaker 2:These existential questions are the tip of the iceberg. On this podcast, we're talking to industry experts, analysts, and influencers to get answers to these questions, identify trends, and best practices. This is Retailgentic, your podcast partner in navigating our AI shopping agent future. And now here's your host,
Speaker 3:Wingo. Welcome to this episode of Retailgentic. I have special podcast for all you protocol nerds out there. So... And here in...
Speaker 3:In an Augptic ConverseLink, we sling around, phrases like ACP, UCP, ATP, MCP. For those that aren't familiar with the content covers, it can sound a lot like vegetable soup. I get a lot of interesting pitches for the podcast, and I filter them all for you, goodness, the audience. I wanna make sure that we're bringing you something that's really interesting and different than you're gonna find anywhere else. When this next group approached me, the pitch was so unique and out there.
Speaker 3:I wasn't sure I believed it until, they walked me through a demo. And I was like, holy cow. This is real. And I was excited to get them on podcast to kinda share some of what they built. I think it tells us a lot of things, although you form your own conclusions.
Speaker 3:But it's one of those that may make your head swim around a little bit, but it's really fascinating. So, also, this is a little bit more complicated area, because there's there's two founders for this company. It's called Vendee, v e n t e d, And that's the opposite of Vendee d. So that's like the the kind of hook up their mind to all. And so first, we have one of the founders as Ross, and he joins us from Brisbane to Australia.
Speaker 3:But... So he is... This was recorded his morning, early afternoon. And then the other co founder is all the way off the coast of France, so he is in the European time zone, which it's more like, probably, I don't know, midnight or one a. M.
Speaker 3:List. So this spans three... I'm in North Carolina. So this spans three different time time zones that is what recorded, which is my blogging experience scheduled. And, you know, I don't wanna spoil it, but you will see on this one a, you know, a live demo of something I'm pretty sure that has never been done live before on the podcast.
Speaker 3:So I hope you enjoy it as much as I did. I learned a lot talking to these gentlemen. I think it's pretty interesting what they've built. He's... We will see where it goes.
Speaker 3:It's still early days, but enjoy conversation and the. And they mentioned... I should say everything that they mentioned will... Is in the show notes. So if you miss any of that and you're listening, be sure to check out the show notes for more information.
Speaker 3:Thanks and enjoy the conversation and the bit.
Speaker 4:Doctor. Darryl Carlton and Ross Gardiner, cofounders of Vendee, welcome to Retailgentic.
Speaker 1:Hey,
Speaker 5:Scot. Hey, Scot.
Speaker 4:Let's start with where you guys are. So, Ross, I believe you're in Australia and Darryl, you're always somewhere. Where are you dialing in from today?
Speaker 1:I'm dialing in from Jersey, not New Jersey, old Jersey. Old Jersey is a tiny island that's nine miles by five miles in size, off the coast of France. It's a protectorate of The UK, but not quite part of The UK. It's closer to Europe.
Speaker 4:And I learned that Jersey and there's a couple other islands play a big role in e commerce.
Speaker 1:Jersey and Guernsey, they're part of the Channel Islands, and they're a financial insurance center for all of Europe. And they have very friendly tax and investment environment where corporations don't pay any tax at all. No capital gains, no income tax on the companies. And in addition to that, extremely strong impenetrable data shield, stronger than UK data protection, cannot be pierced. So they're the two reasons we're setting up the headquarters and the legal entity in Jersey.
Speaker 4:And, a lot of people do it for the favorable tax treatment, for ecommerce. And then there's something around, is it also tariffs? It's like a tariff free zone or has
Speaker 1:it morphing? It's basically a properly structured, very legal tax free center where some of the tax free zones around the world are a little bit questionable. Jersey is not questionable, but still offers a lot of those same benefits. And so you see a lot of companies are headquartered here, even if they are not operational here.
Speaker 4:Let's start with some backgrounds of you fellows, and then we'll lead into why you decided to start this company. That is, you know, obviously, spoiler alert is gonna be very active in the terms of commerce, which is the whole purpose of our podcast.
Speaker 5:Yeah. I originally started my career, Scot, in in media and government and government. And about twenty five years ago, crossed across into technology planning and delivery, which has been quite a journey for me over that time. And I feel blessed because it's taken me around the world. I've lived and worked in six different countries, Australia, Singapore, China, India, United Arab Emirates, and Qatar.
