The Intelligence Layer Live 📱

Welcome to the Intelligence Lair. I'm Vinny O Brien, and over the next 40 minutes or so, we're going to have the unglamorous conversation around AI. This one is more about plumbing than prophecy. One of the key questions for me today, is our data actually in a state to support any of this? For most B2B businesses, the honest answer is not yet. The ERP says one thing, CRM another, and then the storefront tells a third story. This is a follow-up to the inevitable shift to youth, YouGov fielded guide that we published with Glass Atlas, Commerce, and PayPal, looking at 500 UK B2B decision makers. 3 numbers from that work frame today perfectly. 44% of UK B2B firms are planning no investment in technology over the next 12 months. 42% of them have no e-com stack at all, and 52% of them identify as hybrid, which means in many cases, brochure website, manual quote, and some hope. That's the shape of it. Now, the people here to help me guide me through it. First up, it's Martin Rogers, Director of IT at Sweet Squared, the UK and Ireland's leading distributor of professional beauty. Martin runs a stack that thousands of independent salons rely on, every week to keep their own businesses moving. If anyone in this conversation knows what data orchestration looks like at the cold face, it's Martin. With Martin also is Pete Robertschaugh, Customer Solutions Director at ClassAtlas. Pete has spent 14 years bridging IT, e-com, and operations for retailers, which is to say he's the person who walks into the meeting where IT, e-com, and ops are all blaming each other, and he just quietly smooths it all over. And last but not least, the most ubiquitous man on YouTube when it comes to B2B commerce, Lance O'Wyde, VP of B2B at Commerce. Lance has spent more than a decade watching B2B merchants modernize, and the ones who crack it do it with really good purpose. Guys, longest intro of all time. Thank you so much for joining me.

Nice to see you.

Okay, look, I want to get straight into it and get started into the meat of this conversation. So Martin, I want to come to you first. Being in the industry, How much time and effort did you spend in the background in data preparation when it comes to B2B transformation in general, and in particular, looking at the AI side of things?

So I think for a lot of companies, the data challenges has kind of been there for many years. You know, when we look at the BI layer, you know, a lot of business, whether you be using Qlik, Power BI, or Tableau, to name a few. I think data challenges have always been there. I think it's just more prominent in the current marketplace with the emergence of AI. I think that the main difference for myself is that we're probably all used to open up a BI tool like a Power BI report and seeing that resulting figure and then just taking it as gospel, as the right answer, and expecting the same from AI. But what we're probably missing is It's probably then 3 days that you don't see where a business analyst is like, is crying into a dashboard because they can't make 2 data tables join. And, you know, in particular SweetSquared, we're a SaaS-first business. So we have practically no infrastructure, which, which on one hand, it's great because, you know, we put the reliance on vendors, you know, to, to provide uptime, security, and availability. However, on the other side, it does present clear significant challenges around integrations and data flow. However, with that being said, like I say, over the last few years with our journey around the BI side, I think that has helped us move into the AI. However, it's not without its challenges, that's for sure.

Pete, I want to come to you quickly because there's loads of definitions for what data means to a business. Can you tell me when we think about data, all those different hats that it has to wear within the business?

Oh God. I mean, it depends if you're in an e-commerce type of IT world. I mean, the data starts from whether it's your customer data or the data, like the product data, like, and the sources of this data and the businesses I work with tend to be all over the place. ARP has traditionally been the kind of the go-to, but like, it's not always fit for purpose, but trying to manage and organize like data flows as Martin just touched on between the different systems and then actually trying to analyze that data is just become a nightmare. And I know Martin actually in his business is, is done really well on this and the sort of the BI tool that he uses. But even then, um, trying to get that data into one place, um, and have it accurate as well, the trustworthy data is, is difficult. I mean, getting the data is one thing, but then actually being able to believe it's true is another. So I'm glad I'm not an e-commerce manager or head of IT anymore. I'm glad I'm on the other side of this because the amount of data that's out there now, um, I would not want to, uh, have to, to have that on my head. I mean, don't get me wrong, I work with a lot of businesses who've got like best-of-breed platforms for some of this data. Um, I do a lot of work PIM, which I guess we'll talk about in a bit. But yeah, the way things are now, trying to create data warehouses and put layers on top of it, it's not any mean feat. And I'm glad I'm the idea. I get the data I can do the analysis on, but I'd have to provide the data. Thank God.

Lance, I want to bring you in as well because look, you find businesses at different stages of maturity, not just in terms of their time in business, but also where they are in this journey. Why is data seemingly pushed to that last phase? Like, where is it in people's minds? And how are you helping them readjust to reposition the importance of data and getting the foundations right?

