David Brudenell: It doesn't matter in which way or how you think, but if you think it's going to change the world, there's going to be really two types of nations. There'll be AI consumers and there'll be AI creators. Like, we want to be the Toyota Camry engine of businesses. I think some jobs will go away. Yes, absolutely. That'd be silly to say otherwise. Do I think many jobs will evolve? Absolutely. Do I think many people will be exposed? I do.
Nathan Bush: Hey, it's Nathan Bush or Bushy joining you from the land of the turbo people here in Brisbane, Australia. There's lots of amazing AI tools out there and the promise of AI is so huge, so enormous and so exciting. And sometimes it's easy to think about it as the new Lamborghini or the new McLaren for business. But what if it was actually the secret weapon that made your business run like a Toyota Camry? That's the standard we're looking for here in today's chat. Today I am joined by David Brudenell, who is the executive director at Decider AI. David's had a fascina exciting career journey, moving across fintech, E commerce and digital innovation, and even studied zoology. Who knows how all that fits together? But today, David takes us behind the scenes on Decider AI, which is pioneering agentic AI as a business system. Its idea is to transform messy data from across a whole bunch of platforms such as Shopify, Salesforce Zero, into one agentic single source of truth. The idea being that this should be a system that connects all the data across your organization to help you make better decisions based on the goals that you're setting. Now, they are still in the early stages with beta going out in October. But what they're showing so far is incredibly promising for some clients. They've already achieved a 5% incremental revenue in just eight weeks. But what I love about this solution is that it goes after the goals that you set it, so it goes after the results that you want, not what the software wants. So today we will dive into the Decider AR technology. We talk about how businesses can move beyond using AI just for efficiency, but actually using it to enable growth. We talk about the future economy and the agentic economy and how David believes there will be two schools of both businesses and nations based on how we adopt AI technology. And there's plenty to learn in here for senior leaders who are in organizations and head heading up the E commerce function about what your role might be once AI becomes the decision maker. How do you make sure that you yourself remain valuable and leading the AI through the organization. So, heaps to unpack today. Fascinating episode. Here's my chat with David Brudenell, executive director at Decider. David, welcome to ADD to cart.
David Brudenell: Thanks for having me.
Nathan Bush: Ethan, awesome to have you here. It's a first for us. It's the first time we've had a zoologist on the podcast.
David Brudenell: Oh, the boy. The, the good old university days. It still, still follows what is. And here we are talking about AI.
Nathan Bush: You've been in the field for a long time now. You haven't kind of just stumbled across it. I'm really keen, before we get into Decider, just to hear a bit about your career journey and your career path from studying zoology all the way through to leading decider. How do you describe your career arc so far?
David Brudenell: It didn't start in a zoo, that's for sure. Nathan. I would say like, I have done a few things consistently over the last 25 years. So thing number one is to look at challenges in frontier technology. Number two is always explore things that are intellectually challenging, which for me is not that hard. We know number one and number two are linked together and number three is a goal I set for myself in university, which was I wanted to travel the world and I wanted a business to pay for it. And those three things when combined together, there's many more that I've, you know, added as become more mature and more experienced as a professional. But those were the, those are the three things that continue, even continue to, to pull through my, my career. So yeah, that would be that part of the answer.
Nathan Bush: I like number three. I'm detecting an accent obviously. So how long have you been on the Australian soil?
David Brudenell: Yeah, you detected my amazing Australian accent. I've been in Australia since 2003, so I'm coming up to 25 years. So almost like I'm at the halfway point between, you know, my Canadian upbringing and my Australian upbringing, as I like to say. And you know, I spent a, I had a five year stint in the States, you know, during that 20, 24, 25 years or so. But yeah, Canadian born and Aussie bred, as I like to say.
Nathan Bush: Pretty much the same country. You're at different climate.
David Brudenell: They're extremely similar outside of the climate. Like same multi country, multi party government system, universal health care, great people, Commonwealth country, love sport. You know, Aussies like to go in thawed water, Canadians like to play on frozen water. Like there's a lot of similarities. There's a lot of similarities. I love it.
Nathan Bush: Now we're not here to talk about geography. We are here to talk about Decider and AI. And I was really interested when I started doing a bit of research around Decider. Came on my radar a little while ago. And I love around the claim that Decider is to be your go to decision system for teams everywhere. I love that line. Like, that's a really big promise. What does that mean?
David Brudenell: Sure. Yeah, sure. Well, first of all is many people don't have enough time to make decisions or you're focused on the less important decisions. The decisions that we think people should be focused on are the aspirations and the goals of your company, what your product should be, who your customer ICPs are, and those types of things. Unfortunately, many of the decisions we have to make are, can I afford to hire that additional person? Do I have to do this accounting and bookkeeping on a Saturday morning? The baloney, as I like to call it. The baloney. So in a world where, you know, we have this amazing new technology called artificial intelligence or generative AI that has all of the knowledge of humanity at your fingertips now. And when people, you know, when I say we as, I mean like the royal we here. So as we begin to start to learn how to use this technology, we believe that Decider, when used properly and we can get a bit, we can dive a bit more different, deeper into like what that means is we can start to get rid of the baloney and elevate teams to what we call higher order work. And that is, you know, making those more what we think are the more important decisions as a business runs. The example I love to give is that I'm sure there is a metric within Toyota of how many times the bonnet is opened in one of the Camrys. Right. And that that goal target is zero. Right. Because who thinks about, you know, maintenance of a Camry engine? Nobody. Right. Because they're so reliable. When cars were first invented, yes, we had to go under the bonnet and fix it and crank it and do all kinds of those things. That's what we're doing with. Yeah, that's what we're doing with businesses today. We're literally like in the bonnet, we're, we're cranking the engine, we're doing all those things. But you know, with a system like Decider, we want to be the Toyota Camry engine of businesses, like, of business.
Nathan Bush: Never heard someone so aspirationally wanting to be a Camry.
David Brudenell: I know, I know. Isn't that the worst?
Nathan Bush: But I get it.