Speaker 5:And so it's given me a really great world view, but it's also gave me a lot of exposure to very interesting projects. I ran a large development center, offshore software development center in India. I... I've done nation state level projects with with large systems integration, various things like that. And so that's how I've come to it.
Speaker 5:I've known Darryl all my life. We grew up together, and I feel really blessed to have been given the opportunity to join with him in founding Vendee Labs.
Speaker 4:Where in Australia are you?
Speaker 5:I'm in Brisbane, Queensland. Sorry. The great part of Australia.
Speaker 1:And Darryl, what's your background? I started in IT straight out of high school when I was 17 years old. I dropped out of, well, I did six months of an accounting degree and looked around the room and there were three fifty potential accountants just in my college. And I couldn't stand the idea of being another accountant. Dropped out and got a job in IT and that's all I've ever done.
Speaker 1:I got all of my qualifications part time while working full time. And early in my career, I was a national operations manager for Australia's largest retailer, Carlton, Meyer Group. I've spent most of my career in telco, especially in the customer facing and billing systems areas, running very, very large projects. In total, I've run over $3,000,000,000 worth of IT projects, about 30 projects in all. My doctoral research was in why IT projects fail.
Speaker 1:And my second doctorate is in how the courts treat failure and how they deal with it. And I've published eight books on AI and cybersecurity, and now I provide training and consulting on AI governance.
Speaker 4:What's it like to be one of these longtime AI people? And then suddenly we had that ChatGPT moment with, you know, the advent of GPTs and all of them. And suddenly, you're the cat's pajamas.
Speaker 1:Yes. I got my first AI qualification, degree in 1983. So I was writing in prologue and list back in the early eighties. I was doing things like fault diagnosis of telephone networks using artificial intelligence, airline optimization in Prolog, things like that. It's interesting.
Speaker 1:We... I think back then, we always thought that AI was going to be a back office function. We didn't imagine it was ever going to be something people would interact with directly. It was always going to be behind the scenes doing stuff in the background. ChatGPT was a bit of a shock to me how it became such a consumer darling.
Speaker 4:This is a little bit off topic, but since I am an AI expert, I have to ask this. Do you think, GPTs get us to, you know, agentic, you know, or superintelligence or what are general intelligence? Or do you think they'll have to be something else that kind of has to kick in?
Speaker 1:No, I don't think so. I think most of the real experts generally agree that it won't. Even Darryl Amedee has said that about 70% of what Claude puts out, they can't explain how it does it. And it's not really going to achieve genuine intelligence until the tools can explain their reasoning. That doesn't make the tools any less useful right now, but there are things we have to do.
Speaker 1:I'm of the school of thought that until they can explain how they achieve something, they're not really going to be an intelligent. You think about how we were in primary school. Your teacher would say to you, show me you're working. It's not enough to show me that you can add two plus two equals four. Show me on the side, show me how you arrived at that.
Speaker 1:The problem with Gen AI is it can't show you how it arrived at that. The old expert systems rules based on Bayesian calculus would do backward chain reasoning and would explain how they arrived at a result. We've lost that capability at the moment and that has to come back before we can really achieve intelligence.
Speaker 4:So so you guys knew each other very young. You went in your own ways, and now you've come back to start a company. What... Tell us the founding story of Vendee and and what inspired you to do this, and you know, we can also go into what is Vendee going to do?
Speaker 1:Yeah, so during COVID, I was asked to help my then wife with buying things. And she was complaining that she was finding it very, very difficult to buy online. She had to go to too many websites, had to do too many things. And I started exploring this and talking to other friends. And I had a friend who was very successful in online gaming and he was very big into trading sneakers and trading sporting cards, playing cards.
Speaker 1:I mean, really high end playing cards. Yeah. Million dollar cards.
Speaker 4:Sneakerhead Yeah. And a card guy. Yeah.
Speaker 1:Oh my god. Yeah. Just incredible. And what I realized was there was a big gap in not looking at e commerce from the buyer's perspective. So, I had a first version, I had a Vendee version one, which was about setting up a buyer's marketplace.