Yeah, I think that's a great question. I love what Martin said. I got this idea in my head of like, the outer join is dead, long live AI, or something like that. Like, it's the tables, we're going to get out of SQL and into something new. But to your point, Vinny, you know, when we are advising businesses, we always start with what we call a crawl, walk, run. You know, you don't have to bite off all of the elephant at once or all of the hamburger at once. You need to take little bites and little, little steps. And part of that is, is because, you know, ultimately we, you know, when we're working with, with customers and clients, it's to drive a business outcome. That business outcome might be to drive operational efficiency, right? And therefore reduce cost. It might be to drive new revenue, and that could be through new customers or to drive up share of wallet. But that's ultimately what we're trying to get to. And so when businesses start to plan transformation, they often want to skip everything because they're just trying to get to that outcome as fast as they possibly can. I have told the board that we are implementing a new e-commerce solution because that's going to drive new revenue because we're now accessible to a broader customer base. Sure. Yeah, that might be the outcome. But if you haven't built the foundations that mean that site has the data that a customer expects to see, you might as well not have done it. And you're going to run up against problems in the future, especially now as we start talking about AI. Because the issue with AI, and I think people often don't realize this, is it is going to expose those gaps. And I like to use a lot, and we'll get into lots of, I'm sure, detailed examples here. But I'm gonna pick a really simple one. If you think about semantic search, right? Semantic search, this idea that now as a buyer, instead of searching for a specific SKU, which I might be doing often when I know the SKU, I'm gonna start to search using a business problem, right? And that business problem might be, I have an issue with this HVAC unit, I'm looking for this part, or I'm looking for a replacement part for this automotive. That list goes on and on and on. How do you answer that question? The only way an AI can answer that question is with the data that helps it understand how your product solves the problem. And if on your website you don't have detailed descriptions, you don't have explanations of how that product can be used, you don't have metadata on that product and metafields that might be, you know, alternative descriptors, alternative SKUs, alternative identifiers. There's no point implementing these AI tools because they aren't going to help. And so whenever they're not able to answer the question, in fact, they're going to frustrate customers more because they're going to get really bad answers out of the tools. And so we are advising customers that they start by improving that data. And that's lots of different data. I mean, get into it, whether that's customer data, product data, quoting, invoicing, and so on. But that's really what I, what I see. You have to build a strong foundation. Otherwise, you're putting your customer-facing storefront on top of a weak foundation that ultimately is not going to support it.

Martin, on your side, was there a eureka moment or a tipping point at which you realized we need to start planning for this a little bit better? Can you remember that, or were you always working towards this type of project anyway?

Yeah, I mean, we've always been working towards better data, better data visibility. One of the challenges in a lot of businesses is understanding what data lies where and how it's useful. I mean, you know, we've probably all sat in them meetings where, you know, someone's quoting one sales figure from the e-commerce platform, a different figure from the GA, as an example, and then how that compares to the source of truth that might sit in the ERP system. You know, we see it quite often, a customer might place a an order on the website, but by the time it's hit our warehouse management system, they've phoned up to cancel the order, edit the order. We might have stock availability issues and, you know, resulting in a different value. And having that, that visibility and understanding what part each application plays is kind of crucial to us because the answer in that instance then is, well, well, we might take our sales value from our ERP system because that's the resulting true invoice value. However, our e-commerce platform still has a part to play because our ERP system won't tell us things like impressions, conversions, even down to the basic analytics of time of day, then overlay other aspects of, you know, if it would get into the CRM system and customer behaviors, that's going to drive our and influence our sales performance moving forward. And what we've always aspired to do is try and build all these applications and systems together. But like I mentioned, in, in a SaaS world that we live in, you know, it can be quite challenging because quite often with SaaS applications, unfortunately, you have to use the dreaded C word, which is compromise. And, and that unfortunately businesses have to learn to adapt to the way of applications rather than, you know, finding applications that are right for businesses, especially with, you know, SuiteSquared's a multi-divisional business. We we service our customers through multiple channels. So obviously Lance mentioned there about all the great things we can do from an e-commerce perspective, but it's paramount to us to make sure we've got that consistency across all our channels because it's all well and good us going to market in one way and providing them services, but if that's then lost in other parts of the business, it then has a knock-on effect to the customers and how customers are seen in the applications, then they're almost seen as different customers In total.

Lance, I want to bring you back in here because at Commerce Look, you have a huge data set globally, and that informs you really well to be able to help people get through decisions. And the four of us, we're all drinking the Kool-Aid anyway. How can you get into some of those specifics? You said there might be some examples, but are we seeing better performance for companies who are adopting early? Are we seeing them getting maybe better efficiency through the business? What are those two or three key things that you're seeing at a macro level in the commerce world today?