David Brudenell: But you get it, right? So no one cares. But you know what you're doing when you're in your Camry, you're taking the people who mean most in your life, you're taking them safely, you're going on journeys, you're experiencing life. You know all the great things of every car commercial that's ever been made. Right. That's what you're doing. And we're today in businesses, we're doing all the mechanic work and we're not driving the cars. So that's what we believe Decider can do for businesses and business owners.
Nathan Bush: That's very different to a lot of other AI. I wouldn't even call them startups, but AI technologies that are in market at the moment where they're trying to be flashy and innovative, but it sounds like you're almost trying to solve old problems with new tech. Is that fair?
David Brudenell: We are, we are absolutely doing that. We have created like some of Decider's technology is what we call an agentic schema. So, you know, the extra Y would be like, we're the palantir for SMEs. So palantir is creating an ontology for government. Right. We've created like a schema for a business, the whole thing. Right. And each node in that is agentic. And we'll get into what agentic means, our interpretation of that. But yeah, absolutely. Where everyone is focused on point solutions, which I think personally is, you know, being in the industry and working on, you know, many of the large language models and knowing what it looks like underneath the hood of building these things, it's not surprising because, you know, trends, we all do the same things as humans over and over again. And, you know, we're the tailwinds of SaaS and VC. Investment in SaaS are strong. So it's not surprising that we're taking this inhuman technology and putting it into very human businesses that's like, you know, bricks and mortars to online. Like when websites were first created, you don't like literally have the same walkthrough as a David Jones or a Walmart or even worse. You then take the terminal and you put it into the Walmart. So you have people going like, that's kind of what we're doing now when we're jamming these point solutions into human businesses where we believe you need to like any new technology, you need to build an AI first or an agentic first system and then have humans to go into it. And that's what Decider's about. And you're absolutely right. It's a little bit Boring in the sense that we're just, we actually manufacture businesses. That's we're a manufacture. We manufacture agentic businesses and everyone comes out looking exactly the same. And then it. You tune it and based on your higher order or you're very human, which is taste, timing, tacit knowledge your way. Right. The je ne sais quoi of how you do things in the world. But if every business has the same foundations, that's when it gets really cool. Because they're all networked, they're all nested. There's a whole bunch of economic and other transactional benefits that come from the technology. But that like what we're all about here at the solder. And I started with the Camry engine, which is kind of funny. But the Camrys are everywhere. That's, that's a good thing. But we want to build the agentic economy. And the agentic economy is where these agentic businesses transact between them. Like they're actually, when we create them, they're all already networked. And so what does that mean for humans? It means that you have prospects immediately in your agentic business. You don't have to search for prospects anymore. You have vendor recommendations, you have these
Nathan Bush: things before we unpack agentic and we go down that path, which I'm really keen to do. Talk us through the foundation model. So if we kind of bring it back to those basic systems or those operating systems within a business, especially for e commerce businesses. One of the things that I understand about Decider is that you pull together all the platforms that hold the data to allow you to make decisions. What are the common platforms that you see come together especially for e commerce SMEs.
David Brudenell: Yeah. So I think what we look to is, you know, Zapier or the most popular connectors out there. You know, pretty much the top 20 in Zapier are the top 20 most popular apps or SaaS platforms and businesses. So they're already out there and they're hyper successful SaaS companies. So Shopify to Google Workspace to Xero to Salesforce to, you know, everybody would know the list, right? Yeah. That you know, with an e commerce company. And those are the ones that we bring together and like different from point solutions or what we call vertical AI. Decider is a horizontal platform. And what I like to always give the example of is, you know, all of US have used ChatGPT or Gemini or Claude before and when we put in a small prompt versus you put in a big prompt, you get a better result with a big prompt. Right. No one's going to argue, argue that. So that's context, right? That's the context window. And so AI is fundamentally a horizontal technology. So what we're doing is we're bringing all the horizontal data of a business and I'm going to simplify it. We don't chuck it into a prompt window. But it's the same type of logic. If when you give an instruction, but it has the entire context of your business, the instruction is much more valuable. It's your, you know, it's accuracy versus precision that comes out of it. So, you know, that's the fundamental difference between like Decider and many of these point solutions. But again, to answer your question, we connect with most and there's lots of API integrators out there. So every month we're adding more and more integrations to our platform.
Nathan Bush: And is it cross referencing and checking for accuracy across those platforms?
David Brudenell: Yeah, that's the cool part. That's the cool part. So as part of our, our schema, so I talked about that Agentix schema. So if you think about a customer, right? And you have Xero, which is your accounting platform, and you have Shopify, right? If you think about the customer name, which one do you want to take? You want to take your 01, right? And so our system creates a single source record for every customer, every product, every vendor and our system. And again, humans are involved with it. You know, agentic means humans and generative AI and conditional logic in our world. And so you go through that, we have a very interesting like, evaluation method that happens at every single cell in the system. And so kind of an outcome is with Decider, you get an agentically managed database. So and it extends the database, so it adds more columns. And it can do things like it's quite cool, so it can look between the nodes, so our nodes talk to each other. So zero nodes talk to shopify nodes. It can go into the horizontal data because it could find a proposal, right? Or an abandoned cart. It could go into the large language models of which we use all of them and it could find whatever's in there and go out onto the web and find public data. Or it could go to the human who put the record in because it has metadata so we know who is the owner of the record. And so imagine that going through scheduling, that using kind of gen AI at the right points, not all points, but at the right points, then you get this like essentially it creates an agentic single source database for you and it's constantly managed which is pretty exciting. And that's really why our agentic apps that sit on top of them are so smart out of the box. It's like it has perfect data to traverse all the time.
Nathan Bush: Yeah. And tell us more about those agentic apps. So how do they work? So assume that I'm running an E commerce business. What are the most common agentic uses that you find that sit on top of that data?
David Brudenell: Yeah, well I mean we cross the entire business spectrum. So from like front end sales all the way through like bank reconciliations, I say our, with E comm, our E comm customers, the natural first place to start is the, you know, is the front end and with customers. And those were some of our first deployments. Right. With E comm companies where I think the Luddite would call them a chatbot, I don't like calling them chat. We call them like sales apps or agentic sales apps. But they manifest in a chat style or a natural language interface. And today most of the chatbots or what's called AI agents that exist on websites, they're principally task based. So they have to follow an orchestration flow. Right. Which a human has to put together. And there's these big trees of like branching. Ours is not that because it sits on the agentic database. What we're about is goal engineering. So there's no prompting or engineering with our apps. What they just are is you give them a goal. So one app for an E comm company, we said we want it to grow revenue by 20% over 12 months time. That's the only goal, that's the only thing we told this thing. And because it can see across the entire horizontal stack and that's three pl, that's Klaviyo, that's Google Ads, that's Meta, you know the E comm stack.