Speaker 1:And I built that during COVID. And my dev team was in The Ukraine, and then they had to run. And for various reasons, that fell apart and I moved on to doing other things. And as ChatGPT and other Gen AI tools came about, I started to see another way to do this. And so I started working on trying to figure it out, and looking at Agentic Commerce from the buyer's perspective.
Speaker 1:I've spent about nine months trying to understand how to do this. I've submitted four patents for examination on this. And when I realised I was close, I contacted Ross, who I knew had incredible skills at driving team development and is very highly regarded at building and delivering complex systems. And I couldn't do this alone. And so we've started to assemble a team, build the prototypes, put the design together.
Speaker 1:And with my background in large scale systems, I mean, some of the data is quite interesting because this is a big project. You know, when you get up to the big dollars of big systems, 90 to 95% of IT projects fail to deliver on time, on budget with required features. The staggering huge failure rate. Projects that spend over at least 25% of their time on getting the design right have an 82% likelihood of successful outcome. So we've been working really hard to get the design right.
Speaker 4:Ross, what did you think when the phone rang?
Speaker 5:It really resonated with me. It was a story that I had experienced myself as a buyer, you know, that the fragmented discovery, all the friction of trying to find things, the hours and hours spent researching and comparing. I knew with the arrival of AI, there had to be a better way and one that could be focused on on what I needed. And when Darryl spoke to me, I could see that there was a real opening there in the market where we could we could refine things and deliver a better outcome for for buyers, but also at the same time, you know, do the linking and matching up with the needs of merchants as well.
Speaker 4:The story started during codes. You started in 2020, basically. And so you've been at this four or five years before even ChatchBT. Okay. Wow.
Speaker 4:Did ChatGPT's emergence, cause you to pivot or it was like right in line with what you predicted?
Speaker 1:I had tried back in twenty twenty one, twenty two to figure out how to apply artificial intelligence to it. And I couldn't I couldn't make that leap. I couldn't quite get it to work. ChatGPT and then Claude dramatically changed the approach. I've disposed of version one.
Speaker 1:It's gone.
Speaker 4:We'll call it a useful prototype.
Speaker 1:A useful prototype. I I did keep I did keep the graphics. That was all I kept.
Speaker 4:Sometimes you have to do it wrong to figure out how to do it right. Yep. And, give us the pitch, of... Okay. So we understand the problem you're solving.
Speaker 4:So go ahead and tell us what you've built and where you are.
Speaker 1:The pitch is we're changing the way the world buys. We're building the infrastructure layer for ecommerce from the consumer's perspective. Right now, are the numbers are 25,100,000 shops online.
Speaker 4:Globally. That's going to be a global number.
Speaker 1:It's a global number. And what's going to happen is there's going to be, you know, out of 25,100,000 shops online, there's going to be 75,000,000 AI agents trying to deal with that. So, how are people going to people still have the problem of how do I find what I want to buy? Then I looked at an interesting statistic out of The UK. It took a few years, not that long, but over a period of eight to ten years, insurance purchasing went from 95% direct with the insurance company to 85% being through aggregators or compare the market type operators.
Speaker 1:It's clear that consumers will vote with their wallets to for a better buying experience when it's directed at solving or removing friction. And so, what we've been trying to figure out is how to remove friction for the buyer in e commerce. What has surprised both Ross and I is as we've solved buyer problems, we have solved huge problems for the seller. At the moment in the seller market, you have 70% cart abandonment, and that figure hasn't changed in fifteen years. 50% of the seller's costs are in promotions, advertising, referrals.
Speaker 1:Most sellers don't get to a high Google ranking and they never, you know, on the front page. And there was another figure that really staggered us in the payments area. Mastercard, alone, just Mastercard, from their own data, spends $41,690,000,000 that's with a B, every year in chargebacks. And most of that cost occurs because reconciliation of the transaction that occurred with the consumer is very complicated and very expensive. And a fully audited, governed AI environment can solve those problems, solves all of those problems.
Speaker 1:And those were sort of unintended consequences of creating a better buying experience. Those weren't the problems we were trying to solve, they're extras.
Speaker 4:So, your user... So consumers will go to your website or app and they will use this to search for things and then agents go and buy on their maps. Is that that kinda how it works?
Speaker 1:It will. Now there's a lot of stuff that has to happen for that to occur. It's not just a technical change. So, even if the technology, even if the AI is able to do it, there's legal infrastructure that has to be put in place to allow payments to be automatically applied. So right now, there's a couple of little ways you can have the AI agent fully execute on your behalf.