Yeah, well, there's a lot to unpack there in that question as well, right? Because everything is evolving so quickly. So we can answer that from kind of a traditional e-commerce perspective, but also from an AI perspective. But maybe starting, you know, from the more traditional, I think Martin touched on something that's really important there, which is, I, you know, I work for an e-commerce platform, we sell e-commerce. But we know that we are often, for many of our customers, a small sliver of their overall offering. To Martin's point, the business Sweet Squared is getting, you know, orders through, probably through phone, potentially through text, over the phone. They're getting them in, in, in, in lots of different channels across lots of different channels. Lots of businesses then, you know, manufacturers then selling on Amazon, advertising on marketplaces. We are a cog in that big chain of, of things that are happening. And so what we advise and what we see being successful here is that e-commerce can become what we like to think of as the, the, the center of those, those channels. So regardless of the channel that a customer is interacting through, they should be able to self-serve the information that they would need. Right? So if I've placed an order, maybe I've gone into the warehouse and I placed an order, I should see that order online. The minute I get home, it should be there. That means I can replace— I can, you know, if I need to return it, I can initiate that online. If I want to reorder that product, I can do that online. The invoice should be there and visible so that I can pay it online. And so you're creating this kind of unified experience across all of those channels such that it doesn't start to feel disjointed. And that drives a really great customer experience. And that's what we see being really successful. Vinny O Brien, one of our really big customers, $6 billion distributor, to me articulated it really well. He said, look, if we build the experience that customers love to shop with, ultimately that will drive share of wallet. And that's what we are looking for. We want to be the easiest place possible to place an order, but then actually also to follow up on that order, to request a return, to pay the invoice, to ask questions about the product. That's what you're trying to drive. Because that drives loyalty and that will drive share of wallet. And so that is really important. Putting on an AI hat just for a second, because I think it's Intelligence there.

I'm Vinny O Brien and over the next—

helps answer the position even further behind because there's never really been that need because B2B a lot of the time, like for Martin, your customers, you've got a lot of repeat buyers. They probably know the SKUs. They just come on and repeat buy. So they know what it is they're after. So having a load of detailed information about products, not, well, for B2B hasn't always been that important. Like again, what Lance talked about for like spare parts, those kind of queries you're getting now, like it's, it's not just about the SKU anymore. You have to have that, that product data, but not even just having like specs of products, like having just the weight of a product and just having the dimensional stuff. And again, it doesn't, it doesn't answer the questions that customers are asking for. If you use like a technology example of. I know my business travel, they need like a lightweight laptop because I'm always abroad and I need to have a long battery life. Having a load of specs on a table doesn't really mean a lot. But if you give that context of like, hey, this laptop's a 3kg laptop and it's got the battery life is 12 hours, 15 hours, rather than it's like 12 ohms or whatever, that means nothing to an LLM. So product data is important, but it's involved in terms of now taking that data and taking it to the next step in terms of how you're describing it and how you're, you know, to answer questions and your product copy is even more important. LLMs, like, you, like, look at citations of, like, there's data. So we're looking at sources of data. They typically look at things like 67% of queries that return answer a question in the first paragraph of data. So, like, having, like, your sort of spec tables buried at the bottom of your product data is not going to get you surfaced in LLMs. So yeah, I think that's been a shift. So I think it's, it's, it's the importance has grown over time, but now the shift has now been around, like, why should we invest in your product data, as an example. And in terms of customer data, I think, um, that's customer data in terms of in the context of AI. I guess where we've seen that's been more important is actually understanding who the customer is and serving more personalized experience and using AI to then to serve that experience. So matching up sales history and purchase history with like who the customer is, that personal, personalized experience. I know something Martin thinks we've been looking at as well. Is getting the right products in front of the right people. How that kind of translates to Agentic, that's a different matter. I have some questions and some challenges around how Agentic's actually going to perform a logged-in user role. Martin will know, like, you know, he doesn't expose all his product data necessarily to logged-out users. So I'm interested to see over the next few years how Agentic's going to handle this kind of like unified login layer. So if I'm logged into my ChatGPT or logged into Gemini or whatever, It does not log me into the Sweetscrab website, so it knows I'm logged in. But God, Lance has got his hand up. So, guys, this is happening already. Lance, is it you working on this already?

As soon as you said agentic, his ear twitched.

Well, well, I agree. What I will say, a lot of unknowns in that agentic piece specifically, but I agree with everything Pete says. And there's so much I want to add to that as well. But I just wanted to touch on that first point. And I know Martin came off mute, so I will hand him as soon as I've got this point out because I want to hear from him more than I think folks want to hear from me. I agree with you. Like, we don't know what that will look like yet, but what folks like we are doing at BigCommerce. So right now we are building exactly what you just spoke about, Pete, which is the logged in MCP, such that when you're in Gemini, you can be logged into SuiteSquared through our B2B APIs. And when I say we're building that now, that we're, we're building that now and testing it right now, such that the end of this month that will be publicly available and people can start using it. So, but we know that's where you've got to go because to your point, it is not the same in B2C and B2B. Your customers, Martin, have unique prices, unique catalogs. There's specific things they should only be able to see when they're logged in. And that means it's not available in Claude or Gemini. But what you can do is once the MCP is available, Gemini can speak straight to that, get you logged in and get that information. And that could be a public Gemini or that could be your own internal, you know, tool that you might build for yourself. So that's where that's going to go. But to your point, we're going to see that evolve. Pete, we're not quite there yet, but we're building the tools to make that happen because we know that's, that's what you, what you want. I saw you come off mute, man. So I know there's some great things.