Nathan Bush: The growth stack.
David Brudenell: Yeah, the growth stack. Within eight weeks it grew incremental revenue by 5% and after 12 weeks it's nearly 10%. Like the Sales Agentix sales app converts at 10 times higher than the website. What kind of things is it picking up? Well, I think so. It knows the products better than any human. It knows the ingredients in the products. It knows every competitor, all of their ingredients, all of their products. It has nuance. Like because it has, you know, the company's marketing documents and their vision documents and it has nuance. It has all the LLMs ability and grasp of language. It knows stock levels from the 3 PL right. It knows if a customer, a prospect who's on site looks like another prospect who is more likely to convert or less likely to convert. So those are all the really cool technical things, but the other non technical things are it's available 24 hours a day, you can transact within the chat. Right. It speaks 36 languages, it remembers you like. So there's just some basic stuff. I mean, I think we've all been on some sort of support call with a telco and you know, the telcos, they're throwing it out in the world that they have domestic humans who are managing those support queries. I feel sorry for those people. They're like handling 10 concurrent chats. You asked about like your router, which you got five years ago and what's wrong with it? And you can tell that they're just like, that's a painful. I'd rather just talk to this thing that's smart, that's read all the manuals from my router.
Nathan Bush: Well, when you think about E commerce, like the majority of customer service queries is where's my order? And I'm assuming that Decider is hooked up to a shipping platform as well as a E Com platform to be able to pull that information pretty easily. That would be bread and butter, Correct?
David Brudenell: Correct. Where's my order? You get a result snappy right away, right? And then there's a, there's a, there's service which comes from it. It's like, oh, would you like me to. If it hasn't arrived by the deadline, would you like me to notify you? Yes. Ari knows your email, right. It knows your phone number to text you. It knows all these things, right. And remember that's just a natural language interface. No one had to orchestrate anything in the system for that what we call agentic flow now to be deployed to that particular customer. So you know, we went from bricks and mortars from human LED full service to online, which was self service. And we're going back to AI full service, agentic full service. And so like E comm as well, like we have some stuff coming out later in the year where as you are chatting, the entire website is refactoring based on your chat, who you are and even your profile. And the product descriptions may be written differently for you in real time based on. It could be your demos, it could be your Personas, it could be whatever is the system believes is going to achieve that goal which you've instructed your system to do.
Nathan Bush: So you're bleeding across a few technology players in the E commerce space there still their lunch.
David Brudenell: Yes,
Nathan Bush: unapologetically, yeah.
David Brudenell: Well, I think if we take a minute to talk about the point solutions out there, Right. So there's a couple very clear mathematical issues that I would love to talk about. Then there's the actual kind of how a business needs to use these things. Like I'm a sales leader, I've been a CEO, I've done all the jobs. Right. And the stack that you have to put into E commerce is bananas. Right? How many layers, how many point solutions that you need to put in and how much wiring you need and how much, you know, data reconciliation, all that stuff, like it's a lot. So that's pain. The second thing is when you start to look at these point solutions, there's a few things I like to point out. Right. Number one is, remember 10 years ago when Ubers were really cheap? It was a great time, wasn't it?
Nathan Bush: And they were clean.
David Brudenell: Yeah, they were clean. Where they weren't all Toyota Camrys.
Nathan Bush: Nothing wrong with a Toyota Camry, by the way.
David Brudenell: Yeah, exactly. I'm not sponsored by Toyota, although I should be. I mean, I talk about it all the time, so. Well, why were they cheap? They were subsidized by venture capital. Right. And fast forward 10 years when Uber was like, we're going to have a profit. It was like, I don't know, how much? Six times, eight times? It's so much more expensive than it was. Right? Yeah. I took a Waymo when I was in San Francisco and It was like $9 to get from Pier 19 through to the city and I was like, oh, this is great. I was like, boy, this is not going to last how many years till I have to pay, you know, full rack.
Nathan Bush: And now they've set your standards on what it should be and you're like, anything below that is like, oh, I've got to take one of them.
David Brudenell: Yeah, yeah, exactly. But so venture capital subsidized Uber so that we download the apps, we be, we build the habit, right? And they get surface area tokens are heavily subsidized today. So that's number one. Right. Number two is when you use a point solution and many of them are just fine tuned. They're fine tuning the models, right? A generative AI, so they're almost pure gen AI going through. And so. Well, Genai is relatively unreliable. And well, what does that mean? So if you put a prompt in five different times, you're going to get five different results, right? Well, if you have a workflow and you have a 95% reliability rate on each step, right. After 20 steps, that's a 40%, 36% reliability rate. So that doesn't work for business. Right. You need 99% reliability per step. And so even at 99% reliability after 20 steps, that's only 80%, 83%. And so these point solutions are fine tuning and some of them are quite good. But when it comes to enterprise grade, the reliability goes down. So that's number one. And number two is the mathematical problem, which is big prompts are better. Right. Get better results. Well, the longer the conversation and think when I say conversation, that's also tasks. And a workflow is it has to bring the memory from each previous task into the next one. And many people when they say, oh, I just want to upload a customer record into my CRM, our human brains go, oh, that's four steps. No, like the heuristics in our brain are doing a lot of other things that you need to do multiple, multiple steps for a process. Right. And so it's not unreasonable to have a hundred steps, little tiny steps that go into just doing that. If you do 10 steps, it'll cost you 10 cents. If you do 50 steps, it'll cost you 250. If you do 100 steps it costs you $100. So there is a quadratic increase in token cost because of prompt chaining and token nesting that goes into it. And so I think where you in the future you get the push of like VC subsidization going away and you get prompt chaining. So naturally things workflows become more complex. Many of these point solutions, their business models actually won't, won't work. So we'll go through a similar phase that we went through like 20, 15 years ago is like whole explosion of all these point solutions. There'll be consolidation, there'll be consolidation groups and then kind of it all shake out. But again, the bet that we're making is not model native AI, which is what I just explained, but system native AI which is AI as infrastructure or build your AI infrastructure first, then figure out where, what points, what junctions do you want to have a generative AI working?