Speaker 1:Give them a prepaid credit or debit card. But the agent isn't fully autonomous yet. By the end of twenty twenty seven, a lot of the legislation will be in place in Europe and The UK and a few other countries to allow fully autonomous agent execution.
Speaker 4:Okay. I think I'm a visual person, and I think you guys came with a demo. So... And we do have some audio folks. So go ahead and do a little voice over while you do this so that the people are just listening.
Speaker 4:They can kinda, like, get an idea of what you're doing.
Speaker 1:Alright. So this is literally a demo. We're on the browser. This is a demo. So I'm gonna...
Speaker 1:There are things that will come up in the demo that won't be in the production version. So we've got a lot of it canned. So I'm just going to go straight in the application itself. And what you can see here is we've gone into what are you looking for? And you get just a normal dialogue screen exactly the same as you would in any AI Check GPT style of environment.
Speaker 1:So, we get to look at that screen. So, I'm going do a couple of simple things here and show you how you would have a conversation. So, I might start off just with a very simple request exactly as you would do in Google, just putting in some keywords. Like
Speaker 4:running shoes.
Speaker 1:I typed in running shoes. Now, that's exactly as you would do it in a Google product search. Now, so it's going off. You can see in the left window there, I've got the... It's got the ACP protocol occurring.
Speaker 1:Now we are live. So even though this is a demo product, all the connections are live and actually happening. So we're going out to something like 500 retailers, 50,000,000 products being evaluated. And you can see now it comes back, it's gone out, it's made connections, it showed you the ACP handshake, it showed you that it's requested the product catalog and we've retrieved a set of products. And I get a little bit of analysis at the top saying the top running shoes for twenty, twenty five, twenty six include versatile high performance options like Adidas, etcetera, etcetera, etcetera.
Speaker 1:And we get some agent recommendations down here. Now these agent recommendations are limited. We only have up to nine recommendations in here that we get all the products. And so there's hundreds of products have been searched and we've retrieved them here and you get a fair bit of information about them. But what I wanted to show you here is how you would use just the normal Gen AI conversation.
Speaker 1:And so now I can put in a much longer, I can put in a more descriptive field, which you couldn't do and get a response in a normal Google search. So I need shoes, and I can't spell, suitable for all weather conditions and road surfaces. And I'm gonna refine that search and the retrieval based around a more descriptive conversation. So now it comes back and it tells me, I prioritize shoes with durable outsold, weather resistant features and versatile design suitable for various surfaces. Premium models with advanced rubber compounds and engineered uppers.
Speaker 1:So now it comes back with a different set of ranking conditions. So you're able to do this sort of dialogue and get information on the products. If we go into the products, we get even more information, the price, etcetera, etcetera. So if you don't mind, I'm going go back and do a new search and show you some other features in a different way. I'm just going to take One of my standard searches there is sustainable basics.
Speaker 1:Now remember, while the product is a demo, these are live searches that are acting on real vendor, real seller websites. So now it's gone off and it said, based on current sustainable fashion brands, we've identified, their versatile products that combine ethical manufacturing with comfort within your $400 budget. So it's extracted from my avatar some priorities and some preferences that have been applied to the agent search. And so I can go in and I can have a look at And the product card has a lot of information, much more than what you would get from a normal keyword search. So it's got a full description, a stylish, environmentally conscious lifestyle sneaker.
Speaker 1:It's got the estimated delivery times. It tells me which vendor it's coming from. It shows it's fully compostable, high for light, whatever that is, a 100% biased bio based material, coconut fiber. So it's got a lot of information about the product. On the next one, you can see that the retail price was £142.
Speaker 1:It is in British pound at the moment, but it's available discount to 106. So it's drawing discounts, back from the vendor to see what they are. And I can change that currency as well. So that's a 130... 180 US dollars original price are now a 135.
Speaker 1:And it changes the tax. Now this one shows that it can deliver in three to five business days shipping to William Shakespeare's wife, Anne Hathaway, who lives in Los Angeles. So you can see you're interacting with the AI agent while you're shopping. Now, if I buy a few things from here, so I might buy those shoes, buy this one, buy this one.
Speaker 4:So you've got a multicart, you've got an ACP multicart.