Yeah, I was just going to add that, um, obviously as you know, with SweetSquared being a B2B predominantly distributor, part of what you're saying is correct. And you know, especially when it comes to Agentic, I mean, I see Agentic working very well for, especially in the, in the world of beauty for things that I want to shop the look as an example. So yes, that's great from like a B2C. B2C perspective, not as relevant for B2B, like Pete mentioned there, a lot of it's around the supply and demand. They're restocking based on their forward demand set by their customers. So they know what they're wanting. However, there's great use of AI tools to again, help speed that up. And not just that is, particularly in the world of beauty, we're seeing a big shift in customer behavior. You know, yes, we serve the B2B marketplace where, you know, we're selling some, you know, rockstar brands to some professional certified professionals, but we interact with them on a B2C basis. Even though they're B2B. And then it just depends on the type of business that they are and how they want to interact with us, because it's almost a point where for us we no longer see a difference between professional and consumer. And it's like this birth of a prosumer kind of world where, you know, even if I pick on like hair salons of all, pre-COVID, there was more salons across the country. But now we've moved into more of a model where it's rent a chair. So actually I'll go into one salon, I'll just rent a chair, I don't have to worry about managing teams. How I operate with that salon may or may not be different with different salons. Their buying behaviour will then be different compared to what it was prior to them. So again, coming back to the distributor viewpoint is we just need to make sure that we're agnostic to service all of them different viewpoints and all of them demands. So definitely one that we're keeping a mindful eye on then if we can start to, especially like Pete mentioned, you know, you have to be logged in and we also get our products as well through certification. So there's definitely some key challenges around there, but, you know, we definitely welcome any support as we said.

I'll go back to the agentic piece in just a minute. It's such an interesting conversation in general. The fact that as well, you're seeing the blending of behaviors between B2B and B2C. I think we could be entering a point where the B2B side of commerce is going to inform B2C. I think where typically we've kind of followed the other way, but something you described there, Martin. To me feels like the technology's got to be more fluid and agile than it has ever been before, because now we're talking about data accuracy. It means it can flow quicker and easier and has to, and has to behave in a way that people want in that moment. Has your attitude to technology changed? And what does your stack look like today versus where it was maybe 3 years ago?

So it's forever evolving, as we know, the marketplace, new technology always comes along. And obviously we need to ensure that we're ahead of the curve, but it always comes down to that underlying the, the, the so what factor, as we call it within the business. It's, you know, I can move technology today. You know, you can have the right platform, but still have a data problem at the end of the day. You know, I don't believe it's the stack sprawl that's, that's the issue. It's the orchestration of how they're put together. So what we need to understand is it's, it's what are we trying to address? And like I say, we need to understand our customer base in the first instance to make sure that we've got a tech stack supports them and all their different supporting needs.

So all of this is getting to the point that we made at the outset. We're developing this intelligence layer, Lance, and sometimes that can be amazing for us, but there's also times that it holds us back. And we're kind of at this frontier stage where we're forward-looking a little bit. So how are you telling that story without pushing too far into the future? Because we're all saying something similar here, at least what I'm hearing is It's not quite where we think it's going to be yet. It's emerging and we're watching these developments as they go on. And there's kind of a secondary part to that, which is about decision-making. I feel like we've got to make decisions in a much shorter timeframe now, almost like when we were in COVID where you had to make a decision in a snap second. How is the culture and technology choices changing right now as you see it? Because you're on the road every day talking to manufacturers and merchants and you're pretty close to it.

Yeah, I think I love it when I get to disagree with the host a little bit. Always fun and never necessarily a good idea to do that. You know, he's never going to send— he's not going to send the next question your way. I don't think you have to make decisions quickly. I think we feel like we have to make decisions quickly because everything feels like it's moving so fast because we're reading the news and Anthropic's IPOing and they're just releasing the next version of Opus and it's 4.8 and it's incredible. And oh my gosh, AI is taking over. And when we actually take a big breath and slow down, Martin says, you know what, the majority of my salons are still purchasing through my website. And they're still sending their emailed POs, and that hasn't actually changed. I feel like I've got a bit of time here. So I think there's a lot of pressure that we're feeling, especially as technologists, especially as e-commerce leaders, to have a plan and to figure out what's happening. 'Cause the board's putting pressure. They're saying, is, you know, there's AI happening. What's our plan? What are we doing? And actually, I think we have to slow down a little bit and say, what's really happening in the business? Are we really seeing change happen today? And for a lot of businesses, there hasn't been that much change. But I do think we have to we have to start to plan out for the future. And I'll paint a view here of what I think that future will look like. The timelines are completely variable, but you need partners like someone fantastic like Pete, or like a, a platform like BigCommerce that can help you through these stages. And so I view the world like this, which is, and this is, I think is Martin's business is a great example of this. Most businesses look like a sort of 80/20 type rule where 80% of your revenue is driven from about 20% of your customers. Those customers are your core customers, especially B2B. They're contracted. You know what they're buying. They have their set price. They're coming to your site or they're coming to your sales teams. That isn't changing. And the other 20% of that revenue is driven by 80% of the customers. Those are the long tail, as Martin sort of described it. They might be there now. They're just renting a chair. They look like a prosumer. They're a business, but really they look and feel more like a consumer. Those businesses, they have less loyalty. They could shift around and move. Why do I, why do I bring that up? Let's start with that, that prosumer. They might go to that LLM, they might go to Gemini, they might go to Claude to start asking questions about where should they get the products, can they compare pricing, etc. Now what do you need to start servicing them? You need to be visible. How do you get visible? Well, here's my one piece of advice for anyone watching. Go and ask the question you think your customer is asking in Gemini, in Claude today, in ChatGPT. Just ask the question and see what it responds. And then ask it what products do you think I should buy and where shall I buy them? If you aren't visible, you're invisible. And then ask it, why did you not pick my business? Why did you not pick Sweetsquared? Why did we not appear? It will tell you why, and it'll tell you why it picked the others, and then go and close those gaps. And I can tell you what it's gonna say. It's gonna say, we thought this other business was the source of truth because it was externally verified. It had reviews, it had testimonials, it was listed on Reddit, and it answered the question on its site. That's what it's going to say. And so you have to start plugging those gaps. And that's going to mean that you are visible. And then for your core customers, right, they're bigger businesses, most likely. What do you do for them? To Pete's point earlier, they're going to start looking at AI internally. How do they improve efficiency? How do they improve their workflows? And they at some point will come to you and say, hey, I know I come to your website today, but I would love it if my Claude agent just come to your site and place that order. How do I do that? And if you're on BigCommerce, hopefully you're going to put your hand up and say, hey, I've actually got an MCP. Here's how you access my MCP, and you can log in via the MCP and place the order or place the quote or find out where your delivery is, because all that information exists. And other ecommerce platforms are going to do the same thing. But that's how I think about the world today and the kind of, you know, the landscape. Are we there at the endpoint? And, you know, definitely not. But that's where we're going. And I think you want to be prepared. Don't be rushed into decisions, plan it out, put together your— not your full plan, because things will change. But how are you getting towards a place where you can support those customers, and where you think their experience and their buyer experience is going going to— is— how is it going to evolve? Can I just add to that?