Nathan Bush: Yeah, it's an interesting point. I actually wrote about this in our newsletter a few weeks ago and I got mixed responses around. Imagine if you GPT and taking it to the most basic level, you ChatGPT addiction now cost you 30 bucks a month. Whatever it is, if it's 300, you're still going to pay it. Most of us have to because we can't work without it. Now if it's 3,000. Is that still worth it? Because it's going to happen. I'm not sure it'd be that extreme, but it's not going to stay the same as what it is now. One of the things that I found really interesting there with how you've set up Decider is around the goal oriented approach rather than a task oriented approach. If we go back to that customer service perspective there around using it as an agentic customer service support, obviously customer service can be, hey, do this as efficiently as possible for our team. Run it quick, run it fast, don't spend much time. It could be actually give the customer the best experience ever, go over the top, go above and beyond. Exactly those points that you said do follow ups, do all this sort of stuff that will cost slightly more, maybe take longer or there's also the option of actually we want you to solve the problem, but then also upsell and be a sales machine as well. When you are approaching something like customer service with Decider, how do you set a goal when you've got so many different options available to you know what I mean? Like how would you set that up?
David Brudenell: So you have to look at now a goal hierarchy within the organization. So at the very top you have aspirations, right? As a business, right? You have aspirations, you have your vision, you have your values. These are all like higher order goals that you set now underneath in the system. So you have to think about some of the things that are going on in our system at all times. So we have our schema, so everything is uniform and standardized. You have our gentic database, so everything is complete and expanding. So those things are now there, you know, data lakes. And the promise that's actually happens in our system now as you're setting and you're installing what we call logentic apps in the organization. Each app which has a series of workflows, they all have goals, right? And so for not giving the example of, you know, we want to achieve $20 million in revenue this year, if that's a goal for an E commerce company, that's not a goal. That's an outcome of many goals that sit underneath, right? And so we have to when we think about customers, so you get a great example of like, well, this customer you want to upsell. Right now we're thinking using our human brains to think about how to do that in Personas. In a system, AI has perfect memory, can see everything simultaneously and from an agentic perspective it will figure out the best way because it understands all the goals that cascade all the way up. Right. And it will figure out what is the best way to do that.
Nathan Bush: So then it'll standalone goals.
David Brudenell: Yeah. Sometimes it may mean to upsell, sometimes it may mean to send to a human customer service to give them the white glove treatment. Sometimes it would be do nothing. And so if you have a hundred thousand customers, you may have 100,000 unique experiences that are happen in there. And so in our system goal engineering and that doesn't mean you have to set a goal for every field in our system. That's not the many come within pre installed goals, but there are inputs in the system which change the goals. So if I use an example of if the keyword groups that you're buying are increased by 20%, the cost goes up by 20%. Like our system will push decision. That's why it's called decider. It pushes decisions to you and it says your strategy cannot be achieved for your goals and aspirations because it's 20% more expensive. Here are different ways that you can achieve the goals. It might take longer. Right. You might have to increase your aov, you might have to do a discount and then you choose the strategy that now the system's changed. And so I like to think because I'm a, I'm a Mac owner, remember the old time machine and you see like all the cuts of all your staves, that's kind of like what our system will give you. It is like here are all the futures that you could potentially choose. You can only choose one at the moment. Choose that one and then go ahead. And an option for that 20% would be do you have an extra N dollars that will allow us to pay more.
Nathan Bush: Right.
David Brudenell: Which is, that's a choice that you can make. Right. Which is I want to achieve my quantity, but I'll take a hit on my margin.
Nathan Bush: I mean that's essentially the role of a head of E Commerce or a senior E commerce manager, when you're answering into senior management is to give those options and those potential outcomes.
David Brudenell: Right? Yes.
Nathan Bush: Do you see it as a sidekick to those roles or do you see it as a replacement to those roles?
David Brudenell: Well, I think it'll be a, you know, a social media manager didn't exist 14 years ago. Right. That, you know, the people who did really well at EDMs, they probably still did really well in social. Right. It was just another, another pathway to do that. So, you know, I think some jobs will go away. Yes, absolutely. That'd be silly to say otherwise. Do I think many jobs will evolve? Absolutely. Do I think Many people will be exposed. I do, I do. So, you know, that creates like, what's interesting is like this AI is potentially an amazing equalizer. And what I mean by that is the best, right? The best of whatever you are. So I don't know about you, I've had. But there's always one person in the organization who's like amazing at customer service. And it doesn't matter if like the person's foot got chopped off, they'll still get, you know, they'll still buy both shoes. Right. But that's their business. Right. And now they can hyperscale themselves and their je ne sais quoi and they become a specialist and you install as a business. The person who I'm thinking about, her name was Christie. I would install Christie in thousands of businesses. Right, yeah. Into that. So I think the best e comm managers who just, you know, have that je ne sais quoi have that like taste, that ability to pick the thing. They'll be superstars.
Nathan Bush: It's also an interesting thing from a business perspective is that if you do have those superstars and say you have them in your business for two, three years because superstars don't last forever. They're usually, if they're that good, they're getting poached pretty quickly. But if you do have them for a good two or three years, you bake their process and their little magic into the system that can live on well long after they're gone.
David Brudenell: Right? Yeah, yeah, we are. You know, another way to think about decider is like we are doing what to knowledge work what manufacturing did to products, Right? You manufacture knowledge work in our system. And so what you're saying just said in another way. But you define what your amazing workflow is. And you know, that's the Nathan Bush workflow for podcasting and that's the best in the world.