Speaker 1:Correct. So the advantage of the ACP multicart is I've now got four products from three vendors. And if I go, I can have a look on these products. I can see there's been some discount applied to them. I get the total price, including the Vendee costs.
Speaker 1:And as I go through in this multicast, these are all coming from different vendors. And I can now go to shipping and package these up and it will go right into payment. And you can see now I've got my combined shipping cost. So even though I bought four items, I've only got three vendors. So I've only got three vendor costs involved in that.
Speaker 1:So I've consolidated the vendor costs. And so this multi vendor cart now will handle the handshake and communication with the independent vendors at the back end. So I don't have to buy as a consumer, I don't have to buy on three separate occasions. I don't have to go to multiple websites to complete the transaction, complete the ordering. I can compare all the products and I can buy from different vendors.
Speaker 1:And it's order confirmed, we can view the order details and see what we're getting. It shows the full price and the discount.
Speaker 4:Did you just really spend £400? We didn't have to do that for our shop.
Speaker 1:Yeah.
Speaker 4:There'll be a
Speaker 1:lot of
Speaker 4:future returns.
Speaker 1:Well, yes. Now, the... Well, that's interesting you rate the future returns. So because it's agent... Because we we set up all of the parameters, all of the fields within the avatar of what the profile of the buyer is.
Speaker 1:And those are passed through to the agent as unbreakable permissions. And of course, because it's all all of these rules, it's fully governed from the design phase. It means that any claims, any refunds have a complete audit trail associated with them. And in line with A. I.
Speaker 1:Legislation that's occurring in most countries around the world, every A. I. System has to have an audit trail legally to show what decisions it made, how it made those decisions, what authority it had to execute, and where requested, where required, it has to be able to produce this information in a log. We can do that. Now where that benefits is this directly starts to improve all of the payment platforms and the vendors performance in chargebacks and complaints.
Speaker 1:We dramatically cut the cost to suppliers and payments processes associated with, I didn't get the right product, I didn't get what I asked for. And we have all of that data in the transaction record.
Speaker 4:It'd be interesting to see the first Agentic chargeback and the... Will the agent's audit trail win that battle or will the consumer win that battle?
Speaker 1:Well, the agent's the the audit trial will show what was requested and what was provided. The... There will be circumstances where the consumer says, no, no, no. I didn't intend that to occur. I didn't mean to buy that or I didn't want to say that.
Speaker 1:But it's a fairly strict audit trail. What we know at the moment is that without hard permissions, Gen AI has been making, it's been exceeding its authority in a large number of experiments at different vendors. You know, MIT have been doing these studies. So, in about 25, 30% of the cases, the Gen AI tools have exceeded their authority. So, we are coding ours.
Speaker 1:So, it will in most cases have hard limits. You will in some cases have variable limits. So, instead of saying you are not to exceed $100 we might say buy it at the best possible price. Now, this is from the consumer's perspective, this is all about creating a system where you build trust, where it does the things that you ask it to do and you can check that it's done them the right way. One of the things that we found in building this prototype, which is very close to your heart, is where the product data on the supplier side is under specified, where it is only keywords, that's coming back with some very interesting results to us, where it returns the wrong product, but it picked up a particular keyword that was associated with what we're looking for.
Speaker 1:So, in order for Agentic Commerce to work, the suppliers are going to have to have rich, full expressive data on their side explaining the products that they're delivering and selling. Keywords are not going be adequate.
Speaker 4:I have a thousand questions. Let's keep nerdy on Agentic Commerce. So, by estimation, that was the first live demo of a multi item ACP checkout. Yeah. Okay.
Speaker 4:So that's pretty cool. So right there, just wanna pause a moment and acknowledge that that's pretty amazing. And, yeah, it's it's easy to talk about stuff. It's harder to build it. You guys have clearly built something pretty unique here.
Speaker 4:Ross was kind enough to share your deck. I'm not gonna show any of that because I don't know what's proprietary and whatnot. But you you you obviously... You know, being an expert at at these things, you've obviously built a layer on top of ACP here that that has a lot of the... It's just a payments protocol.
Speaker 4:And there's a big argument of even... Is it even Agentic or not? I I don't really wanna get into those really just first. But you built a layer on top of this. And then, you know, I'm assuming that somewhere in there, you didn't do anything with UCP probably because it's so early and there's nothing to, like, endpoint to yet.