That was nodding along and I was going to clap you because that's the whole thing about asking chat or whatever Gemini is about, asking the questions that we— again, I encourage my customers to go and get Claude or Gemini to go and scan the product pages, ask them like, why did I not come back in your results for like a suggested search? It will tell you how to optimize your product page so you do get service. It's not always more than one thing, but like, it's the easiest way to get into— if you think about years ago from an SEO perspective, you could ask Google and get honest answers about How can I become number one in your search results page? Like, they're never going to tell you because there's so many different variations, but like, they will tell you what to do. So again, from a readiness point of view, like I'm, I'm constantly telling my customers, if you're not sure, go and ask and you'll, you'll get the results. And there's guidance out there. There's free, you can go and get free geo audits if you really want to, but at a very base level, just go and ask the question about what do I need to do? And they'll tell you, and it's, uh, it's just the easiest win you'll ever get. And again, don't get me wrong, there's work to do and it will give you the answer, give you the steps to take, but. It's at least a starting point. 100% agree, Lance.

Maybe this is the good inflection point that we've reached where the access to the information we need to make change is so easily available now that we can start to make those changes. But Martin, as you started digging in and there was more of an onus on you to be prepared from an AI perspective, from an energetic perspective, were there any surprises that kind of jumped off the page at you straight away that you thought you didn't know about your business before you started doing this type of investigation?

I think Pete touched on a point earlier on the need for PIM. And obviously as we especially move into like the agentic world, having that consolidated centralized area for data to expose to things like LLMs are important. We've always known that we've got the breadth of data that we need. It's just potentially spread out amongst vast amount of systems. And that doesn't help AI. You know, all of a sudden I'm saying I have to build 15 MCPs rather than the one MCP just to give the answers to the customers that they need. That doesn't help me, nor does it help the customer. So one of the actual benefits, or the bit, the only, the biggest benefit in AI within SuiteSquared, and I might rock the boat a little bit because everyone always focuses typically around the LLMs, is we've actually been using AI to build applications that plug the gaps between our SaaS applications. So we've actually built our own PIM using AI. That is its own self-sufficient application. It has AI embedded in it being One of the big challenges that we've got as a business is I mentioned, you know, we represent some global rockstar brands, but most then of other distributors around the world are advertising the same product using the same product descriptions, using all the same dimensions because it's been given by the manufacturer. So that's an SEO nightmare. I mean, it's going to naturally then be an LLM nightmare. That is almost like we're shifting SEO into LLM. So Again, the product management system that we've built is then optimizing, using AI to optimize it based around performance driven from the insights given by our e-commerce. And don't get me wrong, we're no experts yet. It's still evolving and still growing, but it's like, like both Lance and Pete mentioned there, it's, we're using AI also to ask the questions to feed how we need to move forward, you know. Performance marketing is a great tool within AI to conduct our own audits. We've made a significant amount of changes. We're still going through that journey. There's still more to do, but actually through the use of AI, you know, we're able to use, I'll say, the capital that we might have used with a partner agency before to conduct this, to actually then make the investments that are right for the customers.

Lance, coming back to you again, because it feels like we're now having really rich and valuable conversations about how businesses run B2B or B2C at this point in time. The culture, I guess, around technology is really changing and we're getting more comfortable using it. How are you seeing the teams changing their attitudes and forming stronger opinions maybe when they're going to board level and having these conversations? You mentioned about taking a breath and slowing down, which I think is really, really valuable for people. I should try it myself sometimes, but What is the cultural shift that is helping drive all this forward that you can see?