Nathan Bush: No one wants that. It's not that professional. Anyone who's had a personal trainer knows that whatever targets you set them, they're going to raise it. So when Nutrition Warehouse looked to engage Klaviyo, it wasn't just about sending emails effectively. They tasked them with increasing revenue, driving awareness and building customer lifetime value. It's a spicy set. But Klaviyo, they were up for the challenge. By creating the right email flows triggered by customer actions, using geotargeting for their 110 stores and using predictive analytics, they proved that they were up for the challenge. Smashed it, in fact, with a 47% year on year increase in placed order rate on email flows and a Klaviyo Roi of 50 Times Nutrition Warehouse remain ahead of the pack. To explore how Klaviyo can grow your e commerce business and see more case studies, visit klaviyo.com au I'm really interested because you did some research that you released recently around Deciders AI Readiness Index and one of the stats that came out of there is that you found that 57% of businesses are prioritizing efficiency and 25% are prioritizing growth. With AI seems out of whack, why are we leaning towards efficiency rather than growth when we have these amazing new tools available?
David Brudenell: Well, let me start by saying that I am an Australian citizen so I can have a moan about Aussies just as easily as I can about Canadians. Aussies, we are as a country very parochial and we're very risk adverse. So surprise, surprise, right? Efficiency pops up at on top of the list and productivity is at the bottom, right. We live in this great country and we have this, it's both, I call it like the tyranny of distance when it comes to commercial activities. But is it, is it a blessing of distance when it comes to, you know, living here as a, you know, as a, as a citizen? And so because of this distance, we only as Australians think of our own country and the business that we can do in our own country. And so as a consequence, the deal flow is not that high, which means that the perceived risk is higher, which means that decision making is slower, which means that you have a, I believe a kind of a risk aversion first mindset. So I'm not surprised. I'm like, I'm literally, it's a, it's, I think it's a reflection of the Aussie commercial, you know, mind could have
Nathan Bush: saved thousands of dollars by not doing the research and just going, going on your gut feel.
David Brudenell: Yeah, Maybe, maybe hindsight's 20 20, but I think the, like, it's a sobering report, right. If anyone like wants to have a look at it when it comes to artificial intelligence. So like if you do believe, as one of the listeners, like if you do believe that AI will change the world, doesn't matter in which way or how you think. But if you think it's going to change the world is there's going to be really two types of nations. There'll be AI consumers and there'll be AI creators. And I think that in its current form Australia would be tragically an AI consumer. And when you think about again, if anyone's picking up on this like, this concept of like the agentic economy which I threw out before, which is, feels a bit buzzwordy, but it is an economy enabled by artificial intelligence transactions. I think that's definitely happening whether it's decider or somebody else. But if you place that, how that competes against a human native economy, like I know which one's going to win just based on, you know, if you look at the stock market, right. You look at, you know, electronic trading versus paper trading.
Nathan Bush: Right.
David Brudenell: Like obviously which one is better when it comes to that. So it's a sobering report. I'm excited to do the next round. So we'll do it at least twice a year. I'd like to. AI model development moves pretty quickly so we might do it even, even more than that. But I be curious, you know, three data points can indicate a trend. So I'm excited for our third index to come out so we can hopefully see that Aussies are a bit more focused on what they could be getting. Yeah, Globally, right? Yeah. As opposed to kind of what's happening in Melbourne surrounds or Sydney surrounds or out of bris Vegas.
Nathan Bush: I'm really interested in what you said there around the two types of economies going on and the agentic economy. Can you just paint a picture for us? Like what does an economy that is running as an agentic economy look like? 1 or 2 examples compared to an economy that's not. What's your vision for that?
David Brudenell: Well, I think without just getting too much in the detail describing like the mechanism, maybe it's like what are the benefits that come out of these economies? Right. And so if you think about what AI is proposed to unlock here in kind of Asia in, in the world is about 15 trillion worth of additional value. And if you look at GDP over the next kind of 10 years, that's estimated to be around 15%. So. And that's not linear. Right. So if we take a 10 year view it's, it'll probably be parabolic in the way it kind of moves. But if you're looking at GDP growth anywhere north of 2%, you are like a rocket ship, right. Of productivity and gdp. So what does that translate into? Well that translates into a higher tax base, it translates into more positive trade deficits, it translates into all of these really powerful things that make the sovereign nation a really attractive place to live, which in turn brings more people in, et cetera, et cetera. So and elevates the type of work. So it's this like a sovereign nation that has an agentic economy. I believe will have higher life and quality of life satisfaction with its, with its citizens. A country that does not, it'll be slow at first and then scary fast. Because the drain that will happen out of that kind of non agentic sovereign nation, like I think the world will move into like AI haves and AI have nots, to be honest. And like what does that mean for their education system? What does it mean for multigenerational wealth? Like there's a whole bunch of impacts that can come out of it, but I think it's inevitable coming. It's just how federal governments really start preparing for.
Nathan Bush: And obviously you are deep in the weeds in developing an AI solution and rolling that out globally. Do you think in terms of growth and being one of those nations that is able to move forward and being in the category that is AI enabled, do we actually have to lead in developing the technology or is it more around businesses adapting and using technology regardless of where it comes from?
David Brudenell: Yeah, I think it's the latter, Nathan. I think the federal government and state governments need to really move into how to equip essentially small businesses in particular with this type of technology or this type of thinking. And the reason why is because it's agentic, because it's goal seeking, networked, it actually feeds on itself. So if you are an agentic business, you want your suppliers to have your platform. So it's very value chain based. Right. Which is part of, you can't be
Nathan Bush: out there using AI all by yourself. You actually need the network of people using it to actually make it efficient.
David Brudenell: Correct. And but the network looks for the network. Right. It wants to find other ones and it will preference other agentic organizations before human ones because of speed, because of cost, because of reliability. Right. And it does that at an inhuman speed. Right. So you know, just like at the stock market, like it's been proven that there are micro crashes that happen within milliseconds in trading windows. Right. Because of these big algorithmic trading platforms are trading at so fast that no human could even see it. They have to like, you know, get a research scientist to look within a second and go, oh wow, there's like a, a crash we didn't even know, but here we are. Right. So, but yeah, I think it's a, it's a federal initiative. And you know, the federal government today, like just this past week there's been all this talk about copyright. I mean I wrote an opinion piece and that popped up in Capital Brief. And you know, my view is like, yes, copyright is important, like I don't want to diminish the impact of original work, but that's not really the problem. The problem is productivity. And how do we equip to. Your good question was like, how do we equip businesses to look towards productivity? Because the data, at least the coming out of our index suggests that they are not right. And you know what happens to risk adverse people. Like, we're not preppers, we're not, you know, building economic bunkers stocked with canned beans and two is new, but so we got to do something. You're a good prepper.