Speaker 4:But I... You know, in the deck, you talk about ability to bridge these architectures. Let's get a little nerdy
Speaker 3:on that.
Speaker 1:What we're building is what we we call it a universal protocol bridge. So we took a decision early as we were trying to design it, not to pick a winner, not to try and not to make the choice of VHS or beta, if you like. And most of your listeners probably don't even know what beta and VHS is, but, you know.
Speaker 4:Two competing standards that guide long ago.
Speaker 1:And so we wanted to create a translation layer to abstract above those protocols. And we wanted to do this because the buyer doesn't care about those protocols. So on our side of the technology, so on the Vendee product side, it's all of our interface, all of our coding. And then we translate to each of the codes on the other side, ACP, UCP, H2A, whatever it is, even just plain old web search. And that takes a little bit of clever coding, not too much, but a little bit.
Speaker 1:And that means that as new standards come up, sorry, I'm waffling here a little bit. One of the key design principles in the coding standards we have is separation of concerns. And everything is a single task, but very micro, everything is a micro agent. So, talking to ACP is one micro agent. Talking to UCP is one micro agent.
Speaker 1:And so, parcel up our string, our API request to be passed into a micro agent and then receive that transmission back and put it together in the display. So long as we continue doing that, as standards evolve and as new ones come in and as old ones die, all we have to do is unplug something that's changed and plug something new in.
Speaker 4:So you're sitting up here and you can plug in ACP, UCP, MCP? Yep. What's that on? A to A, AP to everything.
Speaker 1:And let's say China comes up with its own or India comes up with its own. Yeah. We can plug into those. It doesn't matter.
Speaker 4:The end result is you have a buyer experience that could look across protocols and do multi... You could do multi UCPA CP checkout. Correct. Not able to demo that yet?
Speaker 1:No. We haven't got all of that running yet.
Speaker 4:So the distribution of this is... Your vision is consumers will find this this, you know, experience amazing and interesting, and they will be drawn to kind of a consumer interface, or are you gonna have partners that will help you with the distribution at some point?
Speaker 1:We're gonna do both. We think that there's a a really obvious market in the business area, which is to replace what are called P Cards, procurement cards. So in most corporations, the data that Ross collected was that around about 20 to 30% of the turnover of an organisation is spent on procurement, and of that between eight to 20% is on P Cards. Small purchases where the staff has a credit card or a debit card, they can go down to the local hardware store or office supplies company and buy stuff. So, we think in the B2B market, that's an immediate opportunity is to replace those P Cards.
Speaker 1:So, that's a big opportunity. So, consumer and small business to business transactions. We're also going to the market white label. So, we see a particular opportunity there with buying aggregators, organizations that want to buy on behalf of groups. And we have been talking in Australia to an indigenous buying aggregator who wants to be able to serve remote communities.
Speaker 1:And we can see that working. And we also have had conversations with large enterprises who are losing, I think Ross was at about 40%, up to 40% price disadvantage. And so they want to be able to price better and buy better and be more aggressive in their acquisition. So we do think there is the possibility of an enterprise solution. Now that would require additional work to fit into procurement gateways, you know, the whole of tendering, responding, integrating with SAP and Oracle.
Speaker 1:So, that's an additional layer of work we would look at at phase two.
Speaker 4:Is there a world where you would license this kind of like protocol bridge to something that needed it? Like, let's say I'm an agent builder and I want to build an agent for a certain vertical, sporting goods or collectibles. And I would look across all the endpoints and have you guys do all that work, and I wanna... I have an amazing new front end I'm developing or something like that. Is is that a use case that you would consider?
Speaker 1:We have, actually. We've talked about that. Ross raised that very early with me to say that he thought that what we had come up with here stood out on its own as being a separate opportunity beyond the whole product. That there was something very big here that could be licensed on its own.
Speaker 4:Cool. And then how big is the team and have you guys had venture capital, whatever you're comfortable saying?
Speaker 1:So we've got three of us as founders at the moment. Ross and me as the first two of the primaries. We've just brought on a guy to run development. We're talking to somebody about CFO role and I've got a team of developers of about, well, there's six people, but they're not full time. So, you know, about three full time equivalent doing the build right now.