Yeah, that's an awesome question. I love that one because it's often the piece that we don't talk about, which is how do I need to change my team to adapt to what is happening to the technology change that's shifting so fast? And actually, I was on stage at Commerce Live in Chicago about a month ago with some awesome, awesome brands and some awesome names, and one in particular Arrow Fasteners, which is kind of like a part of a conglomerate, billion-dollar conglomerate. And I had their head of e-commerce, Will, on stage with me, and he answered this exact question. And how— what is— what has he done with his team? And I thought it was really interesting. What he's done is he's created a safe space to learn and fail and to test and fail. And effectively, he said to the team, I'm going to give you the tools and the budget to go and test and learn. And whether that's plugging in gaps in our product information system, where we don't have the attributes and you think you can do that with AI, or whether that's building a customer-facing chatbot that you would love to put on the site that can answer questions. I'm giving you the team the freedom to go and play and learn and test. And I think it does two really cool things. Firstly, it gives— empowers the team to go and, um, take the initiative to see what they can solve, which I think is really key. But also it gives the team the freedom to go and learn. Because we're all learning today. These tools are new. We don't all know how to use them. We don't all know the edges of what that— of what's possible. But we need the time to play around and learn. And so I thought that was a really interesting way to go about it, because often what happens here is AI gets centralized. Oh, this is the AI team, or this is the AI czar. If you want something, go to them. That doesn't enable— that doesn't empower the rest of the team to go and to discover what's possible. You know, the e-com team are e-com experts. They're the people that will be able to determine how best to use AI to improve e-commerce, just like the product team. If there's a product team, they should be empowered to go and use AI. And so personally, that's how I think you should— you need to create the safe space away from prod, away from production. So we're not taking things down or causing problems. And we've seen all the recent kind of horror stories where we've carved off specific amounts of budget that we're willing to go and test with, not going to spend $500 million on tokens by accident. But that's what I think is really important. And that's the attitude that's changing. There's a lot more empowerment of individuals to go and test and learn. And those folks are coming back with some pretty cool answers to tough problems when given the time to go and play and test and be creative. And I think that's awesome. Like, I think it's an incredible time to be a technologist. Like, it's fantastic.

Pete, I want to ask a kind of follow-on to you, because as much as it's important that we have internal teams invested as well, it feels like it's a really good time to reset agency relationships and to get tighter with your customer base as well. Are you seeing people more open to hearing ideas from you guys for, and feeling more open to receiving insights and ways of using tools coming from, from yourselves on the agency side?

Yeah, I think there's an expectation actually from our customers that we, we're the experts and we know about all of this. I mean, we're learning as well. We are, as I just answered before, we're, I think we're being cautious in terms of what we recommend in our approach. I mean, we use, you know, we use, we're using AI internally for various things, but we are recommending it where it makes sense and for the right use case. Um, but our customers do come to us with, and when it comes to like platform requirements, now we're doing a discovery, uh, to kind of, um, get somebody live. It's like, how, what elements of AI can we use within the platform that's going to help? Most of it, when we're quite open and transparent about it and, um, how we use it. A lot of like data migration we'll use it for, and we're saying, look, you know, we've, we've now got more efficient processes to help you get from place A to place B, which affects how we can deliver the project and we can spend the budget elsewhere. So we're really transparent with how we use it, and I think again there's an expectation from our customers that we are, and the fact that therefore our project cost should come down. Um, it's not about even coming down, it's about them reallocating. They can do more, get more, get better value elsewhere. I mean, yeah, again, like L said, we've given our teams an internal capability to go and test and learn. And we've all got paid Gemini accounts, some ChatGPT, etc. So we are, we are embracing it, but we also, again, with caution that we did with, with every bit of technology that comes along as an agency, we've always not dived in. Like when headless first came around, I don't know how much to get onto this, but it's this talk, but anything that's new, we always approach with caution, take a step back and we advise our customers to do the same. But then equally, like there's questions about we don't want to be left behind. You know, we should be on the forefront of agenda. We need to make sure that people can go and buy our products from ChatGPT. And we're like, hang on a second. This isn't even like we're still like just on the periphery of this. Yeah, there's stories in the US around this, but again, we're talking about the readiness. So our message is like, get prepared for it as much as we can in whatever area it might be. But yeah, there's definitely, definitely for some reason, I guess because we work in the technology space, we all see it as like we should know everything there is to know about AI. And to a certain degree, in certain areas, the performance side of our business, they're kind of all over this. And from a geo perspective, yeah, we are talking to our customers about how to optimize from a for AI engines, etc. So from that side, yes, I think from a technology side where I sit in the commerce side, we are exploring ways that we can advise our customers to use and looking at features and also challenging the platforms as well. Like, this is what we're getting asked by our customers. Why are you guys doing about it? So, Lance, we might have to have a conversation because there's certain things we were getting this coming up time and time again. So yeah, I guess our role as consultants is to try and give the best advice we can, but without giving the wrong advice or bad advice and steering down a path they probably shouldn't go down right now. That's always been our stance.