Nathan Bush: Putting two is new in there.
David Brudenell: Yeah. Phoebe, I don't know. I don't know. It depends on two weeks old. That's what I'd like. I'd put it down there.
Nathan Bush: Especially if you don't have a fridge plugged in. You'll go old.
David Brudenell: Not. Yeah, exactly.
Nathan Bush: Okay. Do you have much hope, do you have much optimism that the private sector is going to get support from the Australian government around growth using AI, or do you see the government sticking to more of a protectionism role?
David Brudenell: It's a good question. Without getting like too, too political. I think that, you know, over the past five years or so there's been a, a real keen focus on sustainability in this, you know, as a, as a national talk track and a national focus there. I mean, amongst other things, Australia doesn't have a great track record for enabling businesses and kind of getting behind businesses as compared to other other economies. So I would say I would be worried about, you know, the government's ability to effectively engage, gauge businesses. Yeah.
Nathan Bush: So don't hope the government's going to save you. It's essentially up to businesses, especially SMEs, to forge their own path.
David Brudenell: Here it is, you know, and I look to, you know, you hear things or hear or you read things, like in the uae, you know, every single citizen is getting a license to ChatGPT right now. Oil money can allow you to do those things however. Right. It's not that much money if you've got. How many citizens does I have? Like 30 million people, let's say cut it in half to the working population, you know, 15 million people is. Those are types of initiatives. Whether or not it's the right one or the wrong one, I don't think that really matters. But it demonstrates that the government's recognizes the potential impact of this and it wants to. It's finding ways to equip, I think that's the important word, you know, equip citizens in this case or businesses and others to be able to, you know, go through the step, you know, the very human steps, which is totally okay, which is like be exposed, learn, understand, apply.
Nathan Bush: The market is shifting, costs are climbing, and the pressure to do more with less is very real. But when resources are tight, you can rely on Shopify, the platform that's consistently first. With new capabilities, Shopify invests significantly in R and D and drops more than 150 updates every year. Evolving with you and absorbing complexity so that you can focus on what moves your business forward from AI powered insights to the world's best converting checkout. Shopify is designed for the next era of commerce, helping you sell more, scale faster and future proof your brand, build for what is next with Shopify. Visit Shopify for Enterprise to learn more. Well, that's really interesting you say that because if it is up to businesses, we need to be very deliberate around our decision making. So if we're choosing not to partake, it needs to be a deliberate decision rather than ostrich in the sand kind of behavior. Your research showed that three quarters of SMEs in Australia actually don't have an AI roadmap. And I think given our conversation earlier around heads of E commerce and senior E commerce managers with tools like Decider who might be laying out pathways for leadership in growth options or efficiency options, which is essentially their job, it's never going to be more important for leaders in E commerce to be able to own that roadmap, whether it, you know, and have AI tools alongside them to have confidence in that roadmap. When you look at what an AI roadmap might look like for an E commerce business, how detailed and laid out do you think that needs to be for an SME E commerce business given the uncertainty and the continual change in AI currently.
David Brudenell: Yeah, yeah, maybe just starting at that stat which you threw out, which was, which was correct. But I'd add one more layer which is we asked the question of everyone, what tools actually comprise your AI? You know, tool set and 9 out of 10 it was just ChatGPT or Gemini or Copilot. So let's also be honest, of that 25%, they're really not deep, deep practitioners of AI, right? Of AI tools and probably a lot
Nathan Bush: of them are being used under personal accounts because it goes against company policy.
David Brudenell: Probably, yeah, yeah, probably to your question is like, what does it look like for an SME? You know, I would go to number one is the people who create the roadmaps now, they're, they're the ones that are most likely to be adopted because there's, there's literally a vacuum of hey I roadmaps, so might as well be first even if it's not right when you first start. So there's tip number one, like observation number two or comment number two would be, you know, where do you start? You know, I would be sketching out on paper. If you want a system map, right. Of your business, I start there. And I would, underneath each system that you have put like what record do you want? You know, for Shopify it has customer records, it has transaction records. Right? It has product records. Yeah, yeah. So I'd be looking, looking at that and then going, okay, wait a second now take my labor, which if you're an SME, you don't have that many people in your organization typically and say like what do people do? How much do they spend in each one of these systems? Tippy tappies, tippy tapping stuff in into it. Right. And then number three is ask yourself the question is like, if I had 10 times more employees, what would I do? Now that's not to hire 10 times more, but that's like, what would you be focused on? That's the benefit of this amazing technology is that you can literally have an enterprise grade ability right now into your business. So it forces that kind of thought process to say, okay, well where is my tippy tapping labor going into, into these systems of which there's usually a lot when it comes to e commerce companies. And then, well, what would it look like? What would my business look like if I had 10 times people? And then, then you're going saying, well if I where is that tippy tapping going? Because how many of the extra people would I put as tippy tappers right into, you know, creating campaigns in HubSpot? And as you don't have to do those things, well, that's where I should probably focus on my, my AI.
Nathan Bush: That's a pretty disruptive thought, isn't it? Because most E commerce managers are asking their teams to do more with less or more with the same. And we've kind of got used to that ingrained thinking of like how do I just make my team do more and get better results with what we've got? We know the boundaries there. There's not too many e commerce businesses that are going, okay, if you triple your team tomorrow, what could you achieve? Yeah, that's interesting. If you could get rid of the tippy tappers at hotel check ins, that'd be great. Feels like they write a novel every time you check in, I'm like, what,
David Brudenell: are they tapping in? Yes, exactly.
Nathan Bush: Tell me more around Decider in terms of setup and commercials, what you can share here. So if people are hearing this and like liking the sound of it and going, hey, I'd like to explore setting goals for different agents to help me uncover new insights. We know we've got systems with some great data in it, but we've never been able to bring them together. What does it look like from a moment, customer puts their hand up and goes, we're keen to jump on board. What does it look like from a timeline, technical and commercial perspective?