Speaker 1:So it's not a big team at the moment and it's all been bootstrapped. We've had no external funding so far.
Speaker 4:Awesome. You built some, pretty cool stuff pretty quickly with a small team. Congratulations.
Speaker 1:Thank you. The tools are great.
Speaker 4:What's your, can you share what your dev stack is that you you guys prefer? Are you a are you a codex camp, clog code, none of the above?
Speaker 1:I'd rather not get into that.
Speaker 4:But this side too. Yeah. Totally understand. You probably... Knowing doctor Darryl, you just probably built your own something or another.
Speaker 1:We're we're writing everything in assembly in assembly code. We're not really.
Speaker 4:I knew that wasn't true. Or another thing that made me laugh is COBOL, so that was really crappy.
Speaker 1:Look, I did train on COBOL very early on, but I was I was a commercial coder in prologue, but I haven't written code myself in a very long time.
Speaker 4:Awesome. If folks want to learn more, where can they go? Well, let... Let's talk a little bit about the future of where this all goes. You know, maybe, you know, you you talked a little bit about it.
Speaker 4:Let's start with where do you see Vendee going? You've got a lot of different options you painted there. There's a b two b path and buying groups, all this stuff. You know, where do you see this going, and what's your goal here? And then let's talk about your broader vision for where you see this objective of Commerce piece going.
Speaker 1:Ross, do you want to address where we're taking the business? I've been dominating the conversation.
Speaker 5:That's okay. Look, we have a taxonomy of buyers, which ranges from, you know, online shoppers, families, communities, small business through enterprise business, other organizations such as buying groups. So I I guess we come back to our original statement about building an infrastructure layer for Agentic Commerce. Right. So we really are doing that in terms of agent to agent interaction negotiation.
Speaker 5:We're already seeing that surface in our demo with pricing discounts. We're also building preferences on behalf of buyers across that taxonomy, which can surface and is already surfacing merchant data. We're finding free returns, delivery, sustainable brands, many other parameters, and we're building, you know, custom rules there. Of course, there's a piece about having a protocol bridge, the building AI governance, and we're also working very hard about... In the area of credentialing and and identity.
Speaker 5:And we think that that's a big part of the overall solution. In the short term, we're we're we're working on our... Building our MVP right now, and we hope to surface that, say, quarter three this year. At the same time, I'm trying to build a an alpha community of interested people to work with us and test our product.
Speaker 4:So so in my role, the forecasters are saying, you know, by 2030, 20% of ecommerce will be ecommerce, which is okay.
Speaker 1:I just
Speaker 3:kinda think we're moving the the
Speaker 4:the buckets, you know, the water between two similar buckets. I get more excited that that... I've been in ecommerce for for twenty plus years. We've really slowed down, at least in The US and I think in the in The UK, I even think Australia. You know, we slowed down to, you know, kinda matching the the offline retail growth rate.
Speaker 4:Here in The US, we've stable... Stated that 20% of of retail is online. In UK, it's more like 30, 40%. In Australia, I think it mimics that at 30 or 40%. So, you know, number one, in The US, kind of laggard.
Speaker 4:I think it's because we have a pretty good physical store experience that a lot of people like it. But but I get excited that I believe that this will actually cause more lift and will start to... Ecommerce will start grow a good bit faster than than retail. You guys... What do you think, Darryl, on either of those statements?
Speaker 1:I've been doing some interesting research lately on speed of adoption versus forecast. So I worked at Gartner for a few years, and so a lot of the stuff that we did at Gartner was about trying to pick trends. And so I read a lot of these trends and I look at them, how they perform. And when you measure them back, were they on target? What we're seeing with AI is things are happening dramatically faster than has been forecast.
Speaker 1:And this is across the board. So AI is having a bigger impact faster than has been forecast. And Ross and I are fairly conservative in the way we put our business case together and our pack. And the message we're getting from people that we talk to is this is all going to explode. It's just going to take off.
Speaker 1:And I think there are going to be a couple of trust events that will change buyer behavior. And the one that I like to speculate on is Taylor Swift.
Speaker 4:That came out of left field.
Speaker 1:At her last concert in Australia, Australia had the biggest... So, we have some of the biggest... Australia has some of the biggest stadiums in the world. Know, a stadium in Australia seats 100,000 people. It's quite common.