And look, I think agencies, platforms, brands, retailers, we should all be able to have a safe place to have conversations like this and understand the space that we occupy in all of this as well, which is why I wanted to get into it a little bit with you. Martin, we're coming close on time and I want to get into a topic that I would love to hear your point of view on first, because there's two things or there's many things that we need to consider in all of this, but we haven't talked about governance and control. One of the areas that I see with all this caution, optimistic caution, let's call it, around AI is placing a really good impetus on us to rethink the roles of governance and control on how we do things and how we document them and how we place guardrails around them. What's been your approach as an organization to kind of compliance and governance when it comes to developing technology like this? And how would you advise others in your space to maybe start thinking about it?

Yeah, no, great, great question. And as an SME, uh, one of the many hats unfortunately I have to wear is as the data protection officer within the business. And data within any application is always— the security of it is always paramount. Uh, obviously I mentioned before about the BI side of the business. You know, within BI, it's great that, you know, we've outputted our data into data lakes and then we can use that data to create sub-datas and then reports off that data. So, so great. So we know what someone's looking at is tailored and designed for that specific user case. In the world of LLMs, you know, I might have an AI platform that's plugged straight into that data lake and looking at that raw data. Yes, it's great from the functional side that I am able to access whatever I need to get the answers I need. However, the other viewpoint to that is you've now given your full access and view to your entire data lake. How are you controlling that from a security and governance perspective? So we have to be very selective about what we're connecting and where. And my advice just comes back to, as I mentioned at the start, is understanding what roles applications play. You know, we might get our source of truth sales values from our ERP, but I wouldn't be getting my website performance from my ERP system and feeding it around. It's, it's understanding finding that tech stack and that orchestration and where they all sit and bringing that together. Obviously in August this year coming, we're going to see across the EU a new EU AI Act, and I think that's the start and a shift. So as much as it's probably presenting guardrails for businesses, so on one side of the coin, you know, there might be some frustration there around some additional steps that we need to However, within my role as director of IT, I am a lot more reassured that actually them guardrails are there for a reason. You know, especially, you know, I remember the years gone by when I always used to have to say to people, don't go on websites that don't have the padlock, as an example, just simple basics. It doesn't have— don't trust the site. Well, as it stands right now, there isn't any guardrails around AI that's out. I think this is probably the first step towards that. I mean, I think it's really baby steps at the minute, and I think it's going to evolve rapidly over the next couple of years, but it'll be in line with how technology is.

Lance, a connected question to that then. There's lots of standards that are emerging for the use of AI technology, all the acronyms you can imagine. But as an organization, how you like commerce are obviously really connected to making sure you're giving that trusted experience to everyone. How do you go about presenting this standard or adopting that standard roadmap for yourselves as as you look forward when it comes to AI?

Yeah.

Yeah, it's UCP, MCP, ACP. Yeah, ACP, AEO is in there, GEO is in there.

We've got—

there's a lot. And it can be difficult to understand and difficult to know where is the market going. As a business, that's our job. And actually, for B2B, I think this is really an interesting time, because a lot of the innovation that's happening in AI is happening on the B2C side of the house still, but I actually think B2Bs, and I think Pete, you mentioned, so I think B2Bs are going to get more value out of this ultimately because in B2B it's all about efficiency and AI is, that's efficiency. That's what it does. That's what it's best at. Whereas in B2C, there's an experience to a, you know, a shopping experience. I want to see the trainers that I'm about to buy. I don't just want to, I want to feel it. I want to get merchandise too. I enjoy that part of a purchase. Whereas in B2B, I, I don't want any of that, right? I just want efficiency. And so I think B2B might catch up and overtake here. But, you know, when unfortunately, when, you know, technologists think about AI, they're thinking about B2C first. And so us as a, you know, a technology company, e-commerce company, it's our job to stay on the forefront. And so personally, we are partnered with Google, with OpenAI, with Microsoft. We've built pipes into Perplexity. And into OpenAI so that the largest businesses out there can plumb their product data straight into those LLMs. Now, right now, a lot of that's walled off. It's sort of a walled garden. You know, we're partnered with Google for UCP. That doesn't do— it doesn't really do much today. We have to be honest and clear about that, but it will at some point. And so what you— I think what's important here is, is picking partners that get you prepared for that, that are willing to stand and have an honest conversation and say, like Pete said, I wouldn't worry about that today. Like, we'll get— let us worry about that for a bit. And when I think it's ready and when I think you should care, I'll come and tell you you should care. And I'll come and give you the roadmap that lets you get ready for it so that actually your customers can start purchasing using UCP through Gemini. But we will stay on top of it for now. And we are doing that because, you know, we own a company called Feedonomics. It is the number one product out there for syndicating product data into advertising channels and marketplaces used by 30% of the AI 1000. We have to stay on top of it. They are asking us, Dell is asking us every day, I want to be in OpenAI, I want to be Google, I want to be in Gemini, I want to be in Copilot. We have to solve it for them. And so when SweetSquared comes to us to say, hey, now I want it, now I care. Like, yeah, we solve this for the biggest. We're good. We've got you covered. And so I think that that's what it's about as a smaller, you know, especially as an SME, but even as larger B2B businesses, you don't need to stay on top of all of this. Find the right partners, whether that's the right agencies in Pete or the right e-commerce platforms or PIMs that are doing that work for you. And then that can advise you because we're doing it. We have those partnerships and we're helping to guide, you know, what does the next iteration look like?