David Brudenell: Yeah. So you know, if you're, I'll put a couple caveats around this. So like if you're an E comm company that has many of the most popular SaaS tools out there, if you have proprietary tools or older systems that may extend the time to connect into our system. But if you're a couple things. One is we're launching our platform called Decider OS in beta. That's in October. So we only have a couple hundred slots that we're launching with. So you can just go to Decider AI and you pay 99 bucks and you, you get into the preview slot, which is good. And so what it looks like is you put in your URL, that's where you start. And then our system, that seeds our system, right? And our system then starts building your AI Org in real time. It goes out onto the web, it finds all the public data that's available. You know, so if you are a service business, it calls it clients. If you're an E comm customer, it calls you customers. If you're a healthcare company, it calls it patients. Like it does all those configs automatically. It looks at all your products, it categorizes all your products. And if you know many businesses, E comms have up tons and tons of variants. Like every color is a variant, every T shirt size is a variant. And so it builds all of that architecture in the background, right, against our standardized schema and we call that pre boarding. So as you're pre boarded you kind of get a view and you'll probably go like, wow, okay, that's probably better laid out than I've ever seen it before. And then you start integrating or syncing your systems. And that is very light, light touch. So in some cases it's a one button push like Google workspace, right? Connect your Google account with Shopify, it's like get your API key you loaded in and then it syncs in Salesforce and things like that. As you add more that syncs the data and that starts preparing now the agentic single source or what we call the decider record of your data. Once you're in there as part of your onboarding now you're going to be asked questions in a natural language format about your goals and aspirations that starts way at the top, which you know, can be really fast if you know them or can be very hard questions because people like I've never been asked that question before for what are my aspirations, right? What is the, like, what is the true value proposition of my product to my customers and who, how would I describe my best customer? So actually what'll come out of it is, you know, many of the couple onboarding questions may be like, you may have to skip them because you're like I gotta come back to this or I gotta go into the chat GPT and actually I was about to say
Nathan Bush: have a second screen with ChatGPT open to give you the answer.
David Brudenell: Perhaps, perhaps. So you go through that onboarding and at that point, you know, through your pre boarding and onboarding that then opens up what we call agentic apps. And so there's an apps marketplace that are available and there'll be dozens in there. And an app is typically, it's not a big thing, it's typically a small thing or workflow. So that online sales app that I talked about earlier, that's actually a collection of a number of different workflows, right? So that would have like order checking, it would have in cart transactions, it would have a thing and you kind of, you link all those together and that becomes an experience for a customer. And again those are all just, you know, press the button and activate those apps and if you want you can go into the apps and you can see the workflows and you can go oh, our way does it this way, right? Instead of like A, B, C, we go acb right in that particular workflow. And so with the first preview customers that's what it will look like as our ecosystem begins to grow. Every app is actually a business. Like they are an agentic organization and so there'll be different versions of the same thing. So I use the example of like LinkedIn enrichment. So if you're at a B2B business you can sync all your contact records in and some of them will have LinkedIn URLs, some won't. And so there's a, an agentic workflow called LinkedIn Enrichment which adds a few columns, right, which costs five agentic workflows for row, right? And so run, all right? And it'll run only the ones that needs to. And it will add those columns, right? And you would only add them if you see value. Now, that from Decider might cost five agentic workflows, right? And it's all done by AI or our system. You may create the Nathan Bush LinkedIn enrichment, and you. You look at every single one of them as a human to make sure that they're accurate, right? And you will charge $5 every one that puts in, but it's verified by a human. And so there'll be this kind of expanding marketplace, just like on your iPhone or Google Play, where you'll be able to choose, right, what housing, like the best Gentex apps will win. Now, the cool part is because of our schema, because every app is also a business. When you create your app, the Nathan Bush LinkedIn enrichment app, it automatically knows how many blank LinkedIn records are in the entire system. And it will say like your target market in the decider ecosystem is 14 million records, right? And you're going to charge $5 for 14 million. That's a big number. But you're probably not going to get all those ones. And your AI org will send proposals automatically to those other businesses, and then you'll have your strategy for your business and you'll say, I want to get. These are my goals, these are my aspirations. And it will say, well, five bucks, it might be a bit too much, right? So if you want to put it at 15 cents, you'll do a million rows a month. And then.
Nathan Bush: And in that, I know this is just one example, we've kind of gone down the rabbit hole a little bit there. But in that example, say you're LinkedIn, right? And you're responding to this new world. Is LinkedIn gonna start charging you for access to that data?
David Brudenell: Probably. Probably. Yep. I mean, we're already seeing it with Reddit and New York Times and stuff like that, so. And I think LinkedIn has a right to do that, and I'll tell you why, right? So I think what the agentic economy is going to bring, it's going to remove information asymmetry. There are copies of me and you in hundreds or thousands of CRM systems, right? That's the asymmetry, right? Some have good data on you or me, some have poor data on you or me, right? The future, there is one you, right? And systems go into the you, right? To access that, that information. Now, LinkedIn is spent, has A great product. They spent lots of money. Of which we volunteer. We volunteer all of our important information. We've helped build that. Yep. So they might have the ability to charge for it now. A CRM system like Salesforce. No way. Right. That's like Swiss cheese to. At least to start. Like a real estate. Real estate domain. Yes. So I always look for what I call like an immutable record. So that would be a house address, that would be a tax record, that would be email. Google would have it. Some. Some. Yep. Not. Not all.
Nathan Bush: Not all. Yep.
David Brudenell: So because, you know, an email address that's not really immutable, like, and you can have multiple email addresses, there's not just like one for you. Social kind of the same. There's many people have many social accounts. Meta is kind of doing it with Instagram and Facebook and WhatsApp. That's like a. That's kind of a mutable ID across it. Okay.
Nathan Bush: So you see a big battle for the ownership of the one true record.