Speaker 1:We have four or five that do those sort of numbers. Those stadiums sold out the tickets for a Taylor Swift concert in forty five minutes. Where trust will grow is when you give your avatar the authority to execute on buying tickets without coming back and asking for your permission because you want your next Taylor Swift tickets. And as soon as your AI agent does that for you and doesn't mistreat your credit card or your debit card or your direct banking transaction and get you what you want, suddenly trust is earned. And now you start using it for everything else.
Speaker 4:So I was thinking you're gonna say trust event that someone would have a massive fraud problem scoring Taylor Swift tickets. That's a good example. You spent a lot of time thinking and talking about governance and security and these types of things. Do you think we will have an event that's kind negative, one that wakes people up to that side of the equation?
Speaker 1:We will. Now, the analogy that I use in all of my training courses and in all my writing is Ford Motor Company versus Toyota. So, MIT University, Massachusetts Institute of Technology wrote an influential book in 1986 called The Machine That Changed the World, where they looked at, vehicle manufacturing in Japan versus America and why the Japanese were winning. And what they found when they did that research and wrote that book was that, Ford invested 25% of its factory resources and personnel into what they called quality control. And in fact, what quality control was, was when the car came off the end of the line, some big guy in Tennessee with a giant rubber mallet beat the crap out of the car until it was back into shape.
Speaker 1:It was fixing things that weren't connected properly. Toyota had nobody in quality control because they chose that quality is everyone's job and quality has to be built in. This is the same we're going to experience with AI governance today. Most people are approaching AI governance on the basis of we're going to check the output. So, they're standing at the end of the production line of an AI system with a rubber mallet and saying, this one's wrong, belt it with a rubber mallet.
Speaker 1:The good systems are going to be based on governance by design, and you'll have it built in from the outset. And those will grow. The others will die because they will do bad things.
Speaker 4:Like, go hog wild buying stuff without your permission or...
Speaker 1:Yep.
Speaker 4:It's kinda like...
Speaker 1:I'm trying
Speaker 4:to think, like, what would happen if the crazy teenager stole my credit card kinda thing. That's my identity. That's my my avatar for a bad agent. Okay. I do wanna be cognizant of time.
Speaker 4:I know you guys are super busy changing the Agentic Commerce world. If folks have found this... You know, these topics to be super stimulating, where can they go to learn more other than buying all eight of your books? Buying the... Number one, go buy all of Darryl's backlog of books.
Speaker 4:And then number two, where do they
Speaker 1:Yeah. My my publisher would like that. Go to our demo, vendeex.ai, and have a play, and then come and talk to us. There's links on there to talk to us. I know
Speaker 4:you're quite a prolific content producer. Do you have a sub stack or LinkedIn?
Speaker 1:LinkedIn is our primary channel. So I publish two or three times a week.
Speaker 4:Ross, anything on your side that you want to throw in here at the end?
Speaker 5:It's a... It's going to be a wild ride, I think, but one that's a journey that's going to be worth worth going on.
Speaker 4:Possible. Thanks for coming on the podcast. I really enjoyed the demo and hearing all the cool stuff you guys are building. Also checking off another one of the world's first demos that we've had here on Retailgentic, a multi item checkout through ACP. And we'll definitely be tracking your guys' success with which you...
Speaker 4:You know, from one founder to two others, wish you nothing but the best. I think it's very interesting, the area you're poking around on, and you've got a lot of optionality in there, and you'll find success along the water in those paths. We'll check-in with you maybe in a year or so and and see how things are going. And if you ever have anything else cool that you wanna show the audience, let us know. For listeners and watchers, we'll put a lot of links to all this stuff in the show notes.
Speaker 4:So if you wanna go find, you know, any of these resources and go play with the demo yourself. So so what what Darryl demonstrated there is live on the site, and that's, you know, that's another testament to what you've built. This is not some some laboratory experiment that's held together with duct tape and chewing gum. It actually works for a lot of different use cases. I I ran it through a little...
Speaker 4:It's paces with some Star Wars things, and it it passed with flying colors. So thanks, gentlemen. Really appreciate you coming on.
Speaker 5:Thank you, Scot. It's a pleasure.
Speaker 2:Follow, rate, and share to stay up to date with Retailgentic, your guide to the future of AI shopping agents.