We're getting close to the final question. Going to go around the horn and ask you one more time. But Pete, before I do, what's the one thing you want your customers to start thinking about or future customers when it comes to AI readiness?

I've got to say product data, focus on product data. That's, it's, that's the one thing you can start preparing for now. There's never a bad time. Like, I mean, also don't just do it for the LLMs or AI, just improve your product data anyway, because like your customers are going to benefit from it. Everything, like the entire experience around buying product comes from the product data you see, whether it's the images or whatever like that. I'm sorry to keep, I'm a massive product nerd. Um, but that, that for me is a bit, you say you should, this is, this for me should be a kick the backside to start actually focusing on it. And again, no disrespect to B2B and marketing specifically, but generally in B2B, it's not been brilliant. And the amount of businesses I work with now that are looking to make the improvements because other businesses in their space are doing it and they're not, again, not just for AI, they've been trying to start on this journey for the last sort of 2 or 3 years, like get the foundations of that right. Because if you don't like, and it's such an easy win, like getting product data, enriching product data or Use, again, use AI to write your product copy. Like, it's funny, there's this whole kind of like weird cycle of like using AI within a PIM to write copy that an AI agent will then consume. This whole thing is going to combust at some point. But like, honestly, just for me, that from Reddi's point of view, focus on that because it's, it's, it's, there's the side benefits to it anyway, without even thinking about AI. I'm quite passionate about that.

So that's the most polite a person from Manchester has ever given a kick in the backside to anyone. Martin, I'm going to come to you next. What are the signals people should be looking for as they consider their journey if they haven't started taking it now? What are those signs they can take to say now is the time for you to start making these changes or being inquisitive?

Yeah, I mean, Pete obviously mentioned there the importance of data, but, and it is no doubt paramount crucial because around our success to be able to deliver on, on what we need to, but for us it's understanding our customers. And, you know, if we don't understand what the customer wants, our competitors will. So what are them signs? It's we need to be aligned to what the marketplace wants and expects from a company like SuiteSquared. You know, we exist to elevate the professional. We can only do that if we've given them the right tools to deliver on that. You know, they want everything quicker. Better, cheaper, you know, that's just the world of B2B. We, we need to be agnostic to the needs of the marketplace because if we don't, we're just going to get left behind.

And, and Lads, I want to come to you finally because you've kind of summarized this really well and some of the discussion points that you've made. And I just want to maybe get a comment from you on this. You've kind of talked about the fact that there's no end point here, that we're always in this constant state of preparedness and getting ready for something. How do you encourage people to continue to not be exhausted by that and to be as enthused as you are about the world of B2B and AI?

For me, it is about taking, and I said this earlier, but it's about taking small bites of the apple. Take the wins where you find them. And I think this is really difficult because there's a lot of pressure coming, and I said this before as well, but there's a lot of pressure coming down from senior management, from the board, deliver, deliver, deliver. But you can do that. Just take small bites of that apple. It's like I— and we said it earlier, if you walk away from this discussion and do one thing, go onto Gemini and ask a question your customer would likely ask it and see what you get out. And then ask it how you would improve and take that one thing, whatever it says, take the top bullet that it gives you and go and do it. And take the other 10 that it gave you and put them on your backlog and do one of those a week. Or try and do one of those every 2 weeks, one of those a sprint, whatever it might be. Just take small bites at the apple and they add up and they will slowly but surely add to a, one, a truly cohesive strategy of where are you going, but to meaningful business impact. And I think that's the really important thing. Never lose sight of what it is are you trying to achieve as a business? Are you trying to drive revenue growth? Are you trying to drive share of wallet? Are you trying to drive operational efficiency? Everything we do has to tie back to that. Do it in small increments that you can manage, that show results as well, and you will feel good about it and you'll keep delivering and you'll build out that roadmap just like everyone else. And also don't think you're alone, right? There are lots of people out there that are facing the same problems, that are facing the same challenges. Find the content out there, find the associations out there that are talking about these. Well, that's like folks like the B2B Commerce Association, or others, find amazing partners that will share their best practices with you. And we're all in it together.

I would encourage anyone who's panicked about the world of commerce or AI to take the last 70 seconds, put it to some massage music and play it on repeat for a while. You will feel instantly calm just after that. But guys, thank you so much for sharing your time and expertise today. And to your point, Lanza, I'd encourage anyone to not just look for content, but to pick up the phone, speak to you guys, try and make time, learn from each other. It's the way we all get better. And I think the way I'd summarize today is that every day is a school day. But thank you all for taking time out. Look forward to seeing you around the block again soon and seeing where we get to in 6 months' time. Cheers, Vinny.

Thanks, everyone.

Cheers, guys.

Cheers. Bye, Drew.

Thanks, guys.