David Brudenell: Yeah, there will be. I mean, there's a piece of information that came up from Cloudflare just recently, which was like 10 years ago, a number of pages consumed per search query was like, it was for every query, one query, there was two pages crawled. Right. Then go kind of 10 years later, it was one to 300. Today, anthropic, it is one for every crawl. There are 60,000 pages crawled. So like, the nature of the web is fundamentally changing. So it actually might get locked down. You never know. Like, yeah, it might get locked down and gated everywhere. Just because these like the current, like model makers are just scouring the web. And also there's so much more information being created, which kind of goes back to the copyright thing that came up earlier, which is, you know, I think we'll start to revalue original content. But that's based on the removal of this information asymmetry. So when there's only one journalist who has a wonderful reputation and who has a unique take on the world is. The systems will then go into that and they'll have incredible value because of their audiences. Like, that's a product. Right. Just as it would be like a face cream or Camry engine. Yes.
Nathan Bush: We're back at Camry now. There is a really great case study on your site. We didn't even get to talk about properly edible beauty. If people want to see how it all comes together for an E commerce business. Because I know sometimes it's better when you get a practical or a real business in front of you and you see it. So I think that's a great one. We'll put a link to that in the show notes and a link to the site as well in case people want to have a click around and see if they can spot where the agentic is playing its role. But for you, David, what's the next 12 months hold? Where's your priorities for yourself and the team at Decider?
David Brudenell: Sure. So we are, I mean we're launching our beta in October and you know, that's the big release of our platform. So we're 5 million percent focused on that currently once we launch it. I think there's kind of, there's three things that we're focused on. You know, thing number one is to what are the agentic flows and what are the apps that everybody wants and what's cool about our system. It's no code, so we can build them. And we use our system like we're consumers of our own technology. So we every day find new gaps of like, oh, we should make that in an app, we should make that a workflow, so on and so forth. So I'm really excited for the agility to come out of our platform and for value to emerge right as this thing gets put out into the world and kind of starts agentically living, so to speak. That'd be number one. Number two is to continue our partnership strategy as a business. We embed our technology in many of the biggest SaaS, companies and cloud providers around the world. So we build their agentic apps for them and but the cool part is like that's an agentic app, but every one of our partners, customers, they get our whole platform. Like whether they know it or not, they get the whole platform. So there's this great kind of cross selling ability that comes out of it. So to really support our partners and to add more partners into our business.
Nathan Bush: What kind of partners are you looking for in particular?
David Brudenell: So again, going back to that immutable record kind of comment I made before, you know, a website is an immutable record. A hosting website at all. Accounting, partner, real estate.
Nathan Bush: Yep.
David Brudenell: We have those few to name off the top. And you know, coinciding with that is then, you know, our international expansion. So you know, we started it over in the States and it's now, you know, that's a. They have 32 million businesses are already in the U.S. that's a lot of businesses to look at. And because we're a horizontal platform, we're sector agnostic. So it's then like that's the exciting part which is like, oh, we have this ability to have a network system out of the box as pre boarding and as a no code platform to build apps. I'm excited for when companies are building apps that we've never even thought about before in specialized areas. Right. That's when I know that our business is really starting to cook. Where, you know, there's a specialist accounting or bank reconciliation, reconciliation workflow for plumbing companies. I'm like, I know that's. I know when we meet it, when the plumbers. Right, are.
Nathan Bush: When they're bringing new ideas. Yes. I love it. David, thank you so much. You've blown my mind on a lot of things that I've never thought about from an AI perspective. And I especially love the focus on growth and productivity rather than efficiencies. I think we get hung up on AI as a replacement rather than a growth engine. So thank you for that. Even as you're speaking then I'm like, it's going to achieve your third goal in your career, isn't it? You're going to be able to travel wherever you want with this thing. As soon as businesses get a taste of it, you're going to be flying all over the world.
David Brudenell: I hope so. I hope so.
Nathan Bush: Thank you so much for joining us on Add to CART today.
David Brudenell: Thanks, Nathan.
Nathan Bush: I really enjoyed that chat with David. It's not often that we get into politics on Add to cart, but I think that was a really, really welcome conversation, one that we need to have around AI because businesses aren't sure exactly on where it's going to fall in terms of government support. What I got out of that is that you do have a choice. Even if you're not participating in AI or you choose to be cautious, that in itself is a strategic business choice. So three lessons that I took out of that chat with David, number one, create an AI roadmap. And David gave some really great ideas on how to do that from starting to sketch out your systems. So mapping together all the technology that you use and all the sources of truth for your data, understanding what data it's holding and who's using those systems. And I love that that David called out the tippy tappers in our organization because we know that they are not going to be jobs that are around in the future. Who knows how long it is. They. But if you're a tippy tapper, you should be worried. But as a manager, we've got to think about the resources available to us and not only the resources available to US today, but what it will look like in an agentic world where it might be feasible to say, actually your workforce is 10 times bigger, even though they're not physical people. How would you use this to set yourself up for growth and to achieve our organizational goals? That in itself is a big question. Number two is about setting agentic goals, not just tasks. This was a big reframe from David. So a lot of us when we're using AI are thinking about, how do I get it to achieve this task that I've always asked it to do? However, what if we approach it from a sense of I'm doing this task to achieve this goal. What if I actually briefed the goal, not the task? And that's what Decider sets out to do to help businesses achieve those goals, regardless of what tasks are needed needed to achieve it. So by taking our thinking up a level into briefing for goals, not just tasks, it will unlock what AI can actually do for us. And the third thing that I took out of that conversation is what doesn't change is that you still need to know the goals of your organization. We often talk around AI is only as good as the data that's behind it at the foundation. We talked a little bit about that today. But it's also about knowing the strategic direction of your organization and where you're heading. If you are not clear on that, if you cannot describe that to your manager, to your team, good luck describing it to an agent. Because at the end of the day, as systems like Decider start rolling out, being able to articulate your strategic goals and outcomes that you want will be way more valuable than asking an AI agent to do a task. If you can't articulate your strategic goals today, make sure you go back and you get this right before you start unboxing new tools or new tech that may take you down the totally wrong path. If you enjoyed what you heard today and if it stretched your thinking and made you think about things in a totally new way, then you're going to love our ADD to CART community. We talk a hell of a lot in there around AI for ecosystem commerce and there's some of the smartest people in Australia joining that community, over 500 e commerce professionals. So come on in, discuss some of those themes and some of those ideas that we got out of today's episode with like minded people. You can join for free over on add to cart.com we'd love to see you in there. I'll see you again next week